Referral and approach to care

National Guideline Centre (UK)

9Referral and approach to care

9.1. Introduction

The update to the heart failure guideline includes topics where new evidence has emerged since the publication in 2010. A review of new evidence published after 2010 was carried out in order to determine whether any changes to current recommendations where likely to be required. The decision on which topics to include in the update of the guideline was made following consultation of the scope.

This section includes multidisciplinary team working,transition and continuity between different heart failure care settings and information and support needs. The following topics were not within the scope of the update. For more information refer to Appendix R, the 2003 guideline:

Discharge planning
Support groups
Anxiety and depression

See See NICE’s guideline on depression in adults with a chronic physical health problem

9.2. Team working in the management of heart failure

9.2.1. Introduction

Heart failure is a complex disorder whose management involves a number of professional groups. Members of different professional groups contribute their experience and expertise to meet the complex needs of the patients. For the care to be optimised, the efforts of these professionals are best delivered through multi-disciplinary team (MDT) working. Multidisciplinary teams are well established to support people with heart failure in the UK.

A variety of models of care exist in heart failure and studies have documented the outcomes achieved by these different approaches. The role of the MDT in the care of patients with heart failure care was recognised in the NICE guidelines published in 2003. Since then has accumulated that has investigated the composition, competencies, needs for support, and the timing of different contributions of members of MDTs over a patient’s journey. This question sought to establish the competencies that ought to be present in a MDT to deliver optimal care for patients with heart failure as care of these patients increasingly moves into a community setting.

9.2.2. Review question: What competencies should be present in the multidisciplinary teams involved in the outpatient or community-based care of people with heart failure?

For full details see review protocol in Appendix A.

Table 97

PICO characteristics of review question.

9.2.3. Clinical evidence

A search was conducted for randomised trials comparing management involving multidisciplinary team care (MDT) with “usual care” without an MDT. Twenty two studies were included in the review⁶^#2495^,²²^,³⁴^,⁵⁰^,⁵⁶^,⁸⁷^,⁹¹^,¹⁰⁵^,¹⁰⁶^,¹¹¹^,¹³⁷^,¹⁵⁷^,¹⁹³^,²¹⁴^,²²⁰^,²⁴⁴^,²⁵⁷^,²⁷⁷^,³⁰⁰^,³⁴¹^,³⁵². Interventions were heterogeneous, and were categorised into 4 main ‘strands’ based on the definitions in a recent Cochrane report³²⁶

Home-based MDT: multidisciplinary teams from secondary care that included an aspect of caring for people in their home.
MDT clinic: multidisciplinary teams forming an outpatient clinic
Nurse-led or pharmacist-led clinic: enhanced outpatient service with MDT working
Case-management: active management of high-risk people with case managers taking responsibility for caseloads working in an integrated care system.³²⁶

Given the heterogeneous nature of the interventions and populations in the included studies, a number of additional tables have been included in this evidence report to summarise the key features of the included studies. Table 98 shows the included studies listed alphabetically with their population risk and length of intervention stratification, as well as their intervention strand. Table 99 lists the included studies by strata and provides details of the methods for each study. Table 100 expands on the population risk stratification for each study. Table 101 gives more information on the composition of the MDT involved in the intervention arm of each study. Table 102 gives more information on the intervention delivered, and on the “usual care” arm.

Table 98

List of studies included in the review.

Table 99

Study details (arranged by strata).

Table 100

Papers according to population risk: risk strata and background rate of admission (IQR 0.6-1.6, values outside IQR are highlighted).

Table 101

Professions / Competencies delivering interventions (studies arranged by strata).

Table 102

Key activities of MDT, studies arranged by strata.

Evidence from the studies is summarised in the clinical evidence summaries below (Table 103-123). No meta-analysis was conducted due to the heterogeneity in study and intervention designs; a separate evidence summary is provided for each study. Studies are grouped by the strata identified in Table 99.

Table 103

Clinical evidence summary: Ledwidge 2003: Short MDT clinic (MDTc) versus Primary +/- secondary care for high risk HF.

Table 123

Clinical Evidence Summary: Varma 1999: Long Pharmacist-led clinic (MDT pharm) vs Primary care (1 control) >6 months for low risk HF.

9.2.3.1. Clinical evidence summary tables

9.2.3.1.1. High risk, Short, Home-based MDT

No studies were identified in this category.

9.2.3.1.2. High risk, Short, MDT clinic

Table 104

Clinical evidence summary: Rao 2007: Short MDT clinic (MDTc) versus Primary care for high risk HF.

9.2.3.1.3. High risk, Short, Nurse-led or pharmacist-led clinic

No studies were identified in this category.

9.2.3.1.4. High risk, Short, Case-management

No studies were identified in this category.

9.2.3.1.5. High risk, Mid-length, Home-based MDT

Table 105

Clinical evidence summary: PREFER (Brannstrom 2014): Mid-length Home-based MDT (MDThome) vs Primary +/- secondary care for high risk HF.

9.2.3.1.6. High risk, Mid-length, MDT clinic

Table 106

Clinical evidence summary: Ducharme 2005: Mid-length MDT clinic (MDTc) vs Primary / secondary care for high risk HF.

Table 107

Clinical evidence summary: Gonzalez-Guerrero 2014: Mid-length MDT clinic (MDTc) vs Primary care +/- Geriatric clinic for high risk HF.

Table 108

Clinical evidence summary: Nucifora 2006: Mid-length MDT clinic (MDTc) vs Primary care for high risk HF.

9.2.3.1.7. High risk, Mid-length, Nurse-led or pharmacist-led clinic

Table 109

Clinical evidence summary: Driscoll 2014: Nurse-led clinic (MDTn) vs Primary / secondary care for high risk HFREF.

Table 110

Clinical evidence summary: Ekman 1998: Mid-length Nurse-led clinic (MDTn) vs Primary care (1 control) 3-6 months for high risk HF.

9.2.3.1.8. High risk, Mid-length, Case-management

Table 111

Clinical Evidence Summary: J-HOMECARE (Tsuchihashi-Makaya 2013): Mid-length Case management (MDTcm) vs Cardiology clinic for high risk HF.

9.2.3.1.9. High risk, Long, Home-based MDT

Table 112

Clinical Evidence Summary: COACH intensive (Jaarsma 2008): Long Home based MDT (MDThome) vs Cardiology clinic in high risk HF.

9.2.3.2. High risk, Long, MDT clinic

Table 113

Clinical Evidence Summary: Auckland-HF (Doughty 2002): Long MDT clinic (MDTc) vs Primary +/- Secondary care in high risk HF.

Table 114

Clinical Evidence Summary: Capomolla 2002: Long MDT clinic (MDTc) vs Cardiology clinic in high risk HFREF.

Table 115

Clinical Evidence Summary: DEAL-HF (De la Porte 2007): Long MDT clinic (MDTc) vs Cardiology clinic for high risk HF.

Table 116

Clinical Evidence Summary: Del Sindaco 2007:Long MDT clinic (MDTc) vs Primary / secondary care for high risk HF.

Table 117

Clinical Evidence Summary: PRICE (Atienza 2004): Long MDT clinic (MDTc) vs Cardiology for high risk HF.

9.2.3.2.1. High risk. Long, Nurse-led or pharmacist-led clinic

Table 118

Clinical Evidence Summary: COACH basic (Jaarsma 2008): Long Nurse-led clinic (MDTn) vs Cardiology clinic in high risk HF.

Table 119

Clinical Evidence Summary: OPTIMAL (Mejhert 2004): Long Nurse-led clinic (MDTn) vs Primary care for high risk HF.

9.2.3.2.2. High risk, Long, Case Management

Table 120

Clinical Evidence Summary: Berger 2010: Long Case-management (MDTcm) vs Primary +/- secondary care (1/2 control), for >6 months high risk HFREF.

9.2.3.2.3. Low risk, Short

No studies were identified in these categories.

9.2.3.2.4. Low risk, Mid-length

No studies were identified in these categories.

9.2.3.2.5. Low risk, Long, Home-based MDT

No studies were identified in this category.

9.2.3.2.6. Low risk, Long, MDT clinic

Table 121

Clinical Evidence Summary: Northstar (Schou 2013): Extended follow-up in MDT clinic (MDTc) vs Primary care (1 control) >6 months for low risk HF (stable HFREF).

9.2.3.2.7. Low risk, Long, Nurse-led or pharmacist-led clinic

Table 122

Clinical Evidence Summary: Agvall 2013: Long Nurse-led clinic (MDTcm) vs Primary care (1 control), >6 months for low-risk HFREF.

9.2.3.3. Low risk, Long, Case-management

Table 124

Clinical Evidence Summary: HICMann (Peters-Klimm 2010): Non-specialist case management (MDTcm) vs Primary care (1 control) >6 months for HFREF.

Table 125

Clinical Evidence Summary: Martensson 2005: Non-specialist case management (MDTcm) vs Primary care (1 control) > 6 months for high risk HF.

9.2.4. Economic evidence

Published literature

Five health economic studies were identified with the relevant comparison and have been included in this review.²²^,²²³^,²⁷⁴^,²⁷²^,²⁹⁰ To maintain consistency with the clinical review these have also been reported according to the clinical categories specified above. One health economic study was identified for the high risk, mid-length intervention, MDT clinic category; 2 health economic studies were identified for the high risk, long intervention, MDT clinic category; 1 health economic study was identified for the high risk, long intervention, nurse led MDT; and 1 health economic study was identified for the high risk, long intervention, case management MDT. No economic evaluations were identified for the remaining categories. These are summarised in the health economic evidence profiles below (Table 126, Table 127, Table 128 and Table 129) and the health economic evidence tables in Appendix G.

Table 126

Health economic evidence profile: Home-based MDT clinic (mid-length intervention) vs usual care in high risk patients.

Table 127

Health economic evidence profile: MDT clinic (long intervention) vs usual care in high risk patients.

Table 128

Health economic evidence profile: Home-based MDT (long intervention) vs nurse led MDT clinic (long intervention) vs usual care in high risk patients.

Table 129

Health economic evidence profile: MDT (long intervention - case-management) vs usual care (primary +/- secondary) in high risk patients.

One economic evaluation was selectively excluded due to the availability of more applicable evidence. This is listed in appendix J, with reasons for exclusion given.

See also the health economic study selection flow chart in Appendix D.

9.2.5. Evidence statements

Clinical

Overall 22 studies were included in the review. No meta-analysis was conducted due to the heterogeneous nature of the interventions and populations in the included studies. The studies were grouped into strata based on the population risk level (high, low) and the length of the intervention (long, mid, short), and were also categorised into one of four intervention types (home-based MDT, MDT clinic, nurse-led or pharmacist-led clinic, or case management).

High risk Short

Home-based MDT

No studies were identified in this category.

MDT clinic

This category included two studies with high risk populations exposed to a short MDT clinic intervention. For the outcome of all-cause hospitalisations, the evidence was inconsistent, with one study (n=98) suggesting a clinically important reduction and the other (n=112) suggesting a clinically important increase (moderate to low quality evidence). The evidence from both studies did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality.One of the studies (n=112) showed a clinically important increase in prescribing of ACE inhibitors and beta-blockers (high to moderate quality evidence).

Nurse-led or pharmacist-led clinic

No studies were identified in this category.

Case-management

No studies were identified in this category.

High risk, Mid-length

Home-based MDT

This category included one study (n=72) with a high risk population exposed to a mid-length home-based MDT intervention. The evidence showed a clinically important reduction in hospitalisations (high quality evidence) and a clinically important improvement in quality of life (very low quality evidence). Mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality.

MDT clinic

This category included three studies with high risk populations exposed to mid-length MDT clinic interventions. The evidence on hospitalisations was low to very low quality and inconsistent, with two studies (n=7117 and n=230) suggesting a clinically important reduction, and the third study (n=200) suggesting no clinical difference. Evidence on mortality was also low to very low quality and did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality.. The third study also provided evidence on quality of life and heart failure medication prescribing and adherence. Very low quality evidence suggested no clinically important difference in quality of life. The evidence on medications was moderate to very low quality and inconsistent, with evidence of a clinically important decrease in prescription of beta-blockers, but no clinical difference in prescribing of ACE inhibitors or taking prescribed medication. Most of the evidence had very wide confidence intervals around the effect estimates.

Nurse-led or pharmacist-led clinic

This category included 2 studies (n=25 and n=158) with high risk populations exposed to a mid-length nurse-led clinic intervention. For the outcome of hospitalisations, low to very low quality evidence suggested a clinically important reduction. Very low quality evidence suggested mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality.. The smaller study provided low quality evidence for quality of life which suggested no clinically important difference. Very low quality evidence from the same study also suggested a clinically important increase in prescribing of optimal beta-blocker doses. Evidence on all of the outcomes except for beta-blocker prescribing had very wide confidence intervals around the effect estimates. The larger study provided low quality evidence for NYHA class change and prescribing of ACE-inhibitors which showed no clinical difference in the former and a clinical benefit for the latter.

Case management

This category included one study (n=161) with a high risk population exposed to a mid-length case management intervention. Very low quality evidence showed a clinically important decrease in hospitalisations. Very low quality evidence on mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality.. Low to very low quality evidence on quality of life suggested no clinically important difference.

High risk, Long

Home-based MDT

This category included one study (n=683) with a high risk population exposed to a long home-based MDT intervention. Low quality evidence suggested a clinically important increase in hospitalisations . Mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality.

MDT clinic

This category included five studies (n=338, n=173, n=240, n=234, n=197) with high risk populations exposed to a long MDT clinic intervention. Moderate to very low quality evidence from the five studies consistently showed a clinically important reduction in hospitalisations or days in hospital. High to very low quality evidence from four of the studies also suggested a clinically important reduction in mortality or cardiac mortality in two studies,no clear effect in two studies and the fifth study suggested a clinically important increase.

Low to very low quality evidence on quality of life reported in three of the studies (n=220, n=210, n=197) suggested a clinically important benefit. Moderate to low quality evidence on prescribing of beta-blockers from two studies (n=153, n=210) suggested a clinically important benefit. Evidence on ACE inhibitor prescribing was inconsistent, with two studies (n=210, n=154, moderate to low quality) suggesting a clinically important benefit and a third (n=153, low quality) suggesting no clinical difference.

Nurse-led or pharmacist-led clinic

This category included two studies (n=679 and n=208) with high risk populations exposed to a long nurse-led clinic intervention. Evidence for hospitalisations was of moderate to low quality and showed no clinical difference in the former study but a clinical benefit for the latter. The evidence for mortality was of low quality and did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality. One study (n=208) also reported quality of life and medication prescription. The evidence for quality of life was of low quality and showed no clinical difference. Evidence for prescribed ACE-inhibitors and beta-blockers were of low to very low quality and showed a clinically important reduction for both outcomes.

Case management

One study (n=186) with a high risk population undergoing a long case-management programme was included in this category. Low to very low quality evidence suggested a clinically important reduction in hospitalisations, deaths and prescription of beta-blockers. Moderate quality evidence suggested no difference in the prescription of ACE-inhibitors or ARB. Quality of life was not reported.

Low risk, Short

No studies were identified in these categories.

Low risk, Mid-length

No studies were identified in these categories.

Low risk, Long

Home-based MDT

No studies were identified in this category.

MDT clinic

One study (n=920) offering an extended follow-up in an MDT clinic for low risk stable HFREF patients was included in this category. High quality evidence showed a clinically important reduction in admissions. Low quality evidence for mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality. The evidence for the remaining outcomes quality of life, prescribed ACE-inhibitors, prescribed beta-blockers, and three adverse events (serum creatinine increase, hyperkalaemia and hypotension) all showed no clinical difference and were of moderate to very low quality.

Nurse-led clinic

This category included one study (n=160) with a long nurse-led clinic for low risk HFREF patients. Low to very low quality evidence showed a clinically important reduction in the number of admissions. Mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality .Low quality evidence also showed a clinically important increase in prescribed ACE-inhibitors or ARB. No difference was found for the outcomes of prescribed beta-blockers and renal function serum creatinine, which were of moderate to low quality. Quality of life was not reported.

Pharmacist-led clinic

One study (n=83) with a long, pharmacist-led clinic for a low risk HF population was included in this category. The evidence for admissions was of very low quality and showed a clinical reduction for both. Mortality did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality .Very low quality evidence for quality of life also showed a clinical benefit for the intervention. Self-reported taking of prescribed medication showed no clinical difference and was of low quality. An automated measure of taking prescribed medication, however, showed a clinically important increase; the evidence of which was of very low quality.

Case management

Two studies with a long, non-specialist case management approach for a low risk HF population were included in this category (n=149 and n=190). The evidence for mortality was of very low quality and did not demonstrate a clear effect with the confidence interval ranging from a decrease and increase in all-cause mortality. One study (n=149) reported two medication outcomes: prescribed ACE-inhibitors at target dose and prescribed beta-blocker at target dose, both of which were of very low quality. The former outcome showed clinical harm for case management while the latter showed no difference. The second study (n=190) also reported a few other outcomes such as hospitalisations, which was of very low quality showing clinical harm for case management. This study also reported quality of life using the Kansas City Cardiomyopathy Questionnaire (very low quality showing clinical benefit for case management) and the SF-36 mental and physical components (low quality evidence demonstrating no difference for both). This study also reported prescribed double therapy of ACEI/ARB and B-blocker which was of low quality evidence showing no difference between the interventions.

Economic

One cost-utility analysis found that mid-length intervention MDT clinic dominates (more effective and less costly) usual care in people with heart failure who are high risk. This was assessed as partially applicable with potentially serious limitations.
One cost-consequence analysis found that a long intervention MDT clinic was less costly with mixed effects on health outcomes (hospitalisations: RR 0.67, mortality: RR 1.80, quality of life (MLWHFQ) MD 6.60 lower, prescribed ACEi: RR 0.97, prescribed BB: RR 1.62) compared to usual care in people with heart failure that are high risk. This was assessed as partially applicable with potentially serious limitations.
One cost-effectiveness analysis found that a long intervention MDT clinic was less costly and more effective (saves £4,042 per death and/or heart failure-related admission avoided; saves £2,155 per all-cause admission avoided) compared to usual care in people with heart failure that are high risk. This was assessed as partially applicable with potentially serious limitations.
One economic evaluation found that a long intervention, basic MDT dominates (more effective and less costly) both usual care and a long intervention, intensive MDT in people with heart failure who are high risk. This was assessed as partially applicable with potentially serious limitations.
One economic evaluation found that a long intervention, case-management MDT is cost effective (ICER: £3,072 per QALY gained) compared to usual care. This was assessed as partially applicable with potentially serious limitations.

9.2.6. Recommendations and link to evidence

Table

a lead physician with a subspecialty interest in heart failure (usually a consultant cardiologist) who is responsible for making the clinical diagnosis a specialist heart failure nurse

9.3. Transition between heart failure care settings

9.3.1. Introduction

Given the chronic nature of heart failure and its occurrence in populations in whom multiple morbidities are common, management of heart failure involves interaction between primary and secondary care services. People with heart failure are often admitted to secondary care settings with acute deterioration in heart failure (see NICE acute heart failure; CG187) and then transferred to the care of the heart failure multidisciplinary team (MDT) for management in the community. Once their clinical care has been optimised, people with heart failure are discharged back to the care of the routine primary care service but may need to re-access care thorugh the MDT. Evidence has accumulated that that the complicated nature of these transfer processes and the complexity of liaison between different teams can affect the quality of care delivered to people with heart failure. This question reviewed patient and staff experiences of these transfer processes and sought to find evidence on how these could be improved.

9.3.2. Review question: What are the experiences/preferences of staff and patients during transition between different heart failure care settings (including primary, secondary and community care)?

For full details see review protocol in Appendix A.

Table 130

Characteristics of review question.

9.3.3. Qualitative evidence

9.3.3.1. Methods

A search was conducted for qualitative studies exploring experiences or preferences of people with heart failure (CHF) or health care professionals (HCP) regarding transition and continuity of care at the interface of different heart failure care settings. Fifteen qualitative studies were included in the review;⁸^,¹⁴^,²⁹^,⁴⁷^,¹²⁴^,¹²⁶^,¹²⁷^,¹³³^,¹⁴⁴^,²⁰²^,²⁰⁷^,²⁰⁹^,²²⁵^,²⁹⁴^,³²⁵ these are summarised in Table 131 below. Key findings from these studies are described in the qualitative evidence synthesis (section 9.3.3.4) and summarised in the qualitative evidence summary (section 9.3.3.5). See also the study selection flow chart in Appendix C, study evidence tables in Appendix F, and excluded studies lists in Appendix I.

Table 131

Summary of studies included in the review. Abbreviations: Pt = patients with CHF, HCP = health care professionals, NYHA = New York Heart Association Functional Classification.

9.3.3.2. Qualitative synthesis

The guideline committee was aware that the issue of continuity is covered in the “Patient experience in adult NHS services” NICE Clinical Guidance February 2012 (CG138)²³³, and wished to build on that guidance. Therefore, this review used the framework of the recommendations from that guidance to synthesise the heart failure specific findings from this review, while also collecting heart failure specific themes not covered in that guidance.

The most relevant recommendations from CG138 are as follows:

Assess each patient’s requirement for continuity of care and how that requirement will be met. This may involve the patient seeing the same healthcare professional throughout a single episode of care, or ensuring continuity within a healthcare team. (Continuity)
For patients who use a number of different services (for example, services in both primary and secondary care, or attending different clinics in a hospital), ensure effective coordination and prioritisation of care to minimise the impact on the patient. (Co-ordination)
Ensure clear and timely exchange of patient information: (Communication)
- between healthcare professionals (particularly at the point of any transitions in care)
- between healthcare and social care professionals (with the patient’s consent).
Give the patient (and their family members and/or carers if appropriate) information about what to do and who to contact in different situations, such as ‘out of hours’ or in an emergency. (Contact and Access)

This review uses these recommendations to define four findings in which to group the subfindings of our review. We fitted evidence to this framework, and separately collected evidence that did not fit into this framework, as shown in the figure below.

Figure 6

Framework for synthesising review results.

9.3.3.3. Summary of included studies

9.3.3.4. Qualitative evidence synthesis

9.3.3.4.1. Narrative review findings

The review findings were grouped within five overarching findings, as follows:

Continuity
Co-ordination
Communication
Contact and Access
Role of specialist heart failure services

Within most findings there were sub-findings relating to (a) challenges/problems, (b) ideas for improvement, and (c) complex transition issues. The sub-findings are presented, along with an evaluation of the quality, including an explanation for the CerQual grading that appears with each sub-finding in the summary tables in 9.3.3.5. Relevancy was judged with respect to making recommendations on this particular issue for current NHS specialist and general services, taking into account that some evidence came from studies were conducted in other countries (studies from US and outside the OECD were excluded) and before major NHS configuration changes.

1. Continuity

Review sub-finding 1a: Lack of continuity in HF care

Both people with heart failure and health professionals felt that heart failure care was fragmented, leading to a lack of continuity. This was contrasted in one study with the more seamless care offered to people with cancer. Patients who experienced lack of continuity found it made it more difficult to form therapeutic relationships, which undermined their confidence in their management plans. Patients aged 65 and under mentioned that lack of continuity sometimes led to encounters with professionals appearing anonymous and meaningless.

This sub-finding was based on four studies¹²⁶^,¹⁴⁴^,²⁰⁷^,²²⁵.

Methodological limitations were rated moderate concerns overall. Two papers were rated serious limitations as they were secondary analyses with insufficient information about the original interviews. One paper was rated as moderate limitations due to lack of context and data richness. One paper was rated as minor limitations.

Coherence was rated no or very minor concerns overall. Agreement between papers was good and appeared to fit well with related findings in other studies. The only findings that tend the other way are regarding the positive influence of GPs in continuity, and this is not inconsistent.

Relevance was rated minor concerns overall. The relevance of the sub-finding to the focus of the review was high. Concerns stemmed from two papers being from outside the UK and in a sub-set of CHF, and the data collection in a UK paper dating back to 2010. However, the finding was found in various settings, and likely to generalise well.

Adequacy was rated as no or very minor concerns. The sub-finding was general, and although explicitly supported by a limited number of studies, other studies found similar findings. The sub-finding was descriptive and the richness of data was sufficient for this.

The overall assessment of confidence was moderate, having been downgraded by one increment due to the combination of concerns in methodology and relevance.

Review sub-finding 1b: Primary care

People with heart failure and professionals working in heart failure care both recognised that primary care, and general practitioners in particular, could provide individualised care with continuity. It was acknowledged that such continuity is important to patients, and there was concern that where HF services take over a patient, this could cause patients to lose touch with primary care, with a consequent loss of continuity. In one study GPs spoke of how they wished for a “consultant” model where specialists answered questions and made recommendations, but did not take over care of the patient. In an assessment of a new heart failure service, it was emphasised by GPs that new services should integrate into the existing primary care/community multidisciplinary team, partly for this reason.

This sub-finding was based on 4 studies²⁰²^,²⁰⁹^,²²⁵^,³²⁵.

Methodological limitations were rated minor overall. One paper was rated serious limitations due to the lack of explanation of methods and lack of depth of data collection. One paper was rated as moderate limitations due to the participants not being well described. Two papers were rated as minor limitations.

Coherence was rated minor concerns overall. Agreement between papers was good, showing that both patients and other health professionals value the continuity provided by GPs, and believe there is danger of that being disrupted. However, there is other data that challenges the primacy of continuity over other important factors in HF care, for instance the in-depth knowledge of the cardiologist, or the easy access to the HF nurse.

Relevance was rated minor concerns overall. The relevance of the subfinding to the focus of the review was high. Both studies that included patients were from outside the UK. A third study (of HCP) was from the Highlands of Scotland, which has rather specific challenges of geography that may affect the importance of primary care. The remaining study however, was conducted with HCP in three different trusts, offering results that seem to generalise well over settings.

Adequacy was rated as no or very minor concerns. The subfinding was fairly general, was expected, and is supported by numbers of studies. The finding was descriptive and supported adequately by the richness of data.

The overall assessment of confidence was moderate, having been downgraded by one increment due to the combined minor concerns about methodology, coherence and relevance.

Review sub-finding 1c: Discharge from HF clinic

One study that interviewed people discharged from HF clinics to primary care found that they reported having gone from a situation where they received appointments for follow-up, to one where they were no longer called for review. For some, this felt like they were no longer ill enough to qualify for help. When asked, they expressed a wish to be called to see the GP or nurse once in a while. One patient said: “I feel a bit left out, I’m not part of the health-care system anymore…” p291¹⁴

A different study speaking to HF nurses echoed this, saying that they were aware that people liked being patients in the HF clinic, and that this led to some unhappiness when it came to discharging patients. Regarding HF clinics, one says: “I see quite a lot of patients who get discharged, and they hate that. They feel safe… someone’s interested and they’re keeping an eye on me… They don’t like to lose that. They hate to lose that.” P301²⁹⁴

It is notable that the focussed qualitative review on continuity in CG138 found the finding of “Feelings of abandonment (when treatment ends or support not available)”, which seems to echo in this specific finding on discharge from HF clinic to primary care.

This sub-finding was based on 2 studies¹⁴^,²⁹⁴.

Methodological limitations were rated serious overall. Two papers were rated serious limitations – the study of patient interviews because of the small number of participants (4), lack of clarity over methods, the researcher’s voice is potentially prominent and lack of richness of data; the study of HCP interviews because of limited discussion and reflection on the role of the researcher and their methods.

Coherence was rated no or very minor concerns overall. Agreement between papers was good, and while it does not directly align with the sub-finding concerning the primacy of primary care in providing continuity, it fits with the overall theme of this section regarding lack of continuity within the system as a whole.

Relevance was rated minor concerns overall. The relevance of the sub-finding to the focus of the review was very high. Concern was due to one study being from outside the UK and the interviews being in Swedish. However the UK study of HCP shows that the issue is likely to extend to NHS services.

Adequacy was rated as moderate concerns. The sub-finding was specific, and so sufficiently supported by lower numbers of studies. The issues are well described and explained, but overall the richness of the data falls short.

The overall assessment of confidence was low, having been downgraded by 2 increments due to methodological limitations, and concerns over adequacy and relevance, while taking into account the support of the external literature.

2. Co-ordination

Review sub-finding 2a: Poor co-ordination between services

People with heart failure experience the healthcare system as being poorly co-ordinated. One consequence of poor co-ordination was increased treatment burdens, as they needed to attend hospital/clinics on multiple occasions. They also experienced being given conflicting advice, for example by the cardiologist and nephrologist, leading to uncertainty and loss of confidence. People with heart failure also found the organisation of the healthcare system unclear and confusing. Individuals recalled the difficulties they had experienced trying to get referrals to rehabilitation services and nursing services that they needed. They suggested a clearer organisational process where they could identify who had responsibility, and where things like rehabilitation were offered automatically.

This sub-finding was based on six studies²⁹^,⁴⁷^,¹²⁶^, ^#2771^,²²⁵^,³²⁵.

Methodological limitations were rated moderate overall. Two papers were rated serious limitations: a secondary analysis which did not give enough detail about data collection or context, and an interview study due to a combination of lack of researcher reflection and inadequate richness. Two papers were rated as moderate limitations due to a lack of researcher reflection, inadequate richness and lack of discussion. Two papers were rated as minor limitations.

Coherence was rated no or very minor concerns overall. The sub-finding of poor co-ordination is present throughout these six studies, and is implicit in other studies. All positive experiences encountered in this review were of individual teams, and most negative experiences in the studies are between settings – although many of these were between inpatient hospital and community, which is outside the scope of this review.

Relevance was rated minor concerns overall. The relevance of three studies was reduced as they were from outside the UK, two of the remaining studies are from before 2010, and the date of data collection from the final study is not reported. There could be concerns as to whether this is occurring in today’s NHS context; however the breadth of the data, being from multiple countries, multiple settings, and multiple time points, increases its generalisability.

Adequacy was rated as no or very minor concerns. The sub-finding was general and supported by numbers of studies. The sub-finding was descriptive and supported by the richness of data.

The overall assessment of confidence was moderate, having been downgraded by 1 increment due to concerns over methodology and relevance.

Review sub-finding 2b: Models to co-ordinate care

Clinicians acknowledged that closer co-operation would improve patient care, and potentially reduce workloads. Different (non-mutually exclusive) suggestions came from within and across papers:

Nominating a single professional to co-ordinate care.
Heart failure nurses working in a ‘cross-boundary’ role to encourage close working relationships.
Co-working between primary and secondary care, possibly through means of a ‘shared-care agreement’. Such protocols were already in place for people who have other long-term conditions such as diabetes and hypertension, and enable general practitioners to manage patients in certain categories, with specialists managing patients presenting with more complexity.

Protocols such as for ‘shared-care’ were seen as having both pros and cons: they can improve transparency and facilitate co-working, but people with heart failure often also have other chronic problems and could end up being on multiple protocols, leading instead to confusion and increased burden of treatment.

This sub-finding was based on five studies¹²⁴^,¹³³^,²⁰²^,²⁹⁴^,³²⁵.

Methodological limitations were rated serious overall. Two papers were rated serious limitations: one due to lack of context and richness in our area of interest, the other due to lack of detail and reflection on the methods. Two papers were rated as moderate limitations: one because the aims, methodology and findings were poorly discussed, the other because there was little description of the participants. One paper was rated as minor limitations.

Coherence was rated minor concerns overall, based on the finding that clinicians see a problem and think it can be improved. Although there were a number of different improvements suggested, they were compatible with each other. It would be impossible however, to conclude in favour of any particular intervention from these data, and it is clear that there is a tension between disease-specific protocols and holistic management.

Relevance was rated minor concerns overall. The relevance of the sub-finding to the focus of the review was high, as these models may have the potential to improve patient experience at transitions. All but one study was from the UK, and they were in a number of different settings, although some are from before 2010.

Adequacy was rated as moderate concerns. The sub-finding was general, and found in a number of studies, but there was inadequate depth (for example, there were no case studies or examples of pathways in action).

The overall assessment of confidence was low, having been downgraded by two increments due to concerns regarding methodology, and adequacy along with other minor concerns.

3. Communication

Review sub-finding 3a: There is poor communication between services

People with heart failure, their informal carers and professionals within and outside HF services all felt that there was poor communication across boundaries, especially between hospital-based services and primary / community care. Inadequate and delayed transfer of information led to burden on patients and waste of resources. In some areas it contributed to lack of / under-treatment, where management plans were not shared with the prescriber.

One patient explained how this happened: “…there were times when there was a bit of a lack of communication, you know. I would go speak to my doctor [GP] and tell him I’m on such and such, and he would say ‘I’ve not had any word about that’. And that’s part of the problem; you get a printed prescription that has, for instance, if I was on 4.5mg of Bisoprolol and they [hospital] were putting it up to 7.5, often the liaison between the hospital and the doctors wasn’t all that good”. p107²⁰⁷

This sub-finding was based on six studies²⁹^,⁴⁷^,¹²⁶^,¹³³^,¹⁴⁴^,²⁰⁷.

Methodological limitations were rated serious overall. Three papers were rated serious limitations, including two secondary analyses and one with limited detail and reflection. Three papers were rated moderate limitations, two due to lack of richness and context, and a third due to poorly defined aims and method.

Coherence was rated no or very minor concerns overall. Agreement between papers was good, and is supported by many of the other studies in an implicit way.

Relevance was rated no or very minor concerns overall. The relevance of the sub-finding to the focus of the review was very high, as communication was flagged as an issue during transitions between care settings. Concern was due to two studies being from outside the UK, and some being more than ten years old – but more recent studies report similar findings, suggesting that this issue has not significantly changed over time and probably generalises well.

Adequacy was rated as no or very minor concerns. The sub-finding is fairly general and descriptive, and was sufficiently supported by the number of studies and the richness of data.

The overall assessment of confidence was moderate, having been downgraded by one increment due to methodological limitations.

Review sub-finding 3b: Barriers to communication

Healthcare professionals identified the following barriers to cross-boundary communication: fragmented and incompatible information systems, and a lack of time. HF nurses describe in one study, for example, how they struggle to speak to GPs when they have concerns about their patients (GPs are in surgery or on house-calls during most of office hours).

This sub-finding was based on 2 studies²⁰²^,²⁰⁹.

Methodological limitations were rated moderate overall. One paper was rated serious limitations due to lack of depth and explanations. One paper was rated as moderate limitations due to little description of participants. More data came from the latter, therefore rated as moderate overall.

Coherence was rated no or very minor concerns overall. Agreement between papers was good. Agreement with related findings in other studies was also good.

Relevance was rated no or very minor concerns overall. The relevance of the sub-finding to the subject of the review was high. Both studies were from different areas of the UK.

Adequacy was rated as serious concerns. There is some concern about both the lack of quantity of data for a finding of generality, and a lack of depth for an explanatory finding.

The overall assessment of confidence was low, having been downgraded by two increments due to concerns over methodology and adequacy.

Review sub-finding 3c: Information after discharge from HF services

Patients reported being well informed while they were in the HF clinic, but received no information after discharge to primary care. Patients interpreted this change in an ambivalent way, as if not confident in the continuing significance of their CHF and whether they needed ongoing care for their CHF.

This sub-finding was based on 1 study¹⁴.

Methodological limitations were rated serious overall. The included paper was rated serious due to small number of participants (4), poor explanation of methods and low richness of data, in which the researchers’ voice was quite prominent.

Coherence was rated no or very minor concerns overall. This specific finding could not be properly assessed, as there was no related data for comparison.

Relevance was rated minor concerns overall. The relevance of the sub-finding to the focus of the review was very high. Concern was due to the study being from outside the UK, and the interviews being in Swedish. It is not clear to what extent the same problems occur in the NHS, but given the similarly pattern of HF clinic discharge, there is potential for the same issue to arise.

Adequacy was rated as serious concerns; it was limited by the number of participants and lack of depth.

The overall assessment of confidence was very low. The overlap between the rating of methodology and adequacy rating was taken into account when downgrading, but it was still felt necessary to downgrade by three increments due to the serious concerns raised.

4. Contact / Access

Review sub-finding 4a: Access to Routine Care

People with heart failure who were not receiving specialist treatment reported that it was challenging to access specialist advice or secondary care, and that they have to spend time trying to access these through their general practitioner. Once receiving specialist care, or after discharge, patients reported that continued access to the support they felt they needed was also not straightforward. Some patients felt that the illness itself made asking for help harder. One patient reports “If you have a heart issue you have to shout and speak for yourself and keep at somebody… And if you’re not feeling well that’s not what you want to be doing”.²⁰⁷ p 108 This is contrasted with patients who have cancer, who found their condition was a ‘door-opener’, so they did not have to assert their needs in the same way.

The inconsistent way care is offered is to people with CHF was also demonstrated by the impression from HF nurses that GPs have differing thresholds to refer back for specialist help, such that some GPs can leave people struggling in the community.

This sub-finding was based on 3 studies¹²⁶^#2749^,²⁰⁷.

Methodological limitations were rated serious overall. Two papers were rated serious limitations as they were secondary analysis with a lack of information about how the data had originally been collected. One paper was rated as moderate limitations due to lack of description of participants.

Coherence was rated no or very minor concerns overall. Agreement between papers was good. Agreement with general, related findings in other studies was also good.

Relevance was rated minor concerns overall. The relevance of the sub-finding to the focus of the review was high, as movements in and out of HF specialist care are essentially about accessing routine care. Concern was due to two studies being from prior to 2010.

Adequacy was rated as minor concerns. This sub-finding was fairly general, and so the number of studies was a little lower than would be preferable to support it, but there were no concerns about the richness of data.

The overall assessment of confidence was low, having been downgraded by two increments due mainly to methodological concerns plus the minor concerns regarding relevance and adequacy.

Review sub-finding 4b: A Primary Contact Person

People with heart failure felt they would like one professional nominated as their primary contact person. The preferred characteristics of the relationship with their primary contact person were “a long-term relationship, characterized by openness, trust, and appreciation” p1400²⁹. Most would choose their GP as primary contact person, but some felt their cardiologist was more suitable. A key role of the primary contact would be to enable easy access for patients to care, for example timely appointments. This finding is related to finding 2b, where professionals suggested a single person coordinating care might improve co-operation.

This sub-finding was based on 1 study²⁹.

Methodological limitations were rated moderate overall as the included paper was limited in detail and reflection.

Coherence was rated no or very minor concerns overall as it compliments some of the other related review findings.

Relevance was rated serious concerns overall as the study was not from the UK, was originally in German, and had a restrictive definition of heart failure (EF<40%).

Adequacy was rated as moderate concerns; although the sub-finding was only reported in one paper, it is well explored in that paper.

The overall assessment of confidence was low, having been downgraded by 2 increments due to relevance concerns, and also for methodology and adequacy.

Review sub-finding 4c: Access to Urgent Care

Patients found that the HF clinic provided them easy access to physicians and nurses, and healthcare professionals noted that HF nurses were able to react more quickly to sudden changes in health status than cardiologists or GPs. This leads to a discrepancy in access between those who were active on the HF clinic caseload, and those discharged to primary care. It was clear that although access to specialists was valued, consistency of care was also needed, and the two were sometimes in tension. For example in one study, a specialist nurse explained: “…patients can ring us up at any time and we can see them on that day if we need to. There is no waiting around and ringing your GP”, but GPs in that study felt patients needed a more holistic approach than the HF clinic offered because “… all they care about is the heart failure.” pp300, 303²⁹⁴

People with advanced heart failure and their carers reported uncertainty about where to seek help. One patient reported: “I rang my GP who said to ring the hospital, rang hospital and was told they couldn’t do anything, you have to ring GP!” p120⁸

This sub-finding was based on 3 studies²²⁵^,²⁹⁴^#2771.

Methodological limitations were rated moderate overall. One paper was rated serious limitations due to lack of discussion and reflection. One paper was rated as moderate limitations due to lack of discussion and reflection on methods. One paper was rated as minor limitations.

Coherence was rated no or very minor concerns overall. Agreement between papers was good, although they revealed another instance of tension between the objectives of ease of access to specialists and continuity of care.

Relevance was rated moderate concerns overall. The relevance of the sub-finding to the focus of the review was high. One patient interview study was not in the UK, and was originally in Swedish. The UK studies are both from before 2010, and the finding is likely to be sensitive to changes in service design.

Adequacy was rated as moderate concerns. This sub-finding was adequately supported by a number of studies, but was lacking in depth.

The overall assessment of confidence was low, having been downgraded by two increments due to combined concerns regarding methodology, relevance and adequacy.

5. The role of specialist HF services

Review sub-finding 5a: Expectations

The expectations of specialist HF services from people with CHF and primary care providers were seen to differ from what the service is delivering. What people providing HF specialist services think patients and primary care want is different again. For example:

People with heart failure value the time the HF nurse has to spend with them, and the psychosocial support they provide; GPs however said they wanted HF nurses to be a resource for primary care, for advice and a point of contact with cardiology; but HF nurses felt a pressure from GPs to fill a gap in service provision. As one HF nurse states: “GPs will try to pass the buck because heart failure patients are a problem because they’re not stabilised… So they get a patient they even suspect has got heart failure, they’ll move heaven and earth to get them through your clinic because they know you’ll sort them out, give them everything they need, sort all their ‘echoes’, all their blood tests, all go back as a lovely little package.” p301²⁹⁴
Healthcare professionals involved in HF services perceived GPs as wanting patients to be taken on (or not discharged) by the HF clinic. In one of the studies, those same GPs were also heard, and on the contrary expressed willingness to manage the CHF for most of their CHF patients, suggesting that the HF clinic could concentrate on the most complex cases. One GP stated: “Now that we have open access echo available [to primary care), I am much more comfortable about being the person who makes the decision, the diagnosis, and who initiates and monitors success.” p298²⁹⁴

This sub-finding was based on 3 studies⁴⁷^,²⁰²^,²⁹⁴.

Methodological limitations were rated serious overall. Two papers were rated serious limitations due to having limited discussion and reflection, and in one case a lack of richness of data in our area of interest. One paper was rated as moderate limitations as there is little description of the participants.

Coherence was rated minor concerns overall. This was one of the least well-defined sub-findings in the review – what was clear is that different groups had certain expectations on services, but the exact nature of the expectations differed between the studies, and few studies were able to look at the difference in expectations between different groups. The finding of conflict between interested groups due to differing expectations was clear in the two service evaluations included here (which, of the study types in the review, were the studies best able to look at this issue).

Relevance was rated moderate concerns overall. The relevance of this sub-finding to the focus of the review was somewhat indirect. The issues identified may seem quite specific to the exact model of delivering HF care, but all studies were in the UK, at different time periods, so there may be a generalisable point.

Adequacy was rated as serious concerns. The sub-finding was a complex issue, not explored in depth.

The overall assessment of confidence was very low, having been downgraded by three increments for methodology, relevance and adequacy concerns.

Review sub-finding 5b: Focus

Healthcare professionals felt that HF services should be focussed on managing CHF rather than its comorbidities. HCP from HF services felt that they did not have the ability to assess and treat some comorbidities (an example condition was iron-deficiency anaemia). They also felt that addressing people’s “primary care issues” could:

take patients out of contact with primary care, potentially causing a loss of continuity; and
leave HF services without capacity to manage HF care across their caseload.

This sub-finding was based on 2 studies¹²⁷^,³²⁵.

Methodological limitations were rated moderate overall. One paper was rated serious limitations due to lack of information on methods and data analysis, and not all findings being supported by data. One paper was rated as minor limitations.

Coherence was rated no or very minor concerns overall. Agreement between papers was good. The sub-finding was compatible with other sub-findings in the review.

Relevance was rated moderate concerns overall. The relevance of this sub-finding to the focus of the review was somewhat indirect. There was also some concern due to both studies being from outside the UK.

Adequacy was rated as moderate concerns. The sub-finding was specific and supported by a number of studies, but is more than just descriptive, so would benefit from a greater richness of data.

The overall assessment of confidence was low, having been downgraded by two increments due to adequacy foremost, and also methodology and relevance.

Review sub-finding 5c: Decision to keep patients on the caseload

One study looked at care networks of patients with advanced heart failure. It found that specialists describe their decision to take over someone’s care in the HF clinic as being based on patient need (primarily from the CHF perspective) and the perceived ability of the general practitioner to fulfil these needs. The paper shows though, that actual practice varied widely, and the research team reflect that: “complexity [of the patient] interacted with the health care providers’ perspective, comfort level and available resources… to influence referral and consultation practice”p372³²⁵ Another study also reflected this, where some HF professionals acknowledged the GPs’ importance in ensuring continuity of care, but took over care despite of this because they doubted the ability of the GPs to manage heart failure.

This sub-finding was based on 2 studies²⁰²^,³²⁵.

Methodological limitations were rated minor overall. One paper was rated as moderate limitations, and one as minor limitations – with the latter contributing more to this sub-finding.

Coherence was rated no or very minor concerns overall. It fits with other sub-findings, in that the decision about where CHF patients are best managed has a number of objectives, which can conflict.

Relevance was rated moderate concerns overall. The relevance of this sub-finding to the focus of the review was high from a staff experience point of view. Concern was due to one study being from outside the UK, however, the second (UK-based) study was conducted in a number of locations, which supports this being a relevant and generalisable theme.

Adequacy was rated as moderate concerns. The finding was specific, but there were still an inadequate numbers of studies. This sub-finding was both descriptive and explanatory, and there was plenty of detail in the papers. One specific concern was that part of the aim of the main study in this analysis was a partly theoretical piece, designed to look at the complexity that exists in healthcare; so the finding of complexity was partially by design.

The overall assessment of confidence was low, having been downgraded by two increments due primarily to concerns over relevance and secondly methodology and adequacy.

Figure 7

Main review findings and connections between them.

9.3.3.5. Qualitative evidence summary

Table 132

Summary of Evidence for Continuity.

Table 133

Summary of Co-ordination.

Table 134

Summary of Evidence for Communication.

Table 135

Summary of Evidence for Contact / Access.

Table 136

Summary of Evidence for Role of Specialist HF Services.

9.3.4. Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix D.

9.3.5. Evidence statements

Qualitative

See the narrative summaries of review findings in section 1.3.3.1.

Economic

No relevant economic evaluations were identified.

9.3.6. Recommendations and link to evidence

Table

ensure effective communication links between different care settings and clinical services involved in the person’s care lead a full review of the person’s heart failure care, which may form part of a long-term conditions review

9.4. Information and support needs regarding diagnosis and prognosis

9.4.1. Introduction

Communication is the one of the key determinants of successful management of chronic diseases. The diagnosis of heart failure is often a shock to patients. Heart failure is a complex disorder which requires active participation of patients in its management. A large literature has developed about the perceptions and misperceptions that the diagnosis entails and about the prognosis of heart failure. Different features of the disease process are highlighted by patients and the various professional groups involved in the management of heart failure. This review aimed to identify the key factors that need to be communicated about the diagnosis of heart failure to patients and their carers, the information needed about the disease and the support required to allow care to be optimised.

9.4.2. Review question: What are the information and support needs to be considered when communicating a diagnosis and consequent prognosis, to people with heart failure, their families and carers?

For full details see review protocol in appendix A.

Table 137

Characteristics of review question.

9.4.3. Qualitative evidence

9.4.3.1. Methods

This evidence review was developed using the methods and process described in Developing NICE guidelines: the manual.²³⁶ Methods specific to this review question are described in the review protocol in appendix A.

Declarations of interest were recorded according to NICE’s 2014 conflicts of interest policy.

9.4.3.2. Summary of included studies

A search was conducted for qualitative studies exploring the communication and support needs of people with heart failure at the time when diagnosis and prognosis are relayed to them by health care professionals. Twelve qualitative studies (15 papers) were included in the review;⁸^,²⁶^,⁵³^,¹⁰²^,¹¹⁷^,¹²³^,¹³³^,¹⁴¹^,¹⁴⁸^,²⁰⁷^,²³¹^,³⁰²^,³¹³^,³²⁸^,³⁷⁰ these are summarised in Table 138 below. Key findings from these studies are summarised in Section 9.4.3 below. See also the study selection flow chart in Appendix C, study evidence tables in appendix F, and excluded studies lists in appendix I.

Table 138

Summary of studies included in the review.

9.4.3.3. Qualitative evidence synthesis

Table 139

Review findings.

9.4.3.3.1. Narrative summary of review findings

The review findings are grouped within four overarching findings, as follows:

Diagnosis of heart failure
Understanding heart failure
Discussion of prognosis
Improving communication/information flow

Each of these four findings contain several key subfindings that are described below along with an evaluation of the quality including an explanation for the CerQual grading that appears with the subfindings in the summary tables in section 1.3.4.

1. Diagnosis of heart failure

Review subfinding 1a: Communication is challenging

Clinicians find the first conversation disclosing the diagnosis of heart failure to the patient to be difficult. Some clinicians described the diagnosis of heart failure itself as challenging, making it even more difficult to relay information to patients.

This subfinding was based on two studies.²⁶^,¹³³^,³¹³

Methodological limitations were rated moderate concerns overall. Both studies were rated as moderate limitations due to lack of context and the role of the researcher, plus data analysis in one study and data richness in the other.

Coherence was rated no or very minor concerns overall. Agreement between papers was good and appeared to fit well with related findings in other studies.

Relevance was rated no or very minor concerns overall. The relevance of the subfinding to the focus of the review was high. The data stemmed from two papers from the UK in a relevant population.

Adequacy was rated as minor concerns. The subfinding was general, and although explicitly supported by only a limited number of studies, other studies found similar themes. The subfinding was descriptive but the richness of data was somewhat lacking.

The overall assessment of confidence was moderate, having been downgraded by one increment due to methodological and adequacy limitations.

Review subfinding 1b: Timing and setting

The reluctance by some GPs to diagnose patients meant that some patients received a shock diagnosis when admitted to secondary care. Receiving the diagnosis during an unplanned hospital admission was deemed “unhelpful and inadequate by clinicians, patients and their carers” (Simmonds 2015) as diagnosis could not be relayed in a sensitive manner to the patient and family who had no time to assimilate the information in this busy environment. Patients with good access to hospital- and community-based heart failure specialist nursing teams reported more positive experiences of communication regarding diagnosis.

This subfinding was based on three studies.²⁶^,¹¹⁷^,³¹³

Methodological limitations were rated moderate concerns overall. Two studies were rated as moderate limitations due to lack of context and role of the researcher, plus data analysis in one study and data richness in the other. One study had serious limitations due to lack of context, role of the researcher, data collection and richness of data.

Coherence was rated as minor concerns overall. Agreement between papers was good and appeared to fit well with related findings in other studies. The only findings that tended to be the other way are regarding the positive experience of patients with good access to specialist nursing teams.

Relevance was rated no or very minor concerns overall. The relevance of the subfinding to the focus of the review was high. The data stemmed from three studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. The subfinding was general, and although explicitly supported by a limited number of studies, other studies found similar themes. The finding was descriptive and the richness of data was sufficient for this.

The overall assessment of confidence was moderate, having been downgraded by one increment due to the combination of concerns in methodology and coherence.

Review subfinding 1c: Gradual process

Diagnosis was seen by healthcare professionals as a gradual process which in turn makes the communication of the diagnosis a gradual process. They emphasised the need to provide information and guidance as part of an ongoing conversation with the patient and family. Heart failure specialist nurses and GPs were seen as key to the success of this process.

This subfinding was based on three studies.²⁶^,¹³³^,²⁰⁷^,³¹³

Methodological limitations were rated moderate concerns overall. Two studies were rated as moderate limitations due to a lack of context and role of the researcher, plus data analysis in one study and data richness in the other. One study had serious limitations due to the secondary use of data and the subsequent inability to assess their methodology in detail, along with vague description of the analysis.

Coherence was rated no or very minor concerns overall. Agreement between papers was good.

Relevance was rated no or very minor concerns overall. The relevance of the subfinding to the focus of the review was very high. The data stemmed from three studies from the UK in a relevant population.

The overall assessment of confidence was moderate, having been downgraded by one increment due to the concerns in methodology.

Review subfinding 1d: Terminology

Clinicians stated that they felt that the terminology affected communication. They wanted patients to understand their condition but also wished to avoid upsetting them and extinguishing their hope. Only very few healthcare professionals described using the term ‘heart failure’; it was regarded by most of them as an anxiety-invoking term (similar to a cancer diagnosis). So instead many resorted to using euphemisms (e.g. ‘ageing heart’, ‘stiff heart’, ‘heart not pumping efficiently’) or paradoxically even more complex terminology (e.g. ‘left ventricular failure’).

These alternative explanations often led to poorer communication, confusion for people with heart failure and lack of interest in their diagnosis as a consequence.

People with heart failure were often unaware of the term ‘heart failure’. Many described that the term was not mentioned initially but introduced later on by specialists, often leading to shock and confusion for the patients.

This subfinding was based on six studies.⁸^,²⁶^,¹³³^,²⁰⁷^,²³¹^,³¹³^,³²⁸

Methodological limitations were rated moderate overall. Five studies were rated moderate limitations: two studies due to lack of context and role of the researcher, plus data analysis in one study and data richness in the other. Another two studies due to lack in data richness, plus missing information on the topic guide in one, and lack of reflection on the role of the researcher and loose link between findings and conclusions in the other. The fifth study due to no reflections on role of researcher and lack of detail on methodology. One study had serious limitations due to the secondary use of data and the subsequent inability to assess their methodology in detail, along with vague description of the analysis.

Coherence was rated minor concerns overall. Agreement between papers was very good except one where healthcare professionals actually used the terminology ‘heart failure’ with all their patients.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from three studies from the UK in a relevant population.

Adequacy was rated as no concerns. The subfinding was explicitly supported by quite a few studies in much detail.

The overall assessment of confidence was moderate, having been downgraded by one increment due to the concerns in methodology and coherence.

2. Understanding heart failure

Review subfinding 2a: Understanding of diagnosis

There is variability in the complexity and depth of patients’ understanding of their diagnosis.

Most patients showed poor knowledge and misunderstanding of their condition and its implications, including treatments, their importance, side effects and limitations. Many did not connect symptoms they had to their heart failure; some even deduced their diagnosis from the medications they were taking. The lack of understanding was often compounded by confusion and short-term memory loss that is associated with heart failure.

However, other patients showed a high level of interest in their illness, were proactive in seeking information (e.g. online), knew how to manage their condition, and were well informed and equipped for informed exchanges with healthcare professionals about heart failure. In one study the authors noted, however, that this was an unusual group of younger patients and/or those with a background in health.

Healthcare professionals were sympathetic to patients’ uncertainty about the meaning of their diagnosis and about treatments, as they were aware that a lack of time for communication contributed to poor understanding.

This subfinding was based on nine studies.⁸^,²⁶^,⁵³^,¹¹⁷^,¹³³^,¹⁴¹^,²⁰⁷^,²³¹^,³²⁸

Methodological limitations were rated serious overall. Five studies were rated moderate limitations due to a mix of reasons including a lack of the role of researcher, data richness and details of methodology. Three studies were rated serious limitations; one study due to the secondary use of data and the subsequent inability to assess their methodology in detail, along with vague description of the analysis. Two studies due to the lack of richness of data, context and the role of the researcher, as well as a lack of information on data collection in one study and the reasoning for the choice of methods in the other. One study had very serious limitations due to limited information on the background and reflection of the researcher, data collection, data richness, relevance of findings and link to conclusions.

Coherence was rated minor concerns overall. Agreement between papers was very good with the majority demonstrating patients’ lack of understanding of their diagnosis. Only very few papers also described that some patients had a very good understanding of their condition. As a certain degree of variability in people’s understanding is expected and all papers were in agreement that the majority lacked the understanding, the concerns were rated as minor overall.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from nine studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. The subfinding was general and supported by numbers of studies; it was descriptive and supported by the richness of data.

The overall assessment of confidence was low, having been downgraded by two increments due to concerns over methodology and coherence.

Review subfinding 2b: Knowledge and management

Some patients did not want to know more about their diagnosis as it would cause them to worry, and they chose to put all their trust into healthcare professionals to make decisions for them. Yet, others described having received too little information about their diagnosis and some received contradicting information that caused them confusion. These patients felt that the lack of knowledge caused them panic attacks and anxiety about the practicalities of what to do in a crisis, for example.

This subfinding was based on five studies.²⁶^,¹⁰²^,¹¹⁷^,¹⁴⁸^,³²⁸

Methodological limitations were rated serious overall. Two studies were rated moderate limitations: one due to limitations in context, the role of researcher and data analysis, and another study due to missing information on the topic guide and the lack of data richness in sections relevant to this review. Three studies were rated serious limitations due to a lack of information on the role of the researcher, data richness as well as lack of rigour in research methods.

Coherence was rated minor concerns overall. Agreement between papers was good in that they all reported a range of people’s attitudes. Variability in people’s attitudes is expected, hence the concerns were rated as minor overall.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from five studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. The subfinding was general and supported by numbers of studies; it was descriptive and supported by the richness of data.

The overall assessment of confidence was low, having been downgraded by two increments due to concerns over methodology and coherence.

3. Discussion of prognosis

Review subfinding 3a: Difficult conversations

Staff noted that discussions of prognosis are difficult, given the difficulty of diagnosis in the first place and due to the unpredictable disease trajectory of heart failure, and they were reluctant to have them. Some healthcare professionals also found it challenging to balance the need to be honest with the patient about their condition (which could raise anxiety) with building trust and maintaining hope and a positive outlook when faced with a life-threatening illness. Some healthcare professionals considered that patients may not want to know everything about their prognosis, perhaps hinting at a degree of paternalism or recognition of denial as a way of coping.

Some felt it was more appropriate to address prognosis over time, given these uncertainties and in response to changing circumstances, particularly when a patient might be approaching the end of their life.

A common professional perception was that these types of exchange between clinician and patient did not happen often enough. In one study cardiac staff confirmed that issues such as future care in the event of an exacerbation or end-of-life preferences are rarely raised with patients. It was suggested that lessons could be learnt from communication in cancer where clear information about prognosis is provided to patients. However, the staff reported having difficulties handling patient denial, discussing poor prognosis and dealing with the emotional involvement of patients and their families. It was said in this study that cardiac staff often lack the communication skills required to handle these sensitive issues.

This subfinding was based on six studies.²⁶^,⁵³^,¹³³^,¹⁴¹^,²⁰⁷^,²³¹^,³⁰²

Methodological limitations were rated serious overall. Three studies were rated moderate limitations due to the lack of the role of the researcher, plus another two out of the following limitations: lack of context, data richness, data analysis or loose linkage between findings and conclusions. Two studies were rated serious limitations, one due to the secondary use of data and the subsequent inability to assess methodology in detail, along with vague descriptions of this analysis; whilst the second study due to a lack of reflection on the researcher role in the study, limited context, reasoning for the choice of methods and richness of data. One study was rated as having very serious limitations due to the limited information on the background and reflection of the researcher, data collection, data richness, relevance of findings and a link to conclusions.

Coherence was rated no or very minor concerns overall. Agreement between papers was very good.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from six studies from the UK in a relevant population.

Adequacy was rated as no concerns. The subfinding was explicitly supported by quite a few studies in much detail.

The overall assessment of confidence was moderate, having been downgraded by one increment due to methodological limitations.

Review subfinding 3b: Understanding of prognosis

Patients often reported that prognosis was rarely discussed. Some patients were aware of the seriousness of their condition but reported a lack of understanding of the prognosis. Some patients were unaware that heart failure is a terminal condition and reported feeling very frightened when informed at the end-stage.

This subfinding was based on five studies.⁸^,²⁶^,⁵³^,²⁰⁷^,²³¹

Methodological limitations were rated moderate overall. Three studies were rated moderate limitations due to limitations on the role of the researcher and methodology, as well as limitations in context for one study and data richness and lack of link between findings and conclusions for another study. Two studies had serious limitations: one study due to the secondary use of data and the subsequent inability to assess methodology in detail, along with vague descriptions of this analysis; and the second study due to the lack of reflection on the researcher role in the study, limited context, reasoning for choice of methods and richness of data.

Coherence was rated minor concerns overall. Agreement between papers was very good with the majority demonstrating patients’ lack of understanding of their prognosis. It was only slightly downgraded to minor concerns overall due to some variability in patients’ understanding.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from five studies from the UK in a relevant population.

Adequacy was rated as no concerns. The subfinding was explicitly supported by quite a few studies.

The overall assessment of confidence was moderate, having been downgraded by one increment due to moderate methodological and minor coherence limitations.

Review subfinding 3c: Patients’ concerns

Some patients preferred not to know more about their prognosis, perhaps using denial as a way of coping. Yet most others, including carers, felt they were inadequately informed by healthcare professionals, and felt uncertainty about what would happen as the disease progressed. “Thinking about the future was a common preoccupation, with patients expressing very realistic concerns that their life expectancy was limited.” (Aldred 2004)

This subfinding was based on seven studies.⁸^,²⁶^,⁵³^,¹¹⁷^,¹³³^,¹⁴⁸^,³⁷⁰

Methodological limitations were rated moderate overall. Four studies were rated moderate limitations: each due to a number of limitations that included some of the following; limitations in context, role of researcher, data analysis, research methods rigour, and/or data richness in some sections relevant to our review. Three studies were rated serious limitations due to a lack of reflection on the role of the researcher and data richness; plus limited context and reasoning for the choice of methods for one study, lack of information on context and data collection for another, and lack of rigour of research methods, study aims, and relevance of findings for the third study.

Coherence was rated minor concerns overall. Agreement between papers was good in that they all reported a range of people’s attitudes and concerns. Variability in people’s attitudes is expected, hence the concerns were rated as minor overall.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from seven studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. The subfinding was general and supported by a number of studies. The finding was descriptive and supported by the richness of data.

The overall assessment of confidence was moderate, having been downgraded by one increment due to moderate methodological and minor coherence limitations.

4. Improving communication/information flow

Review subfinding 4a: Education and joint working

GPs expressed a need for education around the identification and diagnosis of heart failure patients; and that with such an improvement in their own understanding their ability to communicate this information to the patient would improve. They felt that “changes needed to be made within the health profession first” before information on the identification and diagnosis of heart failure can be more clearly communicated to patients (Barnes 2006).

Further, in one study cardiology and palliative care staff reported a lack of clarity regarding what had previously been discussed with and disclosed to patients. They recommended “mutual education and joint working” between the specialties to improve communication to the patient and family (Harding 2008).

This subfinding was based on two studies.²⁶^,¹⁴¹

Methodological limitations were rated serious overall. One study had moderate limitations due to limitations in context, role of researcher and data analysis. Another study had very serious limitations due to the limited information on the background and reflection of the researcher, data collection, data richness, relevance of findings and link to conclusions.

Coherence was rated no or very minor concerns overall. Agreement between papers was good.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from two studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. Although the subfinding was explicitly supported by only two studies, it was provided in much detail.

The overall assessment of confidence was moderate, having been downgraded by one increment due to serious methodological limitations.

Review subfinding 4b: Ongoing relationship

Healthcare professionals reported that because appointments with consultants are too short to relay all the information a patient would need regarding diagnosis and prognosis, patient education was often delegated to specialist nurses in the outpatient or community setting. Healthcare professionals suggested that education was best delivered within the ongoing relationship between the specialist nurse and patient, in particular during home visits where patients are more relaxed and able to assimilate information.

In support of this some patients reported that they find it easier to communicate with nurses than to cardiologists and it was suggested that the specialist nurse environment was a good place to discuss patients’ condition and give information.

This subfinding was based on two studies.²⁶^,¹³³^,³¹³

Methodological limitations were rated moderate overall. One study had moderate limitations due to limitations in context, the role of researcher and data analysis. Another study was rated as having moderate limitations due to a lack of context, reflections on the role of the researcher, and data richness in some sections relevant to this review.

Coherence was rated no or very minor concerns. Agreement between papers was very good.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from two studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. Although the subfinding was explicitly supported by only two studies, it was illustrated in much detail.

The overall assessment of confidence was high, as the moderate limitations in methodology were offset by the richness of data provided by these two studies.

Review subfinding 4c: Tailoring of information

Healthcare professionals suggested that information needed to be tailored to the need of the individual as some people are more able and/or willing to hear information than others. Also, people with heart failure tend to be older, more likely to accept what a doctor says and not proactive in asking questions (e.g. some patients may be unwilling to raise questions about prognosis). They may also be more likely to have short-term memory loss, are too ill to benefit from education or are in denial about their condition etc. Consequently, specialist nurses spoke of the necessity to find a balance between the education they offered patients and their capacity to receive it. As a consequence they tried to identify key issues and personalise the information accordingly. This would involve repeating these messages over time.

This subfinding was based on seven studies.²⁶^,⁵³^,¹¹⁷^,¹³³^,¹⁴¹^,²³¹^,³²⁸

Methodological limitations were rated moderate overall. Four studies were rated moderate limitations: each due to a number of limitations that included some of the following; limitations in context, role of researcher, data analysis, research method rigour, and/or data richness in some sections relevant to this review. Two studies were rated serious limitations due to lack of reflection on researcher role in the study, limited context and data richness, plus lack of reasoning for choice of methods in one study and limitations in data collection for the other. One study was rated very serious limitations due to the limited information on the background and reflection of the researcher, data collection, data richness, relevance of findings and link to conclusions.

Coherence was rated no or very minor concerns. Agreement between papers was very good.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from seven studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. The subfinding was explicitly supported by seven studies, and it was illustrated in much detail.

The overall assessment of confidence was high, as the moderate limitations in methodology were offset by the richness of data provided by these seven studies.

Review subfinding 4d: Improving access to information

Patients who wished for more information sought a “better understanding of the disease process, the practical limitations, how to get help and how to cope with living with heart failure” Horne 2004). Some people with heart failure wished to be told openly and sensitively about their (poor) prognosis by clinicians, with some wanting to know more accurately when they would die. As some patients reported not having received written information of their heart failure diagnosis, ensuring that information pamphlets are passed on during consultations is vital so that patients can go through them at home in their own time.

Patients expressed the desire for information regarding both diagnosis and prognosis to be communicated using lay terms.

Family and informal carers wished to be involved in all communications in order to support the patient in their role of family information providers. Carers also asked for more information about what to do in an emergency, how to recognise when signs and symptoms needed urgent attention and how to perform cardiopulmonary resuscitation.

The provision of information on access to a telephone advice line or support group was also suggested.

This subfinding was based on six studies.⁸^,²⁶^,¹⁴¹^,¹⁴⁸^,²³¹^,³⁷⁰

Methodological limitations were rated moderate overall. Four studies were rated as moderate limitations: each due to a number of limitations that included some of the following; limitations in context, role of researcher, data analysis, research methods rigour, and/or data richness in some sections relevant to this review. One study was rated as serious limitations due to study aims, role of the researcher, rigour of research methods, data richness and relevance of findings. One study was rated very serious limitations due to the limited information on the background and reflection of the researcher, data collection, data richness, relevance of findings and link to conclusions.

Coherence was rated minor concerns overall. Although the suggestions for how to improve access to information varied, overall the agreement between papers was good.

Relevance was rated no or very minor concerns. The relevance of the subfinding to the focus of the review was very high. The data stemmed from six studies from the UK in a relevant population.

Adequacy was rated as no or very minor concerns. The findings were supported by a few studies in much detail.

The overall assessment of confidence was moderate, having been downgraded by one increment due to moderate methodological and minor coherence limitations.

9.4.3.4. Qualitative evidence summary

Table 140

Summary of evidence.

9.4.4. Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix D.

9.4.5. Evidence statements

Qualitative

See the narrative summaries of review findings in section 9.4.3.3.1.

Economic

No relevant economic evaluations were identified.

9.4.6. Recommendations and link to evidence

Table

discuss the person’s diagnosis and prognosis explain heart failure terminology

9.5. Recommendations

9.5.1. Multi-disciplinary teams

71.

The core specialist heart failure multidisciplinary team (MDT) should work in collaboration with the primary care team, and should include:

a lead physician with a subspecialty interest in heart failure (usually a consultant cardiologist) who is responsible for making the clinical diagnosis
a specialist heart failure nurse
a healthcare professional with expertise in specialist prescribing for heart failure. [2018]

72.: The specialist heart failure MDT should directly involve, or refer people to, other services, including rehabilitation services, and tertiary and palliative care, as needed. [2018]

73.

The specialist heart failure MDT should:

diagnose heart failure
give information to people newly diagnosed with heart failure (see section 9.4.6)
manage newly diagnosed, recently decompensated or advanced heart failure (NYHA [New York Heart Association] class III to IV) heart failure
optimise treatment
start new medicines that need specialist supervision
continue to manage care after an interventional procedure such as implantation of a cardioverter defibrillator or cardiac resynchronisation device
manage heart failure that is not responding to treatment. [2018]

9.5.2. Transition between heart failure care settings

74.

The primary care team should carry out the following for people with heart failure at all times, including periods when the person is also receiving specialist heart failure from the MDT:

ensure effective communication links between different care settings and clinical services involved in the person’s care
lead a full review of the person’s heart failure care, which may form part of a long-term conditions review
recall the person at least every 6 months and update the summary and clinical record
ensure that changes to the clinical record are understood and agreed by the person with heart failure and shared with the specialist heart failure MDT
arrange access to specialist heart failure services if needed. [2018]

9.5.2.1. Care after an acute event

For recommendations on the diagnosis and management of acute heart failure see NICE’s guideline on acute heart failure.

75.: People with heart failure should generally be discharged from hospital only when their clinical condition is stable and the management plan is optimised. Timing of discharge should take into account the wishes of the person and their family or carer, and the level of care and support that can be provided in the community. [2003]

76.: The primary care team working within the specialist heart failure MDT should take over routine management of heart failure as soon as it has been stabilised and its management optimised. [2018]

9.5.2.2. Writing a care plan

77.

The specialist heart failure MDT should write a summary for each person with heart failure that includes:

diagnosis and aetiology
medicines prescribed, monitoring of medicines, when medicines should be reviewed and any support the person needs to take the medicines
functional abilities and any social care needs
social circumstances, including carers’ needs. [2018]

78.

The summary should form the basis of a care plan for each person, which should include.

plans for managing the person’s heart failure, including follow-up care, rehabilitation and access to social care
symptoms to look out for in case of deterioration
a process for any subsequent access to the specialist heart failure MDT if needed
contact details for:
- a named healthcare coordinator (usually a specialist heart failure nurse)
- local heart failure specialist care providers, for urgent care or review
additional sources of information for people with heart failure. [2018]

79.: Give a copy of the care plan to the person with heart failure, ther family or carer if appropriate, and all health and social care professionals involved in their care. [2018]

9.5.3. Information and support needs regarding diagnosis and prognosis

80.: When giving information to people with heart failure, follow the recommendations in the NICE guideline on patient experience in adult NHS services. [2018]

81.: Discuss the person’s prognosis in a sensitive, open and honest manner. Be frank about the uncertainty in predicting the course of their heart failure. Revisit this discussion as the person’s condition evolves. [2018]

82.: Provide information whenever needed throughout the person’s care. [2018]

83.: Consider training in advanced communication skills for all healthcare professionals working with people who have heart failure. [2018]

9.5.3.1. First consultations for people newly diagnosed with heart failure

84.

The specialist heart failure MDT should offer people newly diagnosed with heart failure an extended first consultation, followed by a second consultation, to take place within 2 weeks if possible. At each consultation:

discuss the person’s diagnosis and prognosis
explain heart failure terminology
discuss treatments
address the risk of sudden death, including any misconceptions about that risk
encourage the person and their family or carers to ask any questions they have. [2018]

Publication Details

Copyright

Publisher

National Institute for Health and Care Excellence (NICE), London

NLM Citation

National Guideline Centre (UK). Chronic Heart Failure in Adults: Diagnosis and Management. London: National Institute for Health and Care Excellence (NICE); 2018 Sep. (NICE Guideline, No. 106.) 9, Referral and approach to care.

Table 97PICO characteristics of review question

Population	People with heart failure in a community or outpatient setting that is applicable to UK practice Stratification: Risk status at time of randomisation (high versus lower risk patients) High risk includes: new diagnosis recent decompensation (defined as hospital admission due to HF) severe and/or unresponsive disease (defined as NYHA III/IV) and requiring medicine titration, device implantation or other surgical intervention
Intervention / Comparators:	Studies comparing MDTs with ‘usual care’ (regardless of how defined). For a study to be included, there must be a clear description of collaborative working between more than one healthcare profession or discipline. A study involving an intervention delivered by one specified health profession may be included where there is a clear description of multidisciplinary collaboration. The intervention must have included the delivery (on average) of at least two face-to-face meetings. Stratification: Length of intervention delivered in the study ≤ three months (short) > three months, ≤ six months (mid) > six months (long) Where study length varied due to recruitment dates, average length was used. Where study length varied due to the needs of the patient, the shortest duration of protocol was used.
Outcomes	CRITICAL - All-cause mortality (Time to event) - Quality of life (Continuous) - Unplanned hospitalisation (Count rate) IMPORTANT - Medicine optimization and adherence - Dying in preferred place of death (for palliative care patients) - Adverse events – hypotension, hyperkalaemia, and renal function - Patient and carer experience
Study design	Systematic Review of RCTs RCTs (including cluster randomised)
Comments	Studies that were concerned with the care and discharge management of patients hospitalised for decompensation of heart failure come under the remit of the acute heart failure guidelines. Therefore, when patients were recruited in hospital, at least one face-to-face meeting was required after the patient had been discharged for the study to be included. There is no minimum duration of intervention, but the last outcome measure must be at least three months after the intervention begins. Studies that were concerned with interventions covered elsewhere in the guideline were excluded, as these will have separate guidance. Usual care was considered likely to differ significantly from the UK / NHS standard if the study was carried out in a country outside the OECD or in the United States.

Table 98List of studies included in the review

Study	Population risk strand	Length strand	Intervention strand
Agvall 2013⁶	Low	Long	Nurse-led clinic
Aukland-HF trial Doughty 2002¹⁰³ Walsh 2000³⁶⁵	High	Long	MDT clinic
Berger 2010³⁴ Adlbrecht 2011⁴	High	Long	Case management
Capomolla 2002⁵⁶	High	Long	MDT clinic
COACH trial Jaarsma 2008¹⁵⁷ (Jaarsma 2008¹⁵⁶, Postmus 2011²⁷², Jaarsma 2004¹⁵⁵, Jaarsma 2002¹⁵⁸)	High	Long	Basic: Nurse-led clinic Intensive: Home-based MDT
DEAL-HF trial De la Porte 2007⁸⁷	High	Long	MDT clinic
Del sindaco 2007⁹¹ Pulignano 2010²⁷⁴ Del sindaco 2012⁹⁰)	High	Long	MDT clinic
Driscoll 2014¹⁰⁵	High	Mid	Nurse-led clinic
Ducharme 2005¹⁰⁶	High	Mid	MDT clinic
Ekman 1998¹¹¹ Ekman 2003¹¹²	High	Mid	Nurse-led clinic
Gonzalez-Guerrero 2014¹³⁷	High	Mid	MDT clinic
HICMan trial Peters-klimm 2010²⁵⁷ Peters-klimm 2007²⁵⁸	Low	Long	Case-management
J-HOMECARE trial Tsuchihashi-makaya 2013³⁴¹ (Tsuchihashi-makaya 2011³⁴²)	High	Mid	Case management
Ledwidge 2003¹⁹³ Mcdonald 2001²¹⁷ Mcdonald 2002²¹⁸)	High	Short	MDT clinic
Martensson 2005²¹⁴	Low	Long	Case-management
NorthStar trial Schou 2013³⁰⁰ Schou 2014²⁹⁸	Low	Long	MDT clinic
Nucifora 2006²⁴⁴	High	Mid	MDT clinic
OPTIMAL trial Mejhert 2004²²⁰	High	Long	Nurse-led clinic
PREFER trial Brannstrom 2014⁵⁰ Brannstrom 2013⁴⁹ Sahlen 2016²⁹⁰	High	Mid	Home-based MDT
PRICE trial Atienza 2004²² Ojeda 2005²⁴⁶	High	Long	MDT clinic
Rao 2007²⁷⁷	High	Short	MDT clinic
Varma 1999³⁵²	Low	Long	MDT clinic: pharmacist-led

Table 99Study details (arranged by strata)

First Author / Trial name	n.	Country	Recruitment	HFREF or both	Starts inpatient	QoL	Outcomes: Mort/ Hosp	Meds	Length (months)	Last measure (months)	Face-to-face meetings
High Risk - Short interventions
Ledwidge 2003¹⁹³	98	Ireland	Inpatients admitted with primary diagnosis HF	both	Yes		Deaths, Admissions		1.0	3	3 inpt, 2 outpt
Rao 2007²⁷⁷	112	UK	Patients that had been referred by GP for open access echo due to suspected HF, who had confirmed LVSD	both	No		Death, Admissions	Px	flex 3-12	3	2
High Risk - Mid-length interventions
PREFER trial Brannstrom 2014⁵⁰	72	Sweden	Outpatients at geriatrics dept or PC with NYHA III-IV plus one of five markers of need for palliation	both	No	EQ-5D (appears to be EQ-5D VAS)	Deaths, Admissions		6	6	mean 20
Driscoll 2014¹⁰⁵	28	Australia	Patients with HFREF from clinic for complex heart failure patients on submaximal medication doses	HFREF	No	MLWHF	Death, Admissions	Px	flex 3-6	6	3
Ducharme 2005¹⁰⁶	230	Canada	Seen in ED or admitted with primary diagnosis HF	Most HFREF	On dc		Deaths, admissions		6	6	6
Ekman 1998¹¹¹	158 (145)	Sweden	Inpatients aged >65y screened for HF, NYHA III-IV	both	On dc	Change in NYHA	Death, Admission	Px	6	6	median 6 attend
Gonzalez-Guerrero 2014¹³⁷	117	Spain	Admitted under geriatric department with acute HF	both	Yes		Death or admission		6	12	4
J-HOMECARE trial Tsuchihashi-makaya 2013³⁴¹	168	Japan	From cardiology hospital	both	Yes	SF-8	Deaths, Admissions		6	12	4
Nucifora 2006²⁴⁴	200	Italy	Inpatients <86y screened for congestive HF	both	Yes	MLWHF	Deaths, Admissions	Px	6	6	1 inpt, 3 outpt
High Risk - Long interventions
Auckland-HF Doughty 2002¹⁰³	197	New Zealand	Admitted with HF	both	On dc	MLWHFQ	Deaths. Admissions	Px	12	12	7 (incmandate d GP visits)
Berger 2010³⁴	186	Austria	Hospitalised and signs/symptoms of HFREF	HFREF	On dc		Death, Admissions		12	12	6
Capomolla 2002⁵⁶	234	Italy	Patients in inpatient Heart Failure Unit, CHF and LVEF<40%	HFREF	On dc	Time trade-off / Utility	Cardiac death, Admissions	Px	12 (ave.)	12	mean 5.5 attend
COACH trial Jaarsma 2008¹⁵⁷ - basic	683	Netherlands	Patients admitted to hospital (experienced HF centres) with signs and symptoms of HF for which hospitalisation was considered necessary NYHA II-IV, stabilised before entry to study	both	Yes	NR	Deaths, Admissions		18	18	20h
COACH Jaarsma 2008¹⁵⁷ - enhanced	679	Netherlands	Patients admitted to hospital (experienced HF centres) with signs and symptoms of HF for which hospitalisation was considered necessary NYHA II-IV, stabilised before entry to study	both	Yes	NR	Deaths, Admissions		18	18	40h (inc phone)
DEAL-HF trial De la Porte 2007⁸⁷	240	Netherlands	Inpatients and cardiology outpatients recruited, NYHA III-IV	both	No	SF-36^* MLWHF^*	Death, Days in hospital	Px^* Creat-inine	12	12	9
Del sindaco 2007⁹¹	173	Italy	Hospitalised, aged 70 or more, and discharged home. NYHA III-IV and required IV diuretics	both	On dc		Deaths, Admissions		24	24	8
OPTIMAL trial Mejhert 2004²²⁰	208	Sweden	Patients hospitalised with heart failure with NYHA II-IV and LV systolic dysfunction	both	On dc	Nottingham Health Profile	Death, Admissions	Px	flex 6-18m	Ave 37	mean 2.2 attend
PRICE trial Atienza 2004²²	153	Spain	Cardiology ward with diagnosis of primary HF	both	Yes	MLWHFQ	Deaths, Admissions	Px	16 (ave.)	28	3 outpt
Low risk (stable HF)– Long interventions
Agvall 2013⁶	160	Sweden	Patients recruited from PC with CHF and LVEF<50%	HFREF	No	SF-36^*	Death, Admissions	Px	12	12	2
HICMan trial Peters-klimm 2010²⁵⁷	197	Germany	Patients recruited from PC with hospital admission with HF in last two years, objective CHF and LVEF<45%	HFREF	No	SF-36, KCCQ	Death, Admissions	Px	12	12	3
Martensson 2005²¹⁴	153	Sweden	Recruited from PC, NYHA II-IV	both	No	SF-36^*	Death	Px	12	12	mean 8.6
Northstar trial Schou 2013³⁰⁰	921	Denmark	Stable CHF on optimal meds, LVEF<45%	HFREF	No	MLWHF	Death, Admissions	Px, Adv. effects	24 (ave)	30	est 24
Varma 1999³⁵²	83	UK	Elderly patients recruited from inpatient or outpatient with diagnosis CHF	both	No	SF-36^* MLWHF	Death, Admissions	Adherence	12	12	4

*: reported incompletely, therefore not included in analysis

: NR: measured but not reported

: Px: prescription of ACEi+/beta-blocker+/MRA

: dc: on discahrge

Table 100Papers according to population risk: risk strata and background rate of admission (IQR 0.6-1.6, values outside IQR are highlighted)

	Risk strata (proportions are estimates)
First Author / Trial name	New dxa	ADHFb	Severec	Req. titrationd	Other	Overall risk category	Admissions/pt/year in control
High Risk - Short interventions
Ledwidge 2003¹⁹³	46%	All	NS	None		ADHF	1.0
Rao 2007²⁷⁷	All	None	45%	Most		New	1.0
High Risk - Mid-length interventions
PREFER trial Brannstrom 2014⁵⁰	None	NS	All	None	suitable for palliative care	Severe	2.9 (high)
Driscoll 2014¹⁰⁵	None	None	20%	All	from clinic for “complex HF”	Titration	0.5 (low)
Ducharme 2005¹⁰⁶	NS	All	90%	50%		ADHF	2.0 (high)
Ekman 1998¹¹¹	NS	90%	mean NYHA 3.2	60%	Older patients (>65y)	ADHF	2.4 (high)
Gonzalez-Guerrero 2014¹³⁷	40%	All	NS	60%	from geriatric ward (ave 85y)	ADHF	0.8
J-HOMECARE trial Tsuchihashi-makaya 2013³⁴¹	NS	All	5%	50%		ADHF	NS
Nucifora 2006²⁴⁴	None	80%	65%	80%		ADHF	1.6
High Risk - Long interventions
Auckland-HF Doughty 2002¹⁰³	NS	All	All	10%		ADHF	1.6
Berger 2010³⁴	NS	All	All	25%		ADHF	NS
Capomolla 2002⁵⁶	NS	All	35%	None		ADHF	0.6
COACH trial Jaarsma 2008¹⁵⁷ - basic	NS	All	50%	35%		ADHF	0.7
COACH Jaarsma 2008¹⁵⁷ - enhanced	NS	All	50%	35%		ADHF	0.7
DEAL-HF trial De la Porte 2007⁸⁷	NS	30%	All	30%		Severe	NS
Del sindaco 2007⁹¹	NS	All	60%	50%	Older patients (>70y)	ADHF	NS
OPTIMAL trial Mejhert 2004²²⁰	40%	All	40%	40%		ADHF	1.6
PRICE trial Atienza 2004²²	NS	All	90%	NS		ADHF	0.5 (low)
Low Risk (stable HF) - Long interventions
Agvall 2013⁶	NS	NS	35%	25%		Nil – low risk	0.6
HICMan trial Peters-klimm 2010²⁵⁷	None	25%	30%	25%		Nil – low risk	0.4 (low)
Martensson 2005²¹⁴	None	NS	50%	40%		Nil – low risk	NS
Northstar trial Schou 2013³⁰⁰	None	None	11%	None		Nil – low risk	0.8
Varma 1999³⁵²	NS	25%	mean NYHA class 2.2	NS	Older patients (ave 75)	Nil – low risk	0.7

(a): New dx = identified as new diagnosis of heart failure at recruitment.

(b): ADHF = “Acute decompensated heart failure” risk strata signifying recent decompensation, defined by selection to study due to visit to emergency dept or inpatient admission due to heart failure, with or without congestion.

(c): Severe = proportion NYHA III-IV at baseline (note that if recruited while decompensated, this may be higher than their chronic severity).

(d): Req. titration = Requiring medication titration, device implantation or surgical intervention – if not stated, this was estimated by looking at baseline medication NS = not stated in study paper

Table 101Professions / Competencies delivering interventions (studies arranged by strata)

First Author / Trial name	Doctor (secondary care)	Nurse (secondary care)	Pharmacist	Dietician	Occupational therapist	Social worker	Physiotherapist	Primary care	Details
High Risk - Short interventions
Ledwidge 2003¹⁹³		HF nurse		Y					Education from nurse and dietician in hospital, followed by telephone contact weekly for 12 weeks by HF nurse specialist. Appts at clinic after 2 and 6 weeks (unclear composition of clinic)
Rao 2007²⁷⁷	Cardiologist	Cardiac nurse							MDT
High Risk - Mid-length interventions
PREFER trial Brannstrom 2014⁵⁰	Cardiologist + Palliative care physician	HF nurse + Palliative care nurse			Y		Y		HF team plus palliative nurse + palliative physician
Driscoll 2014¹⁰⁵	Cardiologist	Cardiac nurse							Nurse-led clinic in consultation with cardiologist, could make onward referrals to other professionals
Ducharme 2005¹⁰⁶	Cardiologist	HF nurse	Y	Y		Y			Specialist HF clinic for 6 months, contact within 72h discharge and monthly visits for 6/12
Ekman 1998¹	Cardiologist	HF nurse				Y		Physician	Nurse-run clinic in cooperation with doctors, communication with PC and social care
Gonzalez-Guerrero 2014¹³⁷	Geriatrician	Cardiac nurse							Assessment prior to d/c, at 10d, 1m and 6m post d/c by the ‘team’, plus t/c contact 48h post-d/c (nurse) and 3/12 (geriatrician)
J-HOMECARE trial Tsuchihashi-makaya 2013³⁴¹	Cardiologist	Cardiac nurse	Y	Y		Y			Nurse CM in liaison with MDT for 6 months
Nucifora 2006²⁴⁴	Cardiologist	Cardiac nurse							Pre-discharge education by experienced cardiovascular nurse, thereafter, contact with nurse to reinforce message. Face-to-face meetings with a doctor and nurse three times in next six months
High Risk - Long interventions
Auckland-HF Doughty 2002¹⁰³	Cardiologist	Cardiac nurse						Physician	Explicit partnership with primary and secondary care, involving individualised and group education, scheduled visits to clinic (unclear composition) and GP and drugtitration
Berger 2010³⁴	Cardiologist	HF nurse							CHF specialist nurse carried out home visits according to a schedule, or more if deterioration. Also scheduled medical reviews, with extra at request of nurse
Capomolla 2002⁵⁶	Cardiologist	HF nurse		Y		Y	Y	Y	MDT
COACH trial Jaarsma 2008¹⁵⁷ - basic	Cardiologist	HF nurse							Basic: additional visits with HF nurse at outpatient clinic, education protocol given.
COACH Jaarsma 2008¹⁵⁷ - enhanced	Cardiologist	HF nurse		Y		Y	Y		Intensive: weekly telephone calls after d/c, home visit by HF nurse, home visit from MDT and monthly contact.
DEAL-HF trial De la Porte 2007⁸⁷	Cardiologist	HF nurse		Y					MDT
Del sindaco 2007⁹¹	Cardiologist	HF nurse							Cardiologist case-managed, with support from nurse who made regular telephone calls
OPTIMAL trial Mejhert 2004²²⁰	Cardiologist	Cardiac nurse							Nurse-led clinic (under supervision of cardiologist)
PRICE trial Atienza 2004²²	Cardiologist	HF nurse							Formal education package delivered prior to d/c, detailing self-care, with 3/12 opt appts to optimise medical therapy and reinforce self-care messages. Unclear MDT composition
Low Risk (stable HF) - Long interventions
Agvall 2013⁶		HF nurse						Physician	Specialist HF nurse working in liaison with GP
HICMan trial Peters-klimm 2010²⁵⁷	For training primary care	For training primary care						Practice nurse and Physician	Case management by practice nurse with 1.5 day specific extra training in case management of patients with heart failure according to a clinical practice guideline. Nurse worked with primary care physician
Martensson 2005²¹⁴	For training primary care	For training primary care						Practice nurse and Physician	Training provided to all practice nurses and half of physicians - 3x3h sessions on pathophysiology of HF, evaluation and treatment of HF, and self-management. Upskilled practice nurse delivers self-care intervention
Northstar trial Schou 2013³⁰⁰	Cardiologist	HF nurse							MDT, composition unclear
Varma 1999³⁵²	Cardiologist		Y						Pharmacist care to encourage self-care and self-monitoring

Table 102Key activities of MDT, studies arranged by strata

First Author / Trial name	Meds optimisatio n	Deliver DMPa	Deterioration management	Other	Intensity (ave no. patient-professional encounters)	Usual care
High Risk - Short interventions
Ledwidge 2003¹⁹³		Y	To contact clinic, where given protocol-driven advice on increasing diuretic	Carer participation in education	Fixed face-to-face 3x in hospital + 2x outpatient, plus telephone	Primary care, with referral to cardiology at discretion
Rao 2007²⁷⁷	Y		Contact number given		min. 2 face-to-face	Primary care
High Risk - Mid-length interventions
PREFER trial Brannstrom 2014⁵⁰			IV and SC medication, plus some investigations, could be done at home	regular multi-disciplinary meetings, “person-centred care at home” model	min. 1 face-to-face, actual mean 20	Primary care +/- additional measures as usual
Driscoll 2014¹⁰⁵	Y			specifically titration of beta-blocker	min. 3 face-to-face	Primary care +/-other services as usual
Ducharme 2005¹⁰⁶	Y		Given telephone number to contact during business hours. Could receive assessment and IV diuretics if needed at the clinic		Actual ave 6 face-to-face	Follow-up according to decision of treating cardiologist (not MDT)
Ekman 1998¹	Y	Y	Contact number given, could be reviewed and admitted directly from clinic		min. 3 face-to-face, actual median 6 contacts	Primary care
Gonzalez-Guerrero 2014¹³⁷	Y	Y	Could contact geriatrician every morning	Global therapeutic regimen evaluated according to the capacities of the patient	fixed 4 face-to-face, 2 telephone	Primary care, with referral to geriatrics at discretion
J-HOMECARE trial Tsuchihashi-makaya 2013³⁴¹			Both intervention and control groups given emergency contact methods		fixed 4 face-to-face (first 2 months), 4 telephone (next 4 months)	MDT discharge education (as per intervention). Routine management by cardiologist
Nucifora 2006²⁴⁴	Y		Contact number to leave message		Fixed 4 face-to-face, plus telephone	Routine discharge information. Primary care
High Risk - Long interventions
Auckland-HF Doughty 2002¹⁰³	X	Y	Could contact GP or clinic, recommend GP assessment in first instance		min. 7 face-to-face (including GP)	Primary care with other measures as usual
Berger 2010³⁴	Y		Unclear		Min. 6 face-to-face, plus telephone	Primary care with other measures as usual
Capomolla 2002⁵⁶	Y	Y	Open access to day hospital for review and medication for decompensation without admission	Tailored interventions carried out in a day hospital	mean 5.5 (sd. 3.9)	Cardiology clinic
COACH trial Jaarsma 2008¹⁵⁷ - basic		Y	Basic: Encouraged to contact nurse if change		20h	Cardiology opt within two months of d/c, and six-monthly thereafter
COACH Jaarsma 2008¹⁵⁷-enhanced		Y	Intensive: Encouraged to self-monitor and contact nurse if symptoms increased or gained weight		40h	Cardiology opt within two months of d/c, and six-monthly thereafter
DEAL-HFtrialDela Porte 2007⁸⁷	Y	Y	Contact number given		fixed 9 face-to-face	Cardiology clinic
Del sindaco 2007⁹¹	Y	Y	Given contact number		Fixed 4 face-to-face, plus telephone	Primary care with other measures as usual
OPTIMAL trial Mejhert 2004²²⁰	Y		Could vary own diuretic dose		min. 1 face-to-face, actual range 0-10, median 1, mean 2.2 (sd. 2.3)	Primary care plus personalised written HF plan, as established care in the region
PRICE trial Atienza 2004²²	Y	Y	Contact HF clinic by phone if any changes			Cardiology and primary care
Low Risk (stable HF) - Long interventions
Agvall 2013⁶	Y		Contact number given		min. 2 face-to-face, 2 telephone	Primary care
HICMan trial Peters-klimm 2010²⁵⁷	Y			Intensity of monitoring based on current symptoms. Patient reminders for prescriptions and appointments.	min. 3 face-to-face, 3 telephone	Primary care
Martensson 2005²¹⁴			Could vary own diuretic dose	Education of primary care physician in both arms	min. 1 face-to-face, actual mean 9.6 contacts	Primary care
Northstar trial Schou 2013³⁰⁰	Y		Contact number given	Extended follow-up	est. 24	Primary care
Varma 1999³⁵²			Could vary own diuretic dose	Pharmacist discusses with pt’s physician about optimising meds	fixed 4 face-to-face	Primary care

(a): DMP = Disease management programme, usually consists of structured but individualised education about HF and self-help, often including topics such as daily weights and planning for deterioration

Table 103Clinical evidence summary: Ledwidge 2003: Short MDT clinic (MDTc) versus Primary +/- secondary care for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary +/-secondary care	Risk difference with MDTc (95% CI)
Hospitalisations	98 (1 study) 3 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, indirectness	rate ratio 0.15 (0.03 to 0.69)	255 per 1000	217 fewer per 1000 (from 79 fewer to 248 fewer)
Death	98 (1 study) 3 months	⊕⊖⊖⊖ VERY LOW^a^,^c due to risk of bias, imprecision	RR 0.92 (0.2 to 4.34)	64 per 1000	5 fewer per 1000 (from 51 fewer to 213 more)

a: Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias

b: Downgraded by one increment for indirectness as not protocol outcome of all-cause hospitalisations

c: Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both

Table 104Clinical evidence summary: Rao 2007: Short MDT clinic (MDTc) versus Primary care for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTc (95% CI)
Hospitalisations	112 (1 study) 9 months	⊕⊕⊕⊖ MODERATE^a due to imprecision	Rate Ratio 1.59 (0.81 to 3.14)	245 per 1000	145 more per 1000 (from 47 fewer to 525 more)
Death	112 (1 study) 9 months	⊕⊕⊖⊖ LOW^a due to imprecision	RR 0.45 (0.04 to 4.81)	38 per 1000	21 fewer per 1000 (from 36 fewer to 144 more)
Prescribed ACE-I	112 (1 study) 3 months	⊕⊕⊕⊖ MODERATE^a due to imprecision	RR1.32 (1.05 to 1.66)	642 per 1000	205 more per 1000 (from 32 more to 423 more)
Prescribed beta-blocker	112 (1 study) 3 months	⊕⊕⊕⊕ HIGH	peto OR 11.29 (4.95 to 25.77)	19 per 1000	194 more per 1000 (from 75 more to 467 more)

(a): Downgraded by one increment as confidence interval cross 1 MID or 2 increments as the confidence interval crosses both MID.

Table 105Clinical evidence summary: PREFER (Brannstrom 2014): Mid-length Home-based MDT (MDThome) vs Primary +/- secondary care for high risk HF

Outcomes	No of Participant s (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participant s (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary +/-secondary care	Risk difference with MDThome (95% CI)
Hospitalisations	72 (1 study)	⊕⊕⊕⊕ HIGH	rate ratio 0.28 (0.16 to 0.5)	1472 per 1000	1060 fewer per 1000 (from 736 fewer to 1237 fewer)^a
Death	72 (1 study) 6 months	⊕⊕⊖⊖ LOW^b due to imprecision	RR 2.00 (0.66 to 6.06)	111 per 1000	111 more per 1000 (from 38 fewer to 562 more)
Quality of life EQ-5D final score (appears to be EQ-5D visual analogue scale), higher=better. Scale from: 0 to 100.	72 (1 study) 6 months	⊕⊖⊖⊖ VERYLOW^b^,^c due to risk of bias, imprecision		The mean quality of life in the control groups was 52.3	The mean quality of life in the intervention groups was 8.1 higher (better) (2.03 lower to 18.23 higher)

(a): Manually calculated as Risk Difference above 1000.

(b): Downgraded by one increment as confidence interval crossed 1 MID or 2 increments as confidence interval crossed both MIDs.

(c): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

Table 106Clinical evidence summary: Ducharme 2005: Mid-length MDT clinic (MDTc) vs Primary / secondary care for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary/sec ondary	Risk difference with MDTc (95% CI)
Hospitalisations	230 (1 study) 6 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.68 (0.51 to 0.91)	983 per 1000	314 fewer admissions per 1000 (from 88 fewer to 481 fewer)
Death	230 (1 study) 6 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 0.63 (0.32 to 1.24)	165 per 1000	61 fewer per 1000 (from 112 fewer to 40 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment if the confidence interval crossed 1 MID or by 2 increments if the confidence interval crossed both MIDs.

Table 107Clinical evidence summary: Gonzalez-Guerrero 2014: Mid-length MDT clinic (MDTc) vs Primary care +/- Geriatric clinic for high risk HF

outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with Control	Risk difference with Gonzalez-guerrero 2014 (95% CI)
Hospitalisations	117 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.92 (0.6 to 1.4)	776 per 1000	62 fewer per 1000 (from 310 fewer to 310 more)
Death	117 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 0.58 (0.32 to 1.04)	379 per 1000	159 fewer per 1000 (from 258 fewer to 15 more)
Quality of life - not reported	-	-	Not estimable	-	-

(a): Downgraded by one increment as rated to be at high risk of bias

(b): Downgraded by one increment as confidence interval crosses one MID or two increments as confidence interval crosses both MIDs

Table 108Clinical evidence summary: Nucifora 2006: Mid-length MDT clinic (MDTc) vs Primary care for high risk HF

Outcomes	Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTc (95% CI)
Hospitalisations	200 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	rate ratio 1.00 (0.73 to 1.36)	802 per 1000	0 fewer per 1000 (from 217 fewer to 289 more)
Deaths	200 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 1.79 (0.78 to 4.07)	79 per 1000	63 more per 1000 (from 17 fewer to 243 more)
Quality of life Minnesota LWHFQ (change score) lower=better. Scale from: 0 to 105.	150 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision		The mean change MLWFQin the control groups was 10 (worse)	The mean in the intervention groups was 4 higher (worse) (1.82 lower to 9.82 higher)
Prescribed ACE-I	178 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a due to risk of bias	RR 0.99 (0.86 to 1.15)	806 per 1000	8 fewer per 1000 (from 113 fewer to 121 more)
Prescribed beta-blocker	178 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 0.73 (0.37 to 1.42)	194 per 1000	52 fewer per 1000 (from 122 fewer to 81 more)
Taking prescribed medication	178 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias	RR 1.04 (0.92 to 1.17)	839 per 1000	34 more per 1000 (from 67 fewer to 143 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment as confidence interval crosses 1 MID or 2 increments as confidence interval crosses both MIDs.

Table 109Clinical evidence summary: Driscoll 2014: Nurse-led clinic (MDTn) vs Primary / secondary care for high risk HFREF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary/secondary	Risk difference with MDTn (95% CI)
Hospitalisations and Emergency department attendances	25 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.67 (0.07 to 6.41)	231 events per 1000 person years	76 fewer events admissions per 1000 person years (from 215 fewer to 1000 more)
Death	25 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	peto OR 8.03 (0.16 to 406)	^c	80 more per 1000 (from 120 fewer to 280 more)^d
Quality of life MLWHFQ (change score) lower=better. Scale from: 0 to 105.	25 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias		The mean change MLWHFQ in the control groups was 9.5 (deterioration)	The mean quality of life in the intervention groups was 2.80 lower (better) (13.68 lower to 8.08 higher)
Prescribed “optimal” dose beta-blocker	24 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 2.13 (1.01 to 4.47)	385 per 1000	435 more per 1000 (from 4 more to 1000 more)

(a): ^a Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded 1 increment as confidence interval crossed 1 MID or 2 increments as confidence interval crossed both MIDs.

(c): Cannot be estimated as no events in control arm.

(d): Absolute difference calculated by RevMan.

Table 110Clinical evidence summary: Ekman 1998: Mid-length Nurse-led clinic (MDTn) vs Primary care (1 control) 3-6 months for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary/secondary	Risk difference with MDTn (95% CI)
Hospitalisations	158 (1 study) 6 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.92 (0.68 to 1.22)	1203 per 1000	96 fewer per 1000 (from 385 fewer to 265 more)
Death	158 (1 study) 6 months	⊕⊖⊖⊖ VERY LOW^a due to risk of bias, imprecision	RR 1.24 (0.71 to 2.16)	215 per 1000	52 more per 1000 (from 62 fewer to 250 more)
NYHA class change (a proxy for QoL) mean level (I-IV), lower=better. Scale from: 1 to 4.	158 (1 study) 6 months	⊕⊕⊖⊖ LOW^a^,^c due to risk of bias, indirectness		The mean NYHA class change in the control groups was −0.3 (better)	The mean NYHA class change (a proxy for qol) in the intervention groups was 0.10 higher (less improvement) (0.15 lower to 0.35 higher)
Prescribed ACE-I	145 (1 study) 6 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 1.12 (0.89 to 1.41)	627 per 1000	75 more per 1000 (from 69 fewer to 257 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by one increment as confidence interval cross one MID or downgraded by two increments as confidence interval crosses both MID.

(c): Downgraded due to indirectness as a proxy measure of quality of life was reported by the study.

Table 111Clinical Evidence Summary: J-HOMECARE (Tsuchihashi-Makaya 2013): Mid-length Case management (MDTcm) vs Cardiology clinic for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with cardiology clinic	Risk difference with MDTcm (95% CI)
Hospitalisations	98 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b^,^c due to risk of bias, indirectness, imprecision	HR 0.52 (0.28 to 0.98)	estimated risk: 213 per 1000^d	96 fewer per 1000 (from 4 fewer to 148 fewer)
Death	161 (1 study) 12 months	⊕⊖⊖⊖ VERYLOW^a^,^c due to risk of bias, imprecision	RR 1.04 (0.41 to 2.63)	98 per 1000	4 more per 1000 (from 58 fewer to 159 more)
Quality of life (physical) SF-8 physical component, final score. higher=better. Scale from: 0 to 100.	138 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^c due to risk of bias, imprecision		The mean SF-8 physical in the control groups was 42	The mean in the intervention groups was 2.00 higher (better) (1.18 lower to 5.18 higher)
Quality of life (mental) SF-8 mental health component, final score. higher=better. Scale from: 0 to 100.	138 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias		The mean SF-8 mental in the control groups was 47	The mean quality of life (mental) in the intervention groups was 2.00 higher (better) (0.67 lower to 4.67 higher)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded for indirectness as not protocol outcome of count rates for hospitalisations.

(c): Downgraded one increment as confidence interval crosses one MID or two increments as confidence interval crosses two MID.

(d): Used estimated control rate.

Table 112Clinical Evidence Summary: COACH intensive (Jaarsma 2008): Long Home based MDT (MDThome) vs Cardiology clinic in high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with cardiology clinic	Risk difference with MDThome (95% CI)
Hospitalisations	683 (1 study) 18 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	rate ratio 1.10 (0.96 to 1.27)	1109 per 1000	111 more admissions per 1000 (from 44 fewer to 299 more)
Death	683 (1 study) 18 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	HR 0.81 (0.6 to 1.08)	292 per 1000	48 fewer per 1000 (from 105 fewer to 19 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both.

Table 113Clinical Evidence Summary: Auckland-HF (Doughty 2002): Long MDT clinic (MDTc) vs Primary +/- Secondary care in high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary +/-secondary care	Risk difference with MDTc (95% CI)
Hospitalisations	197 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.76 (0.6 to 0.96)	1588 per 1000	381 fewer admissions per 1000 (from 64 fewer to 635 fewer)
Death	197 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 0.77 (0.45 to 1.31)	247 per 1000	57 fewer per 1000 (from 136 fewer to 77 fewer)
Quality of life Minnesota LWHFQ (change score) lower=better. Scale from: 0 to 105.	197 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias		The mean change in MLWHFQinthe control groups was −12.5 (improvement)	The mean change in the intervention groups was 7 lower (greater improvement) (7.82 to 6.18 lower)
Prescribed ACE-I	154 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 1.14 (0.96 to 1.35)	726 per 1000	102 more per 1000 (from 29 fewer to 254 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by one increment as confidence interval cross 1 MID or 2 increments as confidence interval crosses both MID.

Table 114Clinical Evidence Summary: Capomolla 2002: Long MDT clinic (MDTc) vs Cardiology clinic in high risk HFREF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with cardiology clinic	Risk difference with MDTc (95% CI)
Hospitalisations	234 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a due to risk of bias	rate ratio 0.18 (0.1 to 0.33)	639 per 1000	524 fewer admissions per 1000 (from 428 fewer to 575 fewer)
Cardiac Death	234 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, indirectness	RR 0.16 (0.05 to 0.51)	172 per 1000	145 fewer per 1000 (from 84 fewer to 164 fewer)
Utility (proxy for Quality of life) higher=better. Scale from: 0 to 1.	210 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^c^,^d due to risk of bias, indirectness, imprecision		The mean utility in the control groups was 0.63	The mean utility in the intervention groups was 0.09 higher (better) (0.04 to 0.14 higher)
ACE-I dose prescribed (long acting only)	210 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a due to risk of bias		The mean ace-i dose prescribed (long acting only) in the control groups was 12 mg	The mean ace-i dose prescribed (long acting only) in the intervention groups was 8 mg higher (5.5 to 10.5 higher)
Beta-blocker dose prescribed	210 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a due to risk of bias		The mean beta-blocker dose prescribed in the control groups was 13 mg	The mean beta-blocker dose prescribed in the intervention groups was 21 mg higher (13.9 to 28.1 higher)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded for indirectness as the outcome was not the protocol all-cause mortality.

(c): Downgraded as not a protocol outcome for quality of life.

Table 115Clinical Evidence Summary: DEAL-HF (De la Porte 2007): Long MDT clinic (MDTc) vs Cardiology clinic for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	cardiology clinic	Risk difference with MDTc (95% CI)
Hospitalisation Days in hospital	240 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a due to indirectness	Rate Ratio 0.56 (0.49 to 0.64)	5279 per 1000	2310 fewer days in hospital per 1000 (from 1890 fewer to 2680 fewer)^b
Death	240 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^c due to imprecision	RR 0.54 (0.28 to 1.03)	189 per 1000	87 fewer per 1000 (from 136 fewer to 6 more)

(a): Downgraded due to indirectness as not protocol outcome for hospitalisation of count rates.

(b): Manually calculated as GRADE cannot process RD above 1000.

(c): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 116Clinical Evidence Summary: Del Sindaco 2007:Long MDT clinic (MDTc) vs Primary / secondary care for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary/s econdary care	Risk difference with MDTc (95% CI)
Hospitalisations - dichotomous	173 (1 study) 24 months	⊕⊖⊖⊖ VERYLOW^a^,^b^,^c due to risk of bias, indirectness, imprecision	RR 0.75 (0.6 to 0.93)	747 per 1000	187 fewer people admitted per 1000 (from 52 fewer to 299 fewer)
Death	172 (1 study) 24 months	⊕⊖⊖⊖ VERYLOW^a^,^c due to risk of bias, imprecision	RR 0.85 (0.56 to 1.29)	314 per 1000	47 fewer per 1000 (from 138 fewer to 91 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded for indirectness as not the protocol outcomes for hospitalisations, count rate.

(c): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 117Clinical Evidence Summary: PRICE (Atienza 2004): Long MDT clinic (MDTc) vs Cardiology for high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with cardiology clinic	Risk difference with MDTc (95% CI)
Hospitalisations	338 (1 study) 16 months	⊕⊕⊕⊖ MODERATE^a due to imprecision	rate ratio 0.67 (0.54 to 0.84)	1144 per 1000	377 fewer per 1000 (from 183 fewer to 526 fewer)
Death	338 (1 study) 16 months	⊕⊕⊕⊕ HIGH	RR 1.80 (1.21 to 2.68)	172 per 1000	138 more per 1000 (from 36 more to 290 more)
Quality of life Minnesota LWHFQ, final score. lower=better. Scale from: 0 to 105.	220 (1 study) 16 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision		The mean quality of life in the control groups was 35.5	The mean quality of life in the intervention groups was 6.60 lower (better) (8.47 to 4.73 lower)
Prescribed ACE-I	153 (1 study) 16 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 0.97 (0.78 to 1.21)	688 per 1000	21 fewer per 1000 (from 151 fewer to 145 more)
Prescribed beta-blocker	153 (1 study) 16 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 1.62 (1.17 to 2.25)	390 per 1000	242 more per 1000 (from 66 more to 487 more)

(a): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

(b): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

Table 118Clinical Evidence Summary: COACH basic (Jaarsma 2008): Long Nurse-led clinic (MDTn) vs Cardiology clinic in high risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with cardiology clinic	Risk difference with MDTn (95% CI)
Hospitalisations	679 (1 study) 18 months	⊕⊕⊕⊖ MODERATE^a due to risk of bias	rate ratio 1.01 (0.88 to 1.17)	1109 per 1000	11 more per admissions 1000 (from 133 fewer to 189 more)
Death	679 (1 study) 18 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	HR 0.88 (0.66 to 1.18)	292 per 1000	30 fewer per 1000 (from 88 fewer to 43 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 119Clinical Evidence Summary: OPTIMAL (Mejhert 2004): Long Nurse-led clinic (MDTn) vs Primary care for high risk HF

Outcomes	Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTn (95% CI)
Hospitalisations	208 (1 study) 37 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.90 (0.79 to 1.02)	4895 per 1000	490 fewer per 1000 (from 1000 fewer to 98 more)
Death	208 (1 study) 37 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 1.20 (0.83 to 1.73)	324 per 1000	65 more per 1000 (from 55 fewer to 236 more)
Quality of life Nottingham Health Profile Part 1, lower=better. Scale from: 0 to 600.	208 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias		The mean quality of life in the control groups was 127	The mean quality of life in the intervention groups was 9 higher (21 lower to 39 higher)
Prescribed ACE-I	208 (1 study) 12 months	⊕⊖⊖⊖ VERYLOW^a^,^b due to risk of bias, imprecision	RR 0.9 (0.75 to 1.08)	733 per 1000	73 fewer per 1000 (from 183 fewer to 59 more)
Prescribed beta-blockers	208 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 0.89 (0.71 to 1.12)	619 per 1000	68 fewer per 1000 (from 180 fewer to 74 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both.

Table 120Clinical Evidence Summary: Berger 2010: Long Case-management (MDTcm) vs Primary +/- secondary care (1/2 control), for >6 months high risk HFREF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary +/-secondary care	Risk difference MDTcm (95% CI)
Hospitalisations - dichotomous	132 (1 study) 12 months	⊕⊖⊖⊖ VERYLOW^a^,^b^,^c due to risk of bias, indirectness, imprecision	RR 0.91 (0.76 to 1.08)	830 per 1000	75 fewer people admitted per 1000 (from 199 fewer to 66 more)
Death	186 (1 study) 18 months	⊕⊕⊖⊖ LOW^a^,^c due to risk of bias, imprecision	RR 0.56 (0.36 to 0.89)	389 per 1000	171 fewer per 1000 (from 43 fewer to 249 fewer)
Prescribed ACE-I or ARB	180 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a due to risk of bias	RR 1.01 (0.96 to 1.06)	967 per 1000	10 more per 1000 (from 39 fewer to 58 more)
Prescribed beta-blocker	186 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^c due to risk of bias, imprecision	RR 1.13 (1.03 to 1.25)	844 per 1000	110 more per 1000 (from 25 more to 211 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded for indirectness as “count rate” was the protocol outcome for hospitalisation, and this is proportion who were hospitalised at least once.

(c): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 121Clinical Evidence Summary: Northstar (Schou 2013): Extended follow-up in MDT clinic (MDTc) vs Primary care (1 control) >6 months for low risk HF (stable HFREF)

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTc (95% CI)
Hospitalisations	920 (1 study) 2 years	⊕⊕⊕⊕ HIGH	rate ratio 0.94 (0.85 to 1.05)	1509 per 1000	91 fewer admissions per 1000 (from 226 fewer to 75 more)
Death	920 (1 study) 2 years	⊕⊕⊖⊖ LOW^a due to imprecision	HR 1.05 (0.74 to 1.5)	139 per 1000	6 more per 1000 (from 34 fewer to 62 more)
Quality of life Minnesota LWHFQ (change score) lower=better. Scale from: 0 to 105.	723 (1 study) 2 years	⊕⊕⊕⊖ MODERATE^b^,^c due to risk of bias		The mean change in MLWHFQ in the control groups was 0 (IQR −1 to 0)	The mean change in the intervention groups was 1 lower/better than change in the control group (IQR 1 lower to 1 higher)2
Prescribed ACE-I	920 (1 study) 2 years	⊕⊕⊕⊖ MODERATE^c due to risk of bias	RR 1.00 (0.95 to 1.04)	885 per 1000	0 fewer per 1000 (from 44 fewer to 35 more)
Prescribed beta-blocker	920 (1 study) 2 years	⊕⊕⊕⊖ MODERATE^c due to risk of bias	RR 1.00 (0.95 to 1.05)	876 per 1000	0 fewer per 1000 (from 44 fewer to 44 more)
Adverse - serum creatinine increase >50% during follow-up	744 (1 study) 2 years	⊕⊖⊖⊖ VERYLOW^a^,^c due to risk of bias, imprecision	RR 0.94 (0.44 to 2.01)	35 cases renal failure per 1000	2 fewer renal failure per 1000 (from 20 fewer to 35 more)
Adverse - hyperkalaemia (potassium > 5.0mmol/l) at follow-up	723 (1 study) 2 years	⊕⊕⊖⊖ LOW^a^,^c due to risk of bias, imprecision	RR 0.56 (0.29 to 1.09)	63 hyperK per 1000	28 fewer hyperK per 1000 (from 45 fewer to 6 more)
Adverse - hypotension	723 (1 study) 12 months	⊕⊖⊖⊖ VERYLOW^a^,^c due to risk of bias, imprecision	RR 1.42 (0.24 to 8.42)	6 per 1000	2 more per 1000 (from 4 fewer to 42 more)

(a): Downgraded by one increment as confidence interval crosses one MID or two increments as confidence interval crosses both MID.

(b): Precision cannot be formally assessed, but interquartile range suggests small confidence interval.

(c): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

Table 122Clinical Evidence Summary: Agvall 2013: Long Nurse-led clinic (MDTcm) vs Primary care (1 control), >6 months for low-risk HFREF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTc (95% CI)
Hospitalisations	160 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.72 (0.47 to 1.11)	630 per 1000	176 fewer admissions per 1000 (from 334 fewer to 69 more) Point estimate suggests clinical benefit from MDTcm
Death	160 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 0.82 (0.23 to 2.94)	62 per 1000	11 fewer per 1000 (from 48 fewer to 120 more)
Prescribed ACE-I or ARB	160 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 1.19 (1.08 to 1.31)	840 per 1000	160 more per 1000 (from 67 more to 260 more)
Prescribed beta-blocker	160 (1 study) 12 months	⊕⊕⊖⊖ LOW^a^,^b due to risk of bias, imprecision	RR 0.94 (0.79 to 1.13)	778 per 1000	47 fewer per 1000 (from 163 fewer to 101 more)
Renal function Serum creatinine at follow-up. Lower=better	158 (1 study) 12 months	⊕⊕⊕⊖ MODERATE^a^,^b due to risk of bias		The mean serum creatinine in the control groups was 111.4 umol/l	The mean renal function in the intervention groups was 1.90 umol/l lower (11.88 lower to 8.08 higher)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 123Clinical Evidence Summary: Varma 1999: Long Pharmacist-led clinic (MDT pharm) vs Primary care (1 control) >6 months for low risk HF

Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDT pharm (95%
Hospitalisations	83 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	rate ratio 0.51 (0.27 to 0.97)	659 per 1000	323 fewer per 1000 (from 20 fewer to 481 fewer)
Death	83 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 0.98 (0.38 to 2.54)	171 per 1000	3 fewer per 1000 (from 106 fewer to 263 more)
Quality of life Minnesota LWHFQ, lower=better. Scale from: 0 to 105.	49 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision		The mean MLWHFQfinal score in the control groups was 19.1	The mean quality of life in the intervention groups was 6.40 lower (better) (0.76 to 12.04 lower)
Taking prescribed medication Self-report	49 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias	RR 1.05 (0.93 to 1.18)	957 per 1000	48 more per 1000 (from 67 fewer to 172 more)
Taking prescribed medication Automated measure	23 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR2.56 (0.95 to 6.92)	300 per 1000	468 more per 1000 (from 15 fewer to 1000 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by one increment as confidence interval crosses one MID or two increments as confidence interval crosses both MID.

Table 124Clinical Evidence Summary: HICMann (Peters-Klimm 2010): Non-specialist case management (MDTcm) vs Primary care (1 control) >6 months for HFREF

Outcomes	No of Participant s (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	No of Participant s (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTcm (95% CI)
Hospitalisations	178 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	rate ratio 1.23 (0.78 to 1.94)	374 per 1000	86 more admissions per 1000 (from 82 fewer to 351 more)
Death	190 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 1.07 (0.32 to 3.56)	51 per 1000	4 more per 1000 (from 35 fewer to 131 more)
Quality of life Kansas City Cardiomyopathy Questionnaire, final score. higher=better. Scale from: 0 to 100.	180 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision		The KCCQ final score in the control groups was 66.3	The mean KCCQ in the intervention groups was 1.70 higher (better) (3.28 lower to 6.68 higher)
Quality of life (physical) SF-36 physical health composite, final score. higher=better. Scale from: 0 to 100.	131 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias		The mean SF-physical in the control groups was 38.3	The mean SF-physical in the intervention groups was 0.3 lower (worse) (3.25 lower to 2.65 higher)
Quality of life (mental) SF-36 mental health composite, final score. higher=better. Scale from: 0 to 100.	131 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias		The mean SF-mental in the control groups was 46.6	The mean SF-mental in the intervention groups was 0.1 lower (worse) (3.5 lower to 3.5 higher)
Prescribed double therapy of ACE-I/ARB and B-blocker	180 (1 study) 12 months	⊕⊕⊖⊖ LOW^a due to risk of bias	RR 1.01 (0.84 to 1.2)	720 ACE+BB per 1000	7 more ACE+BB per 1000 (from 115 fewer to 144 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by 1 increment if the confidence interval crossed one MID or by 2 increments if the confidence interval crossed both MIDs.

Table 125Clinical Evidence Summary: Martensson 2005: Non-specialist case management (MDTcm) vs Primary care (1 control) > 6 months for high risk HF

Outcomes	Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
Outcomes	Participants (studies) Follow up	Quality of the evidence (GRADE)	Relative effect (95% CI)	Risk with primary care	Risk difference with MDTcm (95% CI)
Deaths	149 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 3.20 (0.92 to 11.17)	41 per 1000	90 more per 1000 (from 3 fewer to 418 more)
Prescribed ACE-I at target dose	130 (1 study) 12 months	⊕⊖⊖⊖ VERYLOW^a^,^b due to risk of bias, imprecision	RR 0.84 (0.61 to 1.17)	574 ACE-I per 1000	92 fewer ACE-I per 1000 (from 224 fewer to 97 more)
Prescribed beta-blocker at target dose	130 (1 study) 12 months	⊕⊖⊖⊖ VERY LOW^a^,^b due to risk of bias, imprecision	RR 0.96 (0.51 to 1.8)	235 B-blocker per 1000	9 fewer B-blocker per 1000 (from 115 fewer to 188 more)

(a): Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias.

(b): Downgraded by one increment as confidence interval crosses one MID or two increments as confidence interval crosses both MID.

Table 126Health economic evidence profile: Home-based MDT clinic (mid-length intervention) vs usual care in high risk patients

Study	Applicability	Limitations	Other comments	Incremental cost	Incremental effects	Cost-effectiveness	Uncertainty
Sahlen 2016²⁹⁰ [Sweden]	Partially applicable^(a)	Potentially serious limitations^(b)	Cost-utility analysis (health outcome: QALYs) Within-trial analysis of a RCT study⁵⁰ Population: adults with chronic heart failure with NYHA class III-IV symptoms and a marker of severity Interventions: Usual care MDT - consisting of a heart failure nurse, palliative care nurse, cardiologist, palliative care physician, physiotherapist and occupational therapist 6 month follow-up	2-1: Cost saving of £1,517	2-1: 0.03 QALYs	2 dominates 1 (more effective and less costly)	Uncertainty not reported for cost effectiveness Conclusions robust to sensitivity analyses.

: Abbreviations: QALY: quality-adjusted life years; RCT: randomised controlled trial

(a): Single centre study from a county council hospital in Vӓsterbotten County, Sweden and therefore resource use and 2012 costs may not reflect current UK NHS context.

(b): Short time horizon may not capture full costs and effects of the intervention. EQ-5D reported differently to the clinical trial evidence. Only minimal sensitivity analyses were carried out to quantify uncertainty.

Table 127Health economic evidence profile: MDT clinic (long intervention) vs usual care in high risk patients

Study	Applicability	Limitations	Other comments	Incremental cost	Incremental effects	Cost-effectiveness	Uncertainty
Atienza 2004²² [Spain]	Partially applicable^(a)	Potentially serious limitations^(b)	Cost-consequence analysis (health outcomes: 1 year mortality rate, all-cause readmissions, quality of life as measured by the MLWHFQ) Within-trial analysis of same paper in clinical review Population: people discharged from cardiology wards with a primary diagnosis of heart failure. Interventions: Usual care MDT - consisting of a specialist cardiac nurse, primary care physician and cardiologist. 16 month follow up	2-1: cost saving of £1,719	See clinical review of same paper (mortality, all-cause hospitalisation, quality of life).	n/a	None undertaken.
Pulignano 2010²⁷⁴ [Italy]	Partially applicable^(c)	Potentially serious limitations^(d)	Cost-effectiveness analysis (health outcomes: death or readmission for heart failure and all-cause admission rate) Within-trial analysis of a RCT study⁹¹ Population: people over the age of 70 with heart failure with reduced and normal ejection fraction, discharged home after a hospitalisation. Interventions: Usual care (primary/ secondary care) Multidisciplinary team consisting of a cardiologist, experienced in geriatrics (case managers), two to four specialised nurses, and the patient’s primary care physician. 2 year follow up	2-1: cost saving of £721	See clinical review of Del Sindaco 2007⁹¹	Int 2 saves £4,042 per death and/or heart failure-related admission avoided. Int 2 saves £2,155 per all-cause admission avoided.	None undertaken.

: Abbreviations: QALY: quality-adjusted life years; RCT: randomised controlled trial

(a): Spanish resource use data and unit costs (year not reported, assumed to be 2004) may not reflect current NHS context. QALYs were not used as the health outcome measure

(b): Within-trial analysis and so does not reflect the full body of available evidence available for this intervention. Atienza is 1 of 5 studies included comparing MDT clinic to usual care in high risk patients. No exploration of uncertainty.

(c): Italian national health service resource use and unit costs may not reflect current UK NHS context. QALY data was not reported clearly enough to report and therefore were not used as the health outcome measure. Discounting was not applied.

(d): Within-trial analysis and therefore does not reflect the full body of evidence available for this comparison. Pulignano is 1 of 5 studies included comparing MDT clinic to usual care in high risk patients. No exploration of uncertainty.

Table 128Health economic evidence profile: Home-based MDT (long intervention) vs nurse led MDT clinic (long intervention) vs usual care in high risk patients

Study	Applicability	Limitations	Other comments	Incremental cost	Incremental effects	Cost-effectiveness	Uncertainty
Postmus²⁷² [Netherlands]	Partially applicable^(a)	Potentially serious limitations^(b)	Cost-utility analysis (health outcome: QALYs) Within-trial analysis of a RCT study¹⁵⁷ Population: Patients >18 years old with evidence of structural cardiac dysfunction Interventions: Usual care (cardiology clinic) Basic MDT Intensive MDT 18 month follow-up	2-1: cost saving of £58 3-1: £828 3-2: £886	2-1: 0.023 3-1: 0.019 3-2: −0.004	2 dominates 1 and 3 (more effective and less costly)^(c)	Probability intervention 2 cost-effective (€20K threshold): 62% Conclusions robust to sensitivity analysis.

: Abbreviations: QALY: quality-adjusted life years; RCT: randomised controlled trial

(a): This analysis has been undertaken from a Dutch perspective using 2009 unit costs and therefore may not reflect current NHS context. Does not include important cost aspects such as procedures during hospital admission. EQ-5D was not used.

(b): Probabilistic sensitivity analysis was not presented at £20k/QALY. No discounting was undertaken; however the time-horizon was only 18 months and so is unlikely to have a significant effect. Excluded drug costs, cost of procedures conducted during hospitalisation or short term hospital admission as not rigorously reported in associated trial.

(c): This study also reports results for sub-groups according to NYHA class. In patients with less severe heart failure (NYHA class I-II) intervention 2 dominates interventions 1 and 3. However, in patients with severe heart failure (NYHA class III-IV) intervention 1 is the most cost effective option.

Table 129Health economic evidence profile: MDT (long intervention - case-management) vs usual care (primary +/- secondary) in high risk patients

Study	Applicability	Limitations	Other comments	Incremental cost	Incremental effects	Cost-effectiveness	Uncertainty
Moertl 2012²²³ [Austrian]	Partially applicable^(a)	Potentially serious limitations^(b)	Cost-utility analysis, Markov model Population: Patients with heart failure discharged after a heart failure hospitalisation Comparators: Usual care in community Nurse-led MDT 20 year time horizon modelled	2-1: £2,089	2-1: 0.68 QALYs	ICER: £3,072 per QALY gained^(c)	Uncertainty not reported for cost effectiveness Conclusions robust to sensitivity analyses (including discount rate to 3% and 0%).

: Abbreviations: ICER: Incremental cost effectiveness ratio; MDT: multidisciplinary team; QALY: quality-adjusted life years.

(a): Austrian payer perspective. EQ-5D not used to capture quality of life - utility scores converted from MLWHF questionnaire using previously published algorithm. Costs and effects discounted at 5%.

(b): Cost of GP visits and drug costs were not collected and not included in the analysis of the clinical trial phase.

(c): An additional comparator was also included in the study that involved intensive NT-proBNP guided patient management in addition to multidisciplinary care. This comparator dominates (more effective, less costly) both usual care and the nurse-led MDT alone.

Recommendations	The core specialist heart failure multidisciplinary team (MDT) should work in collaboration with the primary care team, and should include: a lead physician with a subspecialty interest in heart failure (usually a consultant cardiologist) who is responsible for making the clinical diagnosis a specialist heart failure nurse a healthcare professional with expertise in specialist prescribing for heart failure. [2018] The specialist heart failure MDT should directly involve, or refer people to, other services, including rehabilitation services, and tertiary and palliative care, as needed. [2018] The specialist heart failure MDT should: diagnose heart failure give information to people newly diagnosed with heart failure (see section 9.4.6) manage newly diagnosed, recently decompensated or advanced heart failure (NYHA [New York Heart Association] class III to IV) optimise treatment start new medicines that need specialist supervision continue to manage heart failure after an interventional procedure such as implantation of a cardioverter defibrillator or cardiac resynchronisation device manage heart failure that is not responding to treatment. [2018]
Relative values of different outcomes	The critical outcomes were identified as all-cause hospitalisations, mortality and quality of life. Important outcomes included medicine optimisation (including the proportion of people prescribed medication and the proportion taking the medication as prescribed). Also important were outcomes felt by the committee to indicate potential adverse events from over-treatment: renal failure, hyperkalaemia and hypotension. Patient and carer preferences, and dying in preferred place of death for palliative care patients, were also considered important, but these outcomes were not reported by any study.
Quality of the clinical evidence	Twenty two studies (23 comparisons) were identified. The studies were stratified into 2 categories of heart failure risk (high risk and low risk) and 3 lengths of intervention duration (3 months or less, between 3 and 6 months and over 6 months). The population risk (high versus low) was based on the existing recommendations for people who may benefit more from an MDT. The included those with a new diagnosis; requiring medication titration and/or surgical procedure; recent deterioration and severe and/or unresponsive disease. In the current review 14 of the 17 studies in the high risk category selected patients on the basis of recent deterioration, with only 1 selecting new patients, 1 specifically targeting patients requiring medication titration, and 1 selecting on the basis of severe HF with a need for palliation. Population risk in the studies was also described by the average number of all-cause hospital admissions per person per year in the control arms, where this could be calculated. There were only 3 studies that were 3 months or less; and all studies included in the low risk category were over 6 months long. The studies had a heterogeneous mix of populations, interventions and standards of usual care. The studies also ranged in scale from 25 participants followed for 6 months to 920 participants followed for an average of 4 years. Due to the heterogeneity, meta-analysis of the studies was not considered appropriate. The quality was variable, with the main reasons for downgrading due to risk of bias being a lack of detail about randomisation/allocation concealment and imprecision due to small study size. Many of the studies had small numbers of patients or events, with large confidence intervals around the effect estimates ranging from a clinically important harm to a clinically important benefit. Quality of life was downgraded for performance bias in all studies, as it was not possible to blind participants. There was also concern that some larger studies stated intent to measure quality of life in the study protocol, but with limited or no results presented, suggesting selective reporting. The committee considered that overall, the evidence was of low quality, often imprecise, and noted that even the direction of effect was inconsistent between outcomes and between studies. The numbers of hospitalisation and deaths in many of the studies was less than might be expected, and that this may represent selection bias towards healthier patients in these trials. Therefore, there may be problems generalising the evidence to the current heart failure population in the UK. In addition, many of the studies were over twenty years old, and most not in the context of current NHS practice, further restricting their applicability. The committee considered whether further research in this area would improve recommendations in this area. While the current evaluation is hampered by heterogeneous small trials, there were some larger trials such as COACH and NorthStar, which have given some moderate and high quality findings. It was not felt to be worthwhile to prioritise research that aimed to add to or improve on these at the current time.
Trade-off between clinical benefits and harms	There was substantial heterogeneity in the study designs, patient populations, types of intervention and lengths of follow-up, and the evidence was of poor quality for many of the outcomes. Because of this, the committee generally found it challenging to identify which MDT interventions and team compositions were likely to be of the greatest benefit, and for whom. In patients in the high risk categories identified in the last guideline, there was some evidence that the involvement of the MDT may lead to the prescription of more appropriate medication, decrease number of hospitalisations and improve quality of life. It was discussed that in high risk patients, while most interventions did decrease hospitalisations, there were increased deaths in some of these studies. The possibility was raised that the avoidance of hospitalisation could have increased the risk of mortality, but as the evidence on mortality was very imprecise and based on small numbers, it was felt this was likely to be a chance finding. There seemed to be evidence of a general effect of extra hours of contact being associated with improved outcomes in high risk patients, with some caveats. First, there seems to be a ceiling effect: although some shorter interventions delivered improvement, extended follow-up (over years) in a clinic offered little benefit; and delivering 40 hours of contact gave only the same benefit of 20 hours of contact. Secondly, expertise appeared to count, as contact with a primary care nurse delivering a heart-failure intervention without specialist expertise in heart failure had no benefit. The committee agreed that the overall evidence supported the general approach that the MDT be used to stabilise and optimise patients. There was no convincing evidence for substantial changes to the previous recommendation regarding which patients should be cared for by the specialist multidisciplinary heart failure team, though the committee agreed by consensus on some clarifications to the previous recommendation text. These included a referral to the MDT should be considered when there was a need to optimise medication or consider specialist options in medication, and implantation of device therapy. The committee discussed specifying the competencies that should be present in the specialist multidisciplinary heart failure team. Clearly, a lead physician with specialist expertise in heart failure was critical, and this will usually be a consultant cardiologist, but there may be areas where other consultants would also be suitable, such as a consultant in elderly care, . The committee noted that the main professionals involved in delivering or coordinating interventions in the included studies were cardiovascular or heart failure specialist nurses. Based on this evidence, the committee agreed that a nursing element was needed, and emphasised that a specialist heart failure nursing competency was necessary; as the evidence suggested that it was not possible to get the same clinical improvements from generalist nurses. There was also evidence that the presence of prescribing expertise (whether by a nurse prescriber or pharmacist) can improve the prescription of some medications, and the committee agreed that the MDT should also include specific prescribing expertise. The committee also noted that there was evidence that involvement of the patient’s primary care team was shown to be beneficial, and this was in accordance with the negative experiences of some of the committee members when primary care was not involved in the management plans. It was felt that the multidisciplinary specialist HF team members provide input to the management of patients with heart failure defined above in collaboration with their primary care team. In this way, the primary care team can help facilitate communication and collaboration between healthcare professionals wherever the patient’s care is being delivered across different settings. This may include a GP who has completed a course for special interest registration
Trade-off between net clinical effects and costs	Five relevant economic evaluations were included in this review which compared an MDT to usual care. All of these studies were assessed as partially applicable with potentially serious or very serious limitations. One economic evaluation assessed high risk patients receiving a mid-length (3-6 month) home-based MDT clinic including a cardiologist, heart failure nurse, palliative care physician, palliative care nurse, physiotherapist and occupational therapist. This was based on the PREFER trial by Brannstrom et al. 2014 included in the clinical review. This within-trial cost-utility analysis found that this multidisciplinary approach is more effective and less costly, and therefore dominated (more effective, less costly) usual care. This study suggests that an MDT is likely to be cost saving compared to usual care. The remaining 4 economic evaluations assessed high risk patients receiving long term (greater than 6 months) MDT interventions. All of these economic evaluations were within-trial analyses based on RCTs included in the clinical review. One economic evaluation compared a basic MDT, an intensive MDT and usual care. Both MDT approaches involved a cardiologist and a specialist heart failure nurse. The basic MDT consisted of 20 hours of contact time, whereas the intensive MDT consisted of 40 hours contact time. This withintrial cost-utility analysis was based on the COACH trial by Jaarsma et al. 2008 included in the clinical review. The analysis found that the basic MDT intervention dominated (more effective, less costly) both the intensive MDT intervention and usual care. This paper also undertook sub-group analyses for patients with severe (NYHA class III/IV) and less severe (NYHA class I/II) heart failure. For patients with less severe HF again the basic MDT dominated both the intensive MDT and usual care. However, for patients with severe heart failure neither the basic or intensive MDT interventions were cost effective at £20,000 per QALY threshold compared to usual care (Intensive MDT ICER: £44,625). The committee discussed that those with severe heart failure are likely to be having much more input from specialists, similar to an MDT, when receiving usual care and therefore were not surprised that a formal MDT was not cost effective in this group of patients. One economic evaluation compared MDT involving a cardiologist and a specialist heart failure nurse compared to usual care. This was based on a RCT by Berger 2010³⁴ included in the clinical review. This cost-utility analysis found that the MDT was cost effective compared to usual care (ICER: £3,072 per QALY gained). This study also reported a third intervention consisting of NT-proBNP guided management in addition to multidisciplinary care. When this comparator is included in the analysis it dominates both MDT alone and usual care. One economic evaluation compared MDT involving a cardiologist experienced in geriatrics, specialist heart failure nurses, and a primary care physician to usual care. This was based on a RCT by Del Sindaco et al. 2007 included in the clinical review. This within-trial cost-effectiveness analysis found that the MDT approach saves £4,042 per death and/or heart failure-related admission avoided, and saves £2,155 per all-cause admission avoided. One economic evaluation compared MDT involving a cardiologist, specialist cardiac nurse and primary care physician to usual care. This was based on the PRICE RCT by Atienza 2004 included in the clinical review which also reported intervention costs. This study found that the MDT approach was less costly than usual care; however, the overall effects of the MDT were uncertain. The outcomes extracted in the clinical review show moderate quality evidence of a clinical harm of the MDT for mortality, but moderate quality evidence of a clinical benefit from reduced hospitalisations, and low quality evidence of an improvement in quality of life. As mentioned above, the evidence on mortality was very imprecise and based on small numbers. Overall the committee considered that overall the MDT may be cost saving and provide a clinical benefit; however, this is uncertain. No studies were identified that assessed the cost effectiveness of short term MDT intervention in high risk heart failure patients or any length MDT intervention in low risk heart failure patients. The committee agreed that the economic evidence suggests that an MDT reduces costs overall compared to usual care due to reduced hospitalisations, and noted that although each of the studies consisted of different competencies, the committee noted that a cardiologist and specialist nurse were included in all of the economic studies reviewed and therefore agreed that a physician with a subspecialty interest in heart failure, and a specialist heart failure nurse should be included in the core MDT. The committee discussed that the current composition of MDTs in clinical practice varies. The core professionals in most teams currently consist of a cardiologist and a specialist nurse as a minimum and in some cases can include many other professionals from other specialties. The committee therefore considered that the recommendation to have an MDT with a core team consisting of a physician with subspecialty interest in heart failure, a specialist heart failure nurse, competencies to manage prescribing, that could refer to other specialties if necessary, would not have an overall cost impact.
Other considerations	Current NHS practise in this area is variable. It was said that in some areas there is a lack of access to specialist heart failure teams, despite the previous recommendations. The committee was aware of the BHF Heart Failure Audit: this was carried out in a large proportion of NHS hospitals in England and Wales, and shows that following a decompensation 70% of patients receive an appointment for a cardiologist, and 60% of patients have a referral to a specialist cardiac nurse. The committee heard from the cardiologists in the group, who considered the HF nurses to be invaluable in managing their caseloads and clinics. The specialist nurses could often take on the management of the heart failure patients who were in need of more comprehensive support. They reported that the psychological support and care that specialist nurses provide could not be replicated by a cardiologist stand-alone clinic. They also noted that it has been the practice in the UK in many areas for specialist nurses or prescribers to optimise evidence-based medication (that is, ACE-inhibitors and beta-blockers, with MRA where appropriate), but that this is currently determined locally. From another perspective, the committee heard that patients very much valued someone to speak to about their illness, and that this is currently often delivered by HF nurses. They also spoke about how the MDT can encourage self-management, and this was more than just giving information or education, but included giving a patient confidence. They felt this was often not provided to the same level in primary care, and it may not be picked up by the studies. The committee heard from the general practitioners that they recognised that there were high risk HF patients who need more intensive support, particularly patients who had just been discharged from hospital after an acute decompensation. However, they felt that specialist input may not be required for people who are stable. They emphasised the importance of holistic care, which is best delivered by primary care professionals, particularly as most patients with heart failure have several other conditions. They would value a system that would offer timely extra support to patients and GPs when necessary, but kept primary care at the centre. The concept of medicines optimisation in HF was discussed. Often the studies reported adding medication and up-titration, but optimisation is about more: starting, stopping, adjusting, and monitoring, all tailored to the individual patient. It is especially important in this population that the prescriber is adjusting their approach depending on comorbidities, and whether this occurred is not well reported in the studies. The NICE guidelines on multi-morbidity (https://www.nice.org.uk/guidance/ng56) were referred to as helpful in this area. Given the large and increasing population of patients with heart failure, with limited NHS resources, the committee agreed that the recommendations needed to prioritise specialist MDT access for those in greatest need who will gain the greatest benefit. If stable patients are managed in general practice, they will need regular review, informed by their needs. The committee emphasised the importance of a personalised approach, such as that described in the NICE multi-morbidity guideline (https://www.nice.org.uk/guidance/ng56).

Table 130Characteristics of review question

Objective	Explore patient and staff experiences and preferences regarding transition and continuity of care at the interface of different care settings in heart failure. This may enable the identification of barriers (where the problems are) and facilitators (examples of good practice) to continuity of care when transitioning between heart failure care settings.
Population and setting	Patients with heart failure in a primary care, outpatient or community setting. Carers for such patients, both family/informal carers, and health-care professionals (HCP)
Context	Any description of patient or staff member experiences or preferences regarding transition and continuity of care at the interface of different care settings.
Review strategy	Synthesis of qualitative research. Quality of the evidence will be assessed by a GRADE CerQual approach for each review finding.

Figure 6Framework for synthesising review results

Table 131Summary of studies included in the review. Abbreviations: Pt = patients with CHF, HCP = health care professionals, NYHA = New York Heart Association Functional Classification

Study	Design	Population	Research aim	Comments
Aldred 2004⁸	Semi-structured interviews with pt + carers, taped and transcribed verbatim. Data coded and analysed by two researchers. Four themes most relevant to aims presented.	N=10 CHF pt recently discharged from hospital and one person each that they lived with and nominated a carer. Age: mean 72 (SD 5) Severity NYHA II-IV (6/10 NYHA III)	Explore the impact of heart failure on the lives of older pt and their informal carers.	Part of larger study on palliative needs. Setting: UK, Barnsley.
Andersson 2013¹⁴	Semi-structured interviews with pt about daily life with CHF, and their experience of information-giving and follow-up. Informed by grounded theory.	N= 4 pt who had been treated in a HF clinic, and were now discharged to primary care. Ages 60, 62, 63, & 84 Severity: Not stated	Investigating whether pt’ need for information, education and knowledge are met to the same extent in the HF clinic and primary care.	Setting: Small town in Sweden.
Baudendistal 2015²⁹	Pt interviews with open section and a second more focused part. Content analysis by two researchers independently, then discussed within the research team.	N= 17 pt identified by GPs with CHF with LVEF<35%. Age: mean 72 (SD 12) Severity: Not stated	Explore perspectives of pt with CHF on their treatment across multiple care settings and to what extent these perspectives are represented in current quality indicators.	Part of QUALIPAT heart project, which aims to find patient-centric quality indicators for care. Setting: Academic GP practices, Heidelberg, Germany.
Boyd 2004⁴⁷	Semi-structured interviews with pt with advanced CHF, carers, health and care professionals. Pt interviewed every three months for up to a year. Interviews were followed by focus groups. Concurrent analysis using narrative analysis framework.	N= 20 pt with advanced CHF and their carers (family/informal 27 interviews, professional 30 interviews). Five died during follow-up and family gave bereavement interview. Age: mean 70 (range 57-92) Severity: NYHA IV	Provide a patient-centric account of the changing and evolving needs of people with advanced heart failure, and how services address these.	Part of larger palliative care project. Setting: UK, region unclear
Fuat 2005¹²⁴	Semi-structured interview with CHF-HCP. Analysis follows “pragmatic variant” grounded theory with a degree of constant comparison.	N= 12 HCP involved in specialist HF services (cardiologists, geriatricians, general physicians and specialist GPs).	Explore reasons for the variations in the diagnosis and management of heart failure and identify barriers to the provision of uniformly high standards of care.	Setting: UK, Durham and Tees SHA
Gallacher 2011¹²⁶	Qualitative secondary analysis. Use archived interviews with pt collected for related research (unpublished). Analysis used “Normalisation Process”. Two authors designed a coding framework, while a third adjudicated.	N= 47 pt with a CHF diagnosis based on echo, taking ACE-inhibitors and a diuretic, sampled from primary care to mirror demographics of CHF. Age: mean 73 (range 45-88) Severity: Not stated	Identify and understand the components of treatment burden to inform the development of tools to measure this, HF being a condition likely to have high treatment burden and comorbidity.	Setting: UK, not clear where or when.
Gastelurrutia 2012¹²⁷	Semi-structured interviews with HCP about health problems commonly comorbid with CHF (hyperuricemia, anti-platelet agents, anaemia and diabetes). Analysed using a total sample, open coding, constant comparative approach.	N= 5 internal medicine specialists and cardiologists from a tertiary hospital HF clinic.	Explore experiences in the pharmacological management of common comorbid health problems in heart failure in order to help clinical pharmacists provide real and practical help.	Setting: HF clinic in Spanish hospital.
Glogowska 2015¹³³	In-depth interviews with HCPs using a topic guide. Also allowed participants to raise their own issues, which could be carried forward to subsequent interviews. Analysed using the constant comparative method and systematic open coding.	N= 24 clinicians (doctors, nurses and rehab workers) sampled from three healthcare settings: primary, community, and hospital in each of three geographical locations (i.e. nine settings total).	Gain an understanding of the issues facing clinicians as they care for people with heart failure in the light of recent developments including the introduction of specialist heart failure nurses.	Setting: UK. Locations were healthcare networks in South West, South Central and the Midlands with different models for providing HF care.
Heckman 2014¹⁴⁴	Qualitative descriptive study about HF management in care homes, nested in a mixed-methods protocol. Three semi-structured focus groups of HCP. Data was analysed using thematic content analysis by two researchers. Findings were presented back to the participants.	N= 18 HCP: 16 primary care physicians and 2 nurse practitioners who provided care to one of three long-term care facilities chosen to offer variety.	Explore perceptions of HCP regarding HF care in care homes, particularly why these HF pt were less likely to be receiving medication, despite the high burden of disease and acute care episodes.	Part of programme aiming to develop care processes to manage HF in care homes. Setting: Northern Ontario, Canada
Lord 2015²⁰²	Qualitative, service evaluation study. Semi-structured interviews with HCP. Data collated and analysed using Framework Method. Initial findings were fed back to the participants.	N= 21 HCP involved in the delivery of HF from three trusts with different models of providing HF care: 8 nurses, 6 consultants, 2 senior managers, 3 commissioners & 4 GPs.	Understand how HF services are delivered in three different trusts, and especially how 1° and 2° care interact to provide continuity of care for HF pt in a context of increasing demand and financial pressure.	Setting: UK, three settings in Birmingham and the Black Country.
Macdonald 2016²⁰⁷	Secondary analysis of qualitative data of transcripts of interviews with pt. Use form of amplified analysis to fit themes to the Candidacy framework.	N=20 CHF pt (a selection of transcripts from two previous HF studies) Age: range 56-86 Severity: 10 advanced and 10 “stable”	Contrast the helpseeking and access to care in cancer and heart disease in order to extend concepts about illness identity, and its relationship to the concept of “Candidacy”	Compares CHF with cancer. Data was taken from the Colorectal Cancer, End-Stage HF and Stable HF studies. Setting: UK, region not stated
MacKenzie 2010²⁰⁹	Mixed methods service evaluation of a new HF nurse service. Used a questionnaire sent to HCP for gathering quantitative and qualitative data. Free text boxes allowed for responses to four sections. No detail on analysis.	N= 86; 83 GPs (32% of those mailed) and all 3 HF specialist nurses returned questionnaires, although not stated how many added free text.	To assess acceptability and effectiveness of a new community based nurse-led HF service in an area with a dispersed population; assess the knowledge and needs of the GPs and assess the perceptions of national guidance.	Setting: UK, the Highlands Although in context of the NHS, the new HF nurse posts were funded by a charity.
Nordgren 2007²²⁵	Qualitative study from lifeworld perspective. Unstructured interviews with “middle-aged” pt with CHF(65 and under), with focus on eliciting lived experience of care. Used phenomenological analysis.	N= 7 pt currently attending HF clinic aged 65 and under, chosen for richness and variety of experience. Age: range 39-65 Severity: moderate to severe with at least one hospitalisation.	Explore how “middle-aged” people with moderate-severe HF experience and understand formal care.	Setting: HF clinic in Sweden
Sanders 2008²⁹⁴	Qualitative study nested in a larger project. Semi-structured interviews with HCP. Coded according to themes, subsequently explored in relation to the literature, using a variation of the constant comparison method.	N= 33 HCP, including GPs (7), HF nurses (10), cardiologists (8) and geriatricians (8), from ten sites, chosen for variety.	Explore views of HCP on managing heart failure pt.	Part of bigger project around communication in HF (unpublished). Setting: UK, in multiple sites in North of England
Tait 2015¹⁸⁷^,³²⁵	Qualitative, using Complex Adaptive System (CAS) theory. Interviews with pt to identify all other people involved in their care, followed by attempts to interview all in these networks. A constant comparison method to improve interview. Used constructivist grounded theory to build an explanatory theory.	N= 50 pt networks, including specialist HF services and at least two of the pts’ carers. Carers included family/informal care, caring professionals and healthcare professionals from both specialist and general services. Some HCP appeared in more than one pt network). Sampled to enrich applicability to palliative care. Age (pt): Not stated Severity: NYHA III-IV	Better understand the behaviour of the teams providing heart failure care to those with advanced HF in order to plan how best palliative care services can integrate into these teams.	Setting: Five Canadian cities in three provinces.

: Abbreviations: Pt = people with chronic heart failure, HCP = healthcare professional

Figure 7Main review findings and connections between them

Table 132Summary of Evidence for Continuity

Study design and sample size		Findings	Quality assessment
No of studies contributing to the finding	Design	Findings	Criteria	Rating	Overall assessment of confidence
1a Lack of Continuity in HF Care
4¹²⁶^,¹⁴⁴^,²⁰⁷^,²²⁵	Focus groups of HCP in care homes Interview of pt aged 65 and under Secondary analysis of pt interviews	Both people with heart failure and health professionals felt that heart failure care was fragmented leading to a lack of continuity, with negative consequences for patient experience.	Methodology	moderate limitations	MODERATE
			Coherence	no or very minor concerns
			Relevance	minor concerns
			Adequacy	no or very minor concerns
1b Primary Care
4²⁰²^,²⁰⁹^,²²⁵^,³²⁵	Interview of pt aged 65 and under Interviews of HCP Interviews of patient and care network Questionnaire to HCP	People with heart failure and professionals working in heart failure care both identified that primary care, and general practitioners in particular, could provide individualised care with continuity. HF services can cause pt to lose touch with primary care and lose this continuity.	Methodology	minor limitations	MODERATE
			Coherence	minor concerns
			Relevance	minor concerns
			Adequacy	no or very minor concerns
1c Discharge from HF Clinic
2¹⁴^,²⁹⁴	Interviews with pt Interviews with HCP	People with heart failure and HF nurses both spoke of discharge from HF clinic as being a loss to the patient in terms of no longer being proactively followed up.	Methodology	serious limitations	LOW
			Coherence	no or very minor concerns
			Relevance	minor concerns
			Adequacy	moderate concerns

Table 133Summary of Co-ordination

Study design and sample size		Findings	Quality assessment
No of studies contributing to the finding	Design	Findings	Criteria	Rating	Overall assessment of confidence
2a Poor Co-ordination Between Services
6²⁹^,⁴⁷^,¹²⁶^, ^#2771^,²²⁵^,³²⁵	2 x interviews of pt with advanced CHF and their care network 3xInterviews with pt (inc. 1 ≤65y and 1 ≥60y with their carer) Secondary analysis of interviews of pt	People with heart failure experience the healthcare system as being poorly co-ordinated, leading to increased burdens on them and to them being given conflicting advice. This in turn resulted in uncertainty and loss of confidence.	Methodology	moderate limitations	MODERATE
			Coherence	no or very minor concerns
			Relevance	minor concerns
			Adequacy	no or very minor concerns
2b Models to co-ordinate care
5¹²⁴^,¹³³^,²⁰²^,²⁹⁴^,³²⁵	2 x interviews of mix of HCP 2 x interviews of HCP in HF services Interviews of pt and their care networks	Health professionals were aware of several models that could be used to better co-ordinate care. Suggestions included: having a single professional to co-ordinate care; using heart failure nurses to facilitate cross-boundary working; and having conjoint working between primary and secondary care via means of a ‘shared-care agreement’.	Methodology	serious limitations	LOW
			Coherence	minor concerns
			Relevance	minor concerns
			Adequacy	moderate concerns

Table 134Summary of Evidence for Communication

Study design and sample size		Findings	Quality assessment
No of studies contributing to the finding	Design	Findings	Criteria	Rating	Overall assessment of confidence
3a Poor communication between services
6²⁹^,⁴⁷^,¹²⁶^,¹³³^,¹⁴⁴^,²⁰⁷	Focus group with HCP in care homes Interviews with pt with advanced CHF and their care network Interviews with pt 2 x secondary analysis of interviews with pt Interviews with HCP	People with heart failure, their informal carers and professionals within and outside HF services all felt that there was poor communication across boundaries, and this affected patient experience and patient care.	Methodology	serious limitations	MODERATE
			Coherence	no or very minor concerns
			Relevance	no or very minor concerns
			Adequacy	no or very minor concerns
3b Barriers to clear communication
2²⁰²^,²⁰⁹	Interviews of HCP in HF care Questionnaire to HCP	Healthcare professionals felt that some barriers to cross-boundary communication were fragmented and incompatible information systems, and difficulties trying to speak to other health professionals.	Methodology	moderate limitations	LOW
			Coherence	no or very minor concerns
			Relevance	no or very minor concerns
			Adequacy	serious concerns
3c Information after discharge from HF services
1¹⁴	Interviews with pt	People with heart failure experience being well-informed in HF clinic, but receiving no information after being discharged to primary care, which some took to mean they no longer needed care for their heart failure.	Methodology	serious limitations	VERY LOW
			Coherence	no or very minor concerns
			Relevance	minor concerns
			Adequacy	serious concerns

Table 135Summary of Evidence for Contact / Access

Study design and sample size		Findings	Quality assessment
No of studies contributing to the finding	Design	Findings	Criteria	Rating	Overall assessment of confidence
4a Access to Routine Care
3¹²⁶^#2749^,²⁰⁷	2 x secondary analysis of pt interviews Interview of HCP	Whether wanting continued access, a new referral or a re-referral, people with heart failure found that it took them time and effort, and that the illness itself could make it harder to stand up for themselves.	Methodology	serious limitations	LOW
			Coherence	no or very minor concerns
			Relevance	minor concerns
			Adequacy	minor concerns
4b A Primary Contact Person
1²⁹	Interviews with pt	People with heart failure expressed a preference for a primary contact person who offered a long-term relationship, openness, trust and appreciation – most would choose their GP – a key role of whom would be to provide easy access to care, e.g. timely appointments.	Methodology	moderate limitations	VERY LOW
			Coherence	no or very minor concerns
			Relevance	serious concerns
			Adequacy	moderate concerns
4c Access to Urgent Care
3²²⁵^,²⁹⁴^#2771	Interviews with pt 65 and under Interviews with pt 60 and over and carer Interviews with HCP	For people with heart failure that were under the care of HF clinics and nurses, it was noted that they had easy access to specialist doctors and nurses, who were able to react more quickly than in conventional service models, but could not provide continuity.	Methodology	moderate limitations	LOW
			Coherence	no or very minor concerns
			Relevance	moderate concerns
			Adequacy	moderate concerns

Table 136Summary of Evidence for Role of Specialist HF Services

Study design and sample size		Findings	Quality assessment
No of studies contributing to the finding	Design	Findings	Criteria	Rating	Overall assessment of confidence
5a Expectations
3⁴⁷^,²⁰²^,²⁹⁴	Interviews with HCP Interviews with HCP in HF services Interviews with pt with severe CHF and their care network	Expectations of HF services from primary care and pt can differ both from what HF services think primary care and pt want, and from what the service is capable of delivering.	Methodology	serious limitations	VERY LOW
			Coherence	minor concerns
			Relevance	moderate concerns
			Adequacy	serious concerns
5b Focus
2¹²⁷^,³²⁵	Interviews with HCP from HF services Interviews with severe CHF and their carer network	Healthcare professionals felt that HF services should be focussed on heart failure rather than its comorbidities, as addressing people’s “primary care issues” has other consequences.	Methodology	moderate limitations	LOW
			Coherence	no or very minor concerns
			Relevance	moderate concerns
			Adequacy	moderate concerns
5c Decision to keep patients on the caseload
2²⁰²^,³²⁵	Interviews with pt and care network	Healthcare professionals involved in HF specialist services vary widely on whether they take over patients, basing their decisions on the balance of a person’s needs for HF care and resources, e.g. the perceived ability of their general practitioner.	Methodology	minor limitations	LOW
			Coherence	no or very minor concerns
			Relevance	moderate concerns
			Adequacy	moderate concerns

Recommendations	The primary care team should carry out the following for people with heart failure at all times, including periods when the person is also receiving specailist heart failure care from the MDT: ensure effective communication links between different care settings and clinical services involved in the person’s care lead a full review of the person’s heart failure care, which may form part of a long-term conditions review recall the person at least every 6 months and update the clinical record ensure that changes to the clinical record are understood and agreed by the person with heart failure and shared with the specialist heart failure MDT arrange access to specialist heart failure services if needed. [2018] Care after an acute event For recommendations on the diagnosis and management of acute heart failure see NICE’s guideline on acute heart failure. The primary care team should take over routine management of heart failure as soon as it has been stabilised and its management optimised. [2018] Writing a care plan The specialist heart failure MDT should write a summary for each person with heart failure that includes: diagnosis and aetiology medicines prescribed, monitoring of medicines, when medicines should be reviewed and any support the person needs to take the medicines functional abilities and any social care needs social circumstances, including carers’ needs. [2018] The summary should form the basis of a care plan for each person, which should include: plans for managing the person’s heart failure, including follow-up care, rehabilitation and access to social care symptoms to look out for in case of deterioration a process for any subsequent access to the specialist heart failure MDT if needed contact details for a named healthcare coordinator (usually a specialist heart failure nurse) local heart failure specialist care providers, for urgent care or review. additional sources of information for people with heart failure. [2018] Give a copy of the care plan to the person with heart failure, ther family or carer if appropriate, and all health and social care professionals involved in their care. [2018]
Findings identified in the evidence synthesis	The review findings were grouped into findings and subfindings following the structure of the recommendations on continuity of care from the NICE Guideline CG138 “Patient experience in adult NHS services” in which four aspects of continuity of care were identified: continuity, co-ordination, communication and contact/access to health professionals and services. In addition the role of specialist services was identified as an important aspect of continuity of care for people with heart failure. For each of the five subfindings of continuity, the review identified challenges, as well as suggestions for improvement, and examples of how transition in heart failure care in particular could affect patients and patient care. The guidelines “Medicines optimisation: the safe and effective use of medicines to enable the best possible outcomes” [NG5] 2015 and “Multimorbidity: clinical assessment and management” [NG56] 2016 were also identified as covering similar findings on the management of transition and continuity of patient care. Continuity: It was found that heart failure care that is spread across different healthcare settings is fragmented, and this can have a negative effect on peoples confidence in their management plans. The evidence showed that primary healthcare services, and especially general practitioners, were good at providing continuity of care and individualised, holistic care. It suggested that specialist services should, attempt to integrate into existing primary care structures. Evidence showed that discharge from HF specialist services to general practice can be seen as a ‘loss’ by patients. The reasons for this were thought to be multifactorial and concerned the difference between specialist and GP care. These included a real or perceived reduction in the focus on the condition, a reduced awareness of the intricacies of the condition, and unfamiliarity with the history of the condition and its exacerbations. In addition to this there may be an increase in the person with HF’s sense of vulnerability at being away from specialist care or his/her access to specialist care. Co-ordination: The evidence suggested that patients find the healthcare system poorly co-ordinated and confusing to navigate. This leads to increased treatment burden, and sometimes even conflicting advice. Models to improve co-ordination suggested by health care professionals (HCPs) in the studies included a single care-coordinator, a shared-care protocol between primary and secondary care, and using HF nurses as a bridge between primary and secondary care. Communication: There was a finding of poor communication across service boundaries, especially between hospital-based and primary / community services, which can lead to extra burden on patients, and under-treatment. Some of the factors HCPs felt contributed to this were a fragmented IT system and lack of time available within appointments. Patients reported that there was a paucity of information about their condition after they had been discharged from HF services. Contact/Access: The evidence suggested that people with HF can struggle to access the services they require, and referral / re-referral to HF specific-services can be inconsistent. People with heart failure felt that things may be improved if they had a primary contact, with whom they had a relationship, and most would choose their GP. It was reported that HF clinics and HF nurses can provide easy access to urgent assessment and treatment where people are able to access them. Role of specialist HF services: Expectations of HF services, such as whether they will take over “total care” of the patient (that is, including all comorbidities), and at what point they will discharge patients to primary care, differed between professionals working within, compared with those working outside of HF-services. Some professionals found that managing issues beyond HF in the HF clinic may cause an unwanted loss of continuity in primary care and take time away from HF management. Professionals in HF services felt the decision about whether to keep a patient under follow-up in the HF clinic is complex and taken on an individual basis.
Quality of the evidence	The committee acknowledged the varied confidence ratings for the different subfindings in the review, some of which were fairly low. The members weighed the confidence in the individual findings with their own experience as patients and health professionals in the NHS, and agreed that the subfindings were broadly consistent with their own experiences. Continuity: The subfinding of lack of continuity was rated as moderate confidence, due mainly to methodology concerns. The subfinding of primary care as a key player in continuity was rated as moderate confidence due to minor concerns in coherence and relevance. The committee agreed with these subfindings, but also noted that general practitioners were less able to offer continuity than historically, due to different ways of working. The subfinding about discharge from HF clinic was rated as low confidence due to methodological concerns and being limited to two papers, but the committee felt that this subfinding agreed with their experiences. The lay members explained how they built up a relationship with the HF clinic whilst they experienced significant health issues, and they felt vulnerable transitioning away from this to primary care. They were less sure about the major distinction being whether appointments were given or requested, and suggested the issue was more around the certainty of the hospital setting compared with uncertainty in, and continuity through, the GP setting. Co-ordination: The finding of poor co-ordination between services was of moderate confidence, mainly due to methodological concerns, and the finding of models to co-ordinate care was rated low confidence. Lay members explained that there were resources out there, but people with heart failure were not told about them, and they were therefore underutilised. The committee agreed that HF nurses were usefully able to provide a bridge between primary care and hospital-based services. The committee felt that HF specialist nurses would ideally co-ordinate care of HF for people who are being actively managed by specialist services, but that this could be passed to general practitioners once the patient was stabilised and on optimal treatment. Communication: The subfinding of poor communication between services was rated as moderate confidence due to methodological limitations. The subfindings of barriers to clear communication was rated as low confidence due to methodological and adequacy concerns. The professionals on the committee felt that problems with communication occurred in their workplace. It was their experience that they commonly needed to have a three-way conversation or more, and that services can require a full-time administrator in order to facilitate good communication. The committee noted that the system relied upon letters going between care settings, in the absence of a universal NHS information system. The subfinding of information after discharge from HF services being insufficient was rated as very low due to methodology and it having come from only one paper. However, the committee recognised that this was a significant issue. Contact/Access: The subfinding of difficulty accessing routine care was rated low confidence due mainly to methodological limitations. The subfinding of a primary contact person was rated as very low confidence, but was seen to overlap with other findings to a large extent. The subfinding regarding access to urgent care was rated low confidence due to concerns over methodology, relevance and adequacy. Role of specialist HF services: The committee felt that some of these subfindings were inevitable in an organisation like the NHS for a condition that crosses boundaries between services. The subfinding on the expectations on HF services was rated very low confidence due to methodological concerns and lack of depth. However the committee recognised the experience of a member who had been involved in creating a local HF pathway and experienced the same issue: that people involved in different aspects of HF care had different expectations of the service. They also recognised that GPs differ in their desire to have specialist input and facility to manage heart failure and its comorbidities. The subfinding of focus for HF services was rated as low confidence due to concerns in methodology, relevance and adequacy. The committee debated these findings. Some felt that that it was impossible for cardiologists to have the expertise and capacity to provide care for all the healthcare needs of patients with heart failure whose co-morbid conditions are complex, and therefore supported a focussed approach. Others felt that compartmentalisation of the clinical features into different distinct disorders does not help patients, and their own experience is that patients really value the attempt to provide an overview of all of their diagnoses and treatment. The subfinding on the decision to keep patients on the caseload was rated as low confidence due to concerns on relevance and adequacy of the data from only two papers. The committee felt that while it was important that decisions were made on an individual basis, the NICE guidelines should promote greater consistency and transparency in these decisions.
Trade-off between benefits and harms	The guideline committee appreciated the richness of the data in this qualitative review. They recognised that the nature of qualitative evidence offered an opportunity to identify recurring issues and examples of good practice. Continuity: The committee acknowledged that it is sometimes necessary for patients to move between different teams / services. For example, if a patient has been unwell, it is essential for the specialist HF service to follow-up and review their management and adjust any medication as necessary; but that once the patient is stable it is often appropriate to transfer routine care back to the GP. However, GPs and lay members had experience of patients receiving care by the specialist HF service without the involvement of their GP, to whom the patient was subsequently discharged. The committee agreed that an initial care plan should be developed by the HF MDT, and that this plan and subsequent contacts with the MDT should be communicated to primary care. The patient and primary care physicians should also be informed that intense management by the HF MDT is often only needed for a limited period, and will be replaced by monitoring in primary care once their condition is stabilised. The committee considered efficient and easy access back into specialist HF MDT when needed, was very important to ensure continuity of care for the patient. This could be achieved by having an “open appointment” arrangement with the specialist MDT should difficulties arise. Having primary care embedded within the MDT may facilitate this transfer. The committee felt that the role of the general practitioner in offering a longitudinal and individualised approach should be emphasised. They felt that the patient’s clinical record should be reviewed by the general practitioner at least every six months, whether or not the person is being routinely managed by HF specialist services. This should enable active management of comorbidities and early identification of other service needs, and prevent loss of contact, and thereby encourage continuity of management between primary care and heart failure services. The idea of a GP-led review is not to duplicate the work of the MDT, but to address the wider concerns of the person with CHF. This guideline already recommends at least 6 monthly clinical reviews of patients with heart failure, and GP-review of the clinical record could easily be incorporated into this existing review where patients are being managed in primary care. Even where patients are being routinely seen in the MDT, those with other long-term conditions may already be attending primary care for reviews, and the HF review could be usefully added in to existing appointments. Co-ordination: The committee agreed that coordination between services should be a priority. The committee recognised that a single care coordinator was already advocated in the multimorbidity guideline, and this would be helpful in HF. It was felt that the named healthcare coordinator would usually be the HF specialist nurse where a person was being actively managed by the multidisciplinary specialist HF team, but at the point of transfer to routine management in primary care, this would usually change to the patient’s GP. Members also advocated collaborative working between primary and secondary care through the MDT. Lay members emphasised that there needed to be better accountability, so that decisions were not only made, but also followed through and followed up effectively. Therefore future plans, who will do what, and when, as well as where/when the next routine review will take place, should form an essential part of the care plan. Communication: The committee felt that it was important that patients have access to their health information, for their own reference, and also to inform others. There was a suggestion that patients could hold their heart failure notes, but it was felt that this was not yet practical. However, it should be the case that they receive a copy of everything that is sent about them, as is already advised in CG138, including their current care plan after every formal review. The committee attempted to define what would be included in the summary of the patients status and how the care plan would be formulated, but appreciated that this would vary depending on where the patient was in the HF pathway, and other individual factors. The committee agreed that it is essential that the patient is given a named contact and knows who is going to follow them up and when, along with current medication. After contact details and follow-up arrangements, the most important things to include were thought to be the diagnosis, current treatment, red flag signs that HF is deteriorating, and signposting to further information / sources of support. Where appropriate, it would also include aetiology of HF, social circumstances, carer needs, how they can access rehabilitation and other resources. The committee acknowledged that the full summary of status and care plan may not be ready to provide to the patient during their appointment in order that they can take a copy home with them, and agreed that it may comprise a letter to the patient and GP following the appointment. Letters should use appropriate language to communicate to the GP, with a glossary or similar to make them accessible to patients. It is not intended that two sets of letters should be produced. Contact/Access: It was seen as particularly important that patients and their carers know where to get support in both routine and urgent circumstances, as per the current patient experience guideline. Pathways could be devised locally, but the committee felt that it was important that patients with heart failure could contact the heart failure clinic or service if requiring advice or urgent care. It was acknowledged that two way communication between heart failure specialists and GPs is essential to ensure patients are well managed. Specialists need to provide a clear plan to allow medication lists to be updated and GPs also have a responsibility to inform the specialist team when the condition of the patient changes. The ideal system would be to allow all members of the MDTto access hospital, community and GP notes. However this is not yet possible in most areas. Role of specialist HF services: The committee acknowledged that local dialogue about what HF services should provide for people with CHF might be useful, but may not be practical. The reports of the complexity of treating the comorbidities of patients with CHF, the nature of sub-speciality medical training, and the lack of a universal information system, highlighted the fact that the general practitioner, or another generalist such as a geriatrician, should be actively involved throughout the heart failure pathway. It was felt that HF specialists should primarily be involved in HF optimisation, while collaborating with other healthcare providers, who would be providing for the wider needs of patients.
Trade-off between net effects and costs	No published economic evaluations were identified. The committee agreed that ensuring a care plan was in place and shared with the patient and necessary healthcare professionals was important to improve current barriers in co-ordination and communication for people with heart failure. The committee noted that there may be some small costs associated with the time spent on the administrative tasks such as the writing and dissemination of the care plan. However, the committee felt that providing a care plan was already standard practice in many cases, and could be achieved by changes to current paperwork rather than additional time and resource. The committee agreed that ensuring care plans are regularly updated and shared would likely improve patient treatment by reducing cases of under or over treatment and any resultant adverse events. The committee believed that the reduced treatment burden would improve patient quality of life and reduce patients’ anxiety about their condition and treatment. Therefore, overall the committee considered that any additional administrative tasks and care plans are likely to be a cost-effective use of resources.
Other considerations	The current experience of HF care across the country was felt to be variable due to a number of factors.. These included the degree to which GPs were engaged with national programmes such as QOF and NICE Quality Standards programme; the degree to which GPs are involved in regionally agreed long-term conditions frameworks or other benchmarking; locally agreed protocols; and the involvement of the third sector, such as patient advocate groups, in facilitating access to care. The committee felt that clarifying the roles of primary care and the rest of the multidisciplinary specialist HF team would enable commissioners and others to reduce variation by defining the structures for high quality HF care, including effective co-ordination and communication. The GP members of the committee spoke about their current involvement in the care of people with HF. They felt that it was essentially the same as for any other patient with a long term condition, and approached it in the same way. However, HF was among the most challenging conditions to manage, as people with HF tend to have multiple, complex comorbidities. This led the committee to think about including aspects of the multi-morbidity guidance when thinking about reviewing and co-ordinating care for people with HF. The committee highlighted an important sub-population of HF patients, being those that are housebound. It was recognised that they are often not reviewed by HF clinics, and not seen in the GP surgery, and therefore perhaps not routinely seen by the primary healthcare team either. It was felt that some people in these circumstances only received care once they were acutely unwell. The committee intended that these patients should be included in the recommendations regarding six-monthly review. It was reported that, in one example, reviews for housebound patients with HF were sometimes delegated to district nurses. The committee felt that HF monitoring and review should be done by someone with competencies in clinical examination and medicines optimisation.

Table 137Characteristics of review question

Objective	The aim of this review is to identify the information and support needs of people with heart failure, their families and carers, when healthcare professionals are communicating a diagnosis and prognosis.
Population and setting	Patients with heart failure in a UK primary care, outpatient or community setting. Studies that relate to patient/staff experiences of communication regarding diagnosis or prognosis that occur during a patient’s hospitalisation for heart failure will be included, where the issues identified are also relevant to communication in the community/outpatient setting. Patient, family and carer information and support needs will be considered.
Context	Any description of support and information needs of patients, families or carers relating to communication of a diagnosis or the prognosis of heart failure. Views can be provided by patients, families, carers or healthcare staff
Review strategy	Synthesis of qualitative research: thematic analysis- information synthesised into main review findings. Results presented in detailed narrative format. Quality of the evidence was assessed by a GRADE CerQual approach for each review finding.

Table 138Summary of studies included in the review

Study	Design	Population	Research aim	Comments
Ald red 2004⁸	Semi-structured interviews with patients + carers. Thematic analysis.	N=10 CHF patients recently discharged from hospital and one person each that they lived with and nominated a carer. Age: mean 72 (SD 5) Severity NYHA class II-IV (6/10 NYHA class III)	Explore the impact of heart failure on the lives of older patients and their informal carers.	Part of larger study on palliative needs. Setting: UK, Barnsley.
Barnes 2006²⁶	Individual interviews with patients + carer (if they wished to do so) and focus groups with HCPs. Thematic analysis.	N=44CHF patients (NYHA class III or IV) recruited from GP practices. Age: median 77 (IQR 71-83) N=79 HCPs in 9 focus groups Age range: 27-58	Explore the attitudes of older people and primary care professionals towards communication of diagnosis, prognosis and symptoms in heart failure.	Setting: four geographical locations in the UK: East Devon, West Hampshire, Bradford and Barnsley.
Browne 2014⁵³	Individual interviews with patients + carer (if they wished to do so) and interviews or focus groups with HCPs. Thematic analysis using the Framework method. Coding framework linking data categories to an exploratory model provided by a theory known as Normalisation Process Theory.	N=30 CHF pts with advanced heart failure (NYHA class III or IV) and a history of hospital admissions. Age: mean 72 (range 60-86) N=20 Carers N=65 HCPs (14 interviews, 6 focus groups)	Examine patient, carer, and professional perspectives on current management of advanced heart failure and barriers and facilitators to improved care.	Setting: one health board in Scotland, UK.
Doos 2015¹⁰²	Individual interviews with patients + carer (if they wished to do so), taped and transcribed verbatim. Thematic analysis using the principles of Grounded Theory (constant comparison).	N=6 pts with CHF and comorbid COPD recruited at the time of hospital discharge. Age: mean 79 (range 62-91) N=5 Carers	Explore experiences of multi-morbid COPD and HF patients during, and shortly after a hospital stay. Also, to focus on patient and carer information needs on transitions and any perceived gaps in relation to their multi-morbidity.	Mixed methods study: Survey followed by interviews. Setting: a large regional hospital in England, UK.
Field 2006¹¹⁷	Open-ended narrative interviews with patients, taped and transcribed verbatim. Thematic analysis using a modified grounded theory approach, incorporating constant comparison.	N=37CHFptsatall stages of heart failure. Age range: 33-84	Examine whether heart failure patients’ awareness of the purpose and side effects of their medicines equips them to participate in informed discussions about treatments, how they cope with the condition and manage their medication.	Setting: ‘throughout the UK’
Horne 2004¹⁴⁸	Open semi-structured interviews with patients, taped and transcribed verbatim. Thematic analysis using a Grounded Theory approach.	N=20CHF patients recruited by consultants and specialist nurses from two teaching hospitals. Age: mean 73 (range 60-83) Severity NYHA class II-IV (11 class IV; 7 class III)	Explore the experiences of patients with severe heart failure and identify their needs for palliative care.	Setting: Doncaster, UK. Urban and rural communities situated in former coal mining area.
Macdonald 2016²⁰⁷	Secondary analysis of qualitative data of transcripts of interviews with patient. Use form of amplified analysis to fit themes to the Candidacy framework.	N=20CHFpts(a selection of transcripts from two previous HF studies) Age: range 56-86 Severity: 10 advanced and 10 “stable”	Contrast the help-seeking and access to care in cancer and heart disease in order to extend concepts about illness identity, and its relationship to a concept known as “Candidacy”	Compares CHF with cancer. Data was taken from the Colorectal Cancer, End-Stage HF and Stable HF studies. Setting: Scotland, UK.
Murray 2002²³¹	In-depth interviews at 3-month ly intervals for up to a year with patient. Following bereavement their informal caregivers were interviewed. Focus group of key professional carers identified by patients. Thematic analysis using the techniques of narrative analysis.	N=20 pts with end stage heart failure (NYHA class IV) identified by consultants as outpatients. Age: mean 74	Compare the illness trajectories, needs, and service use of patients with cancer and those with advanced non-malignant disease (heart failure).	Setting: Edinburgh and Livingston, Scotland, UK.
Selman 2007³⁰²; Harding 2008¹⁴¹	Semi-structured interviews with patints, their carers and HCPs. Thematic analysis using a constant comparison approach.	N=20CHFpts recruited from outpatient clinics and hospital wards. Age: mean 69 (range 43-83) Severity NYHA class III-IV (14 class III) N=11 Carers N=12 HCP from cardiology and palliative care	Selman 2007: Formulate guidance and recommendations for improving end-of-life care in chronic heart failure. To generate data on patients’ and carers’ preferences regarding future treatment modalities, and to investigate communication between staff, patients and carers on end-of-life issues. Harding 2008: Generate recommendations for the appropriate provision of feasible and acceptable information to chronic heart failure patients and their family carers, in line with UK and international policy guidelines.	Setting: London, UK.
Simmonds 2015³¹³; Glogowska 2015¹³³; Fry 2016¹²³	Qualitative study using ethnographic methods (in-depth interviews, observation, impromptu interviews, field notes, patient and carer diaries, patient medical records). Pts were followed individually throughout their interactions with healthcare for up to 11 months. In-depth interviews with a subsample of patients and/or carers. HCPs were observed delivering care for these pts. In-depth interviews with HCPs who cared for different pt. Thematic analysis involving constant comparison.	N=31 pts with severe or difficult to manage heart failure and who had an unplanned hospital admission forCHFinthe preceding 6 months. Age: mean 72 N=9 Carers N=55 HCPs overall (in-depth interviews with N=23)	Simmonds 2015: Identify critical points on heart failure patient pathways where risk of unplanned admission is increased and identify barriers to the implementation of evidence-based interventions. Glogowska 2015: Explore perceptions and experiences of health care professionals working in multi-disciplinary teams that include specialist heart failure nurses when caring for the management of heart failure patients.	Fry 2016¹²³ is linked to the dataset but did not report findings relevant to this review. Setting: three study sites, UK.
Taylor 2017³²⁸	Semi-structured interviews with patients and their carers (if they wished to do so). Thematic analysis using the Framework method.	N=16 patients with a recent (<1 year) diagnosis of heart failure recruited from heart failure clinic. Age: median 78 (range 52-87)	Explore the experiences of patients with a recent diagnosis of heart failure with a focus on symptom onset and diagnosis parts of the pathway to explore how and when patients realised something was wrong and what the term ‘heart failure’ means to them.	Setting: central England, UK.
Wingham 2015³⁷⁰	Semi-structured interviews (n=22) and one focus group with carers of CHF patients and the person they cared for (on request of the carer). Thematic analysis using the Framework method.	N=26 Carers of CHF patients who had been caregiversfor at least 6 months. Age: mean 66 (range 39-84)	Identify the needs of caregivers supporting a person with heart failure and inform the development of a caregiver resource to be used as part of a home-based self-management programme.	Setting: Cornwall, Birmingham and Leicester, UK.

: Abbreviations: Pt = people with chronic heart failure, HCP = healthcare professional, NYHA = New York Heart Association Functional Classification, COPD = chronic obstructive pulmonary disease

Table 139Review findings

Main findings	Statement of finding
Diagnosis of heart failure
1a. Communication is challenging	Clinicians find the diagnosis itself and the ensuing communication around diagnosis challenging.
1b. Timing and setting	Choosing the timing and setting of communicating the diagnosis is important.
1c. Gradual process	The persons comprehension of their diagnosis is seen as a gradual process which results in information and guidance around diagnosis being relayed to patients in a gradual manner.
1d. Terminology	As the term ‘heart failure’ is seen by healthcare professionals as anxiety-invoking, many make use of euphemisms or even more complex terminology instead which is often confusing for patients.
Understanding heart failure
2a. Understanding of diagnosis	Although there is variability in patients’ understanding of their diagnosis it is generally poor.
2b. Knowledge and management	Patients’ desire for more knowledge about their condition and its management is highly individual.
Discussion of prognosis
3a. Difficult conversations	Staff reported difficulties in discussing prognosis and future care options due to the unpredictable disease trajectory, uncertainty whether patients wanted to know and difficulties handling emotional involvement of patients and their families.
3b. Understanding of prognosis	Patients described that prognosis was rarely discussed and showed that their understanding of their prognosis was generally poor.
3c. Patients’ concerns	Whilst patients’ desire for more knowledge about their prognosis was highly individual many were worried by the uncertainty of what the future held.
Improving communication/ information flow
4a. Education and joint working	Staff identified that communication with patients and family members would be improved through training in diagnosing and prognosticating heart failure, and through mutual education for staff from different specialities.
4b. Ongoing relationship	Staff suggested for information and education to be delivered to the patient within an ongoing relationship and in an appropriate setting.
4c. Tailoring of information	The variability in patients’ ability and willingness to receive information regarding their condition requires staff to tailor the information to the needs of the individual.
4d. Improving access to information	Patients desire more detail, written information and greater access to support.

Table 140Summary of evidence

Study design and sample size		Finding	Quality assessment
Number of studies contributing to the finding	Design	Finding	Criteria	Rating	Overall assessment of confidence
Diagnosis of heart failure 1a. Communication is challenging
2²⁶^,¹³³^,³¹³	Individual interviews with pt + carers and focus groups with HCPs Ethnographic study (in-depth interviews, observation, impromptu interviews, field notes, patient and carer diaries, patient medical records) involving pt and HCPs	Clinicians find the diagnosis itself and the ensuing communication around diagnosis challenging.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	No or very minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	Minor concerns about adequacy
Diagnosis of heart failure 1b. Timing and setting
3²⁶^,¹¹⁷^,³¹³	Individual interviews with pt + carers and focus groups with HCPs Ethnographic study involving pt and HCPs Open-ended narrative interviews with pt	Choosing the timing and setting of communicating the diagnosis is important.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No concerns about adequacy
Diagnosis of heart failure 1c. Gradual process
3²⁶^,¹³³^,²⁰⁷^,³¹³	Individual interviews with pt + carers and focus groups with HCPs Secondary analysis of interview transcripts of pt. Ethnographic study involving pt and HCPs	The diagnosis itself is seen as a gradual process which results in information and guidance around diagnosis being relayed to patients in a gradual manner.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	No or very minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No concerns about adequacy
Diagnosis of heart failure 1d. Terminology
6⁸^,²⁶^,¹³³^,²⁰⁷^,²³¹^,³¹³^,³²⁸	Semi-structured interviews with pt + carers Individual interviews with pt + carers and focus groups with HCPs Secondary analysis of interview transcripts of pt. In-depth interviews with pt and following bereavement interviews with informal carers and FG with professional carers Ethnographic study involving pt and HCPs Semi-structured interviews with pt and their carers	As the term ‘heart failure’ is seen by healthcare professionals as anxiety-invoking, many make use of euphemisms or even more complex terminology instead which is often confusing for patients.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Understanding heart failure 2a. Understanding of diagnosis
9⁸^,²⁶^,⁵³^,¹¹⁷^,¹³³^,¹⁴¹^,²⁰⁷^,²³¹^,³²⁸	Semi-structured interviews with pt + carers Individual interviews with pt + carers and focus groups with HCPs Individual interviews with pt + carer and interviews or FG with HCPs Open-ended narrative interviews with pt Secondary analysis of interview transcripts of pt In-depth interviews with pt and following bereavement interviews with informal carers and FG with professional carers Semi-structured interviews with pt, their carers and HCPs Ethnographic study involving pt and HCPs Semi-structured interviews with pt and their carers	Although there is variability in patients’ understanding of their diagnosis it is generally poor.	Limitations	Serious concerns about methodological limitations	LOW
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Understanding heart failure 2b. Knowledge and management
5²⁶^,¹⁰²^,¹¹⁷^,¹⁴⁸^,³²⁸	Individual interviews with pt + carers and focus groups with HCPs Individual interviews with pt + carer Open-ended narrative interviews with pt Open semi-structured interviews with pt Semi-structured interviews with pt and their carers	Patients’ desire for more knowledge about their condition and its management is highly individual.	Limitations	Serious concerns about methodological limitations	LOW
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Discussion of prognosis 3a. Difficult conversations
6²⁶^,⁵³^,¹³³^,¹⁴¹^,²⁰⁷^,²³¹^,³⁰²	Individual interviews with pt + carers and focus groups with HCPs Individual interviews with pt + carer and interviews or FG with HCPs Secondary analysis of interview transcripts of pt In-depth interviews with pt and following bereavement interviews with informal carers and FG with professional carers Semi-structured interviews with pt, their carers and HCPs Ethnographic study involving pt and HCPs	Staff reported difficulties to discuss prognosis and future care options due to the unpredictable disease trajectory, uncertainty whether patients wanted to know and difficulties handling emotional involvement of patients and their families.	Limitations	Serious concerns about methodological limitations	MODERATE
			Coherence	No or very minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Discussion of prognosis 3b. Understanding of prognosis
5⁸^,²⁶^,⁵³^,²⁰⁷^,²³¹	Semi-structured interviews with pt + carers Individual interviews with pt + carers and focus groups with HCPs Individual interviews with pt + carer and interviews or FG with HCPs Secondary analysis of interview transcripts of pt In-depth interviews with pt and following bereavement interviews with informal carers and FG with professional carers	Patients described that prognosis was rarely discussed and showed that their understanding of their prognosis was generally poor.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Discussion of prognosis 3c. Patients’ concerns
7⁸^,²⁶^,⁵³^,¹¹⁷^,¹³³^,¹⁴⁸^,³⁷⁰	Semi-structured interviews with pt + carers Individual interviews with pt + carers and focus groups with HCPs Individual interviews with pt + carer and interviews or FG with HCPs Open-ended narrative interviews with pt Open semi-structured interviews with pt Ethnographic study involving pt and HCPs Semi-structured interviews and one FG with carers of pt and the person they cared for	Whilst patients’ desire for more knowledge about their prognosis was highly individual many were worried by the uncertainty of what the future holds.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Improving communication/ information flow 4a. Education and joint working
2²⁶^,¹⁴¹	Individual interviews with pt + carers and focus groups with HCPs Semi-structured interviews with pt, their carers and HCPs	Staff identified training needs regarding diagnosing and prognosticating heart failure and mutual education for staff of different specialities that would improve the communication with the patient and family.	Limitations	Serious concerns about methodological limitations	MODERATE
			Coherence	No or very minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Improving communication/ information flow 4b. Ongoing relationship
2²⁶^,¹³³^,³¹³	Individual interviews with pt + carers and focus groups with HCPs Ethnographic study involving pt and HCPs	Staff suggested for information and education to be delivered within an ongoing relationship with the patient in an appropriate setting.	Limitations	Moderate concerns about methodological limitations	HIGH
			Coherence	No or very minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Improving communication/ information flow 4c. Tailoring of information
7²⁶^,⁵³^,¹¹⁷^,¹³³^,¹⁴¹^,²³¹^,³²⁸	Individual interviews with pt + carers and focus groups with HCPs Individual interviews with pt + carer and interviews or FG with HCPs Open-ended narrative interviews with pt In-depth interviews with pt and following bereavement interviews with informal carers and FG with professional carers Semi-structured interviews with pt, their carers and HCPs Ethnographic study involving pt and HCPs Semi-structured interviews with pt and their carers	The variability in patients’ ability and willingness to receive information regarding their condition requires staff to tailor the information to the needs of the individual.	Limitations	Moderate concerns about methodological limitations	HIGH
			Coherence	No or very minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy
Improving communication/ information flow 4d. Improving access to information
6⁸^,²⁶^,¹⁴¹^,¹⁴⁸^,²³¹^,³⁷⁰	Semi-structured interviews with pt + carers Individual interviews with pt + carers and focus groups with HCPs Open semi-structured interviews with pt In-depth interviews with pt and following bereavement interviews with informal carers and FG with professional carers Semi-structured interviews with pt, their carers and HCPs Semi-structured interviews and one FG with carers of pt and the person they cared for	Patients desire more detail, written information and greater access to support.	Limitations	Moderate concerns about methodological limitations	MODERATE
			Coherence	Minor concerns about coherence
			Relevance	No or very minor concerns about relevance
			Adequacy	No or very minor concerns about adequacy

: Abbreviations: pt = people with chronic heart failure, HCP = healthcare professional, FG = focus group

Recommendations	When giving information to people with heart failure, follow the recommendations in the NICE guideline onpatient experience in adult NHS services. [2018] Discuss the person’s prognosis in a sensitive, open and honest manner. Be frank about the uncertainty in predicting the course of their heart failure. Revisit this discussion as the person’s condition evolves. [2018] Provide information whenever needed throughout the person’s care. [2018] Consider training in advanced communication skills for all healthcare professionals working with people who have heart failure. [2018] First consultations for people newly diagnosed with heart failure The specialist heart failure MDT should offer people newly diagnosed with heart failure an extended first consultation, followed by a second consultation, to take place within 2 weeks if possible. At each consultation: discuss the person’s diagnosis and prognosis explain heart failure terminology discuss treatments address the risk of sudden death, including any misconceptions about that risk encourage the person and their family or carers to ask any questions they have. [2018]
Findings identified in the evidence synthesis	The evidence from the review was grouped into four findings, each of which contained multiple sub-findings. These key findings are summarised below. Diagnosis of heart failure: the evidence suggested that clinicians find the diagnosis itself and the ensuing communication around diagnosis challenging. The timing of and setting in which the diagnosis is communicated is important, taking care to avoid busy environments that do not foster the sensitive relay of information. The diagnosis itself was seen as a gradual process allowing the information and guidance around diagnosis to be passed on to patients in a manageable way. The evidence suggested that the term ‘heart failure’ was regarded by healthcare professionals as anxietyinvoking, with many health care providers making use of euphemisms or even more complex terminology instead. These alternative explanations often led to a poorer understanding, confusion for people with heart failure and a lack of interest in their diagnosis as a consequence. Understanding heart failure: it was found that although there is variability in patients’ understanding of their diagnosis; it is generally poor and often misunderstood by people. Many did not connect symptoms they had to their heart failure; and showed poor knowledge of treatments, their importance, side effects and limitations. The lack of understanding was often compounded by confusion and short-term memory loss that is associated with heart failure. Patients’ desire for more knowledge about their condition and its management is highly individual. Some patients did not want to know more about their diagnosis as it would cause them to worry. Yet, others described that having received too little or conflicting information about their diagnosis caused them anxiety. Some people were proactive in seeking information themselves and knew how to manage their condition. Discussion of prognosis: staff reported difficulties in discussing prognosis and future care options due to the unpredictable disease trajectory, uncertainty about whether patients wanted to know their prognosis, and the challenge of handling the emotional response of patients and their families. Some healthcare professionals also found it challenging to balance the need to be honest with the patient about their condition (which could raise anxiety) with building trust and the desire to maintain hope and a positive outlook for patients faced with a life-limiting illness. Patients identified that prognosis was rarely discussed and the evidence showed that their understanding of their prognosis was generally poor. Patients’ desire for more knowledge about their prognosis was highly individual, whilst some preferred not to know more, many were worried by the uncertainty of what the future held. Improving communication/information flow: staff identified training and education needs regarding diagnosing and prognosticating heart failure to improve communication with the patient and family. Staff commented that single consultant appointments are too short to relay all the information a patient would need regarding diagnosis and prognosis. It was suggested that information and education be delivered within the context of a relationship of trust and continuity between the patient and the health care professional, and in an appropriate setting. Healthcare professionals identified the variability in patients’ ability and willingness to receive information about their condition requires staff to tailor the information to the needs of the individual. Many patients desire more detail, preferably in lay terms, written information and greater access to support. Family and informal carers wished to be involved in all communications.
Quality of the evidence	The confidence ratings for the different sub-findings in the review were varied, ranging from high to low. The members weighed the confidence in the individual findings with their own experiences, and agreed that they were broadly consistent. Diagnosis of heart failure: All 4 sub-findings within this finding, namely communication is challenging, timing and setting, gradual process and terminology, were rated as moderate confidence. This was mainly due to methodology concerns, as well as minor coherence limitations in a 2 findings and minor adequacy limitations in 1 sub-finding. Understanding heart failure: Both sub-findings within this finding, understanding of diagnosis and knowledge and management, were rated as low confidence having, in both instances, been downgraded by 2 increments due to concerns over methodology and coherence. Discussion of prognosis: All 3 sub-findings within this finding, namely difficult conversations, understanding of prognosis, and patients’ concerns, were rated as moderate confidence. This was mainly due to limitations in methodology, as well as minor coherence limitations for the last 2 sub-findings. Improving communication/information flow: Two sub-findings (education and joint working and improving access to information) were rated as moderate confidence having been downgraded mainly due to serious methodological limitations. The other 2 sub-findings within this finding, ongoing relationship and tailoring of information, were rated as high confidence as the moderate methodological concerns were offset by the data richness provided by the studies.
Trade-off between benefits and harms	The committee agreed that the findings and sub-findings identified in the review resonated with their own experiences as healthcare professionals or patients within heart failure services in the NHS. The committee acknowledged that some of the sub-findings identified in the review reflect general issues around communication and information needs in health care and are addressed by recommendations in the patient experience guideline CG138. The committee identified several key recommendations in that guideline which, if implemented in heart failure services, would go some way towards addressing the issues identified in the review. Highly pertinent recommendations from CG138 include ‘allow adequate time so that discussions do not feel rushed’ (1.3.4.), ‘clarify with the patient at the first point of contact whether and how they would like their partner, family members and/or carers to be involved in key decisions’ (1.3.10), ‘avoid the use of jargon and define unfamiliar words’ (1.5.6), provide ‘both oral and written information’ (1.5.12), and to ‘give the patient the opportunity to discuss their diagnosis, prognosis and treatment options’ (1.5.21). The committee agreed that there was no need to replicate these recommendations in this guideline and instead decided to cross refer to the patient experience guideline. The committee decided to focus on the issues raised that are specific to the communication of the diagnosis and prognosis of heart failure. In line with the review findings, the committee was unanimous that a single consultation was too short to explain the diagnosis to the patient and the amount of information that would have to be relayed could be overwhelming. This echoed the general recommendation in the patient experience guideline to allow adequate time so that discussions do not feel rushed (section 1.3). Building on this, the committee recommended an extended first consultation followed by a prompt (usually within 2 weeks) follow-up appointment with any member of the specialist multidisciplinary team. The committee considered a 2 week gap between the first and second appointment as reasonable and reflected usual practice whilst acknowledging this may not always be possible or appropriate in all situations. To foster continuity of care and building of the future ongoing relationship, the committee recommended that the patient would be told who they would see at their next consultation (a member of the specialist heart failure multidisciplinary team) and explained what they can expect to happen. The committee also discussed the importance of having family members/carers involved in communication from an early stage as receiving bad news in the form of the diagnosis of heart failure can be overwhelming for the patient. Carers may also be better able to take on board information provided regarding ongoing care and management than the patient who has just been given a diagnosis. Additionally, family members and carers often struggle with how best to help when the patient has conveyed information poorly or partially to them. For this reason, the inclusion of a family member or carer in one or more consultations will improve the understanding of the condition and the potential support provided to the patient. The committee acknowledged that this was covered by the patient experience guideline (particularly section 1.3) and agreed to include a cross reference to that guidance in their recommendations. It was agreed that the following points should be communicated to the patient in the first 2 appointments with the heart failure MDT: Diagnosis. Explanation of the terminology. Symptoms of heart failure and the meaning of proscriptions (for example, dietary, physical activities). Available treatments, their effects and side effects. Prognosis. Mention exacerbations and risk of sudden death. The cardiologists on the committee described that they often have to dispel misconceived ideas at the first consultation with a newly diagnosed patient. This may be the result of misunderstandings occurring when patients are referred for specialist investigation of heart failure symptoms. The committee was also aware of a lack of public awareness of heart failure as a medical condition that contributed to this. The committee did not support the use of euphemisms and suggested that health care professionals use the term ‘heart failure’ in an open and honest manner with the patient, explaining its meaning in detail, dispelling any myths or other fears induced by the language, and providing a balanced picture of the condition. The committee discussed that imagery used to describe heart failure by some healthcare professionals can be received poorly by patients. To avoid confusion for patients it was suggested that healthcare professionals could practice some phrases to explain the condition before seeing the patient. The committee acknowledged that discussion of prognosis should be a gradual process and may not necessarily be discussed in the first consultation as it is highly difficult to predict, particularly in the early stages of the condition. The prognosis needs to be discussed in a sensitive, open and honest manner, ensuring that the unpredictability of the prognosis and any uncertainty regarding life expectancy is conveyed. The positive effects of treatments available should also be discussed. The group acknowledged the difficulty patients may have coming to terms with information around their own prognosis which highlights the importance of an immediate follow-up consultation and the engagement of family/carers (where the patient consents to their presence). This consultation provides reinforcement of the key implications of the diagnosis and prognosis and, especially where a family member or carer is present, clarification of any misperceptions. To facilitate a quick rapport with the patient and to tailor information to the patient’s needs (in line with the recommendations of the patient experience guideline), the consultant committee members suggested including any preferences the patient had, regarding the level and type of information they want to be given, into the referral form from primary care. Another suggestion was for patients to complete a simple questionnaire before the first appointment identifying key areas they would like to discuss during the consultation. The committee acknowledged that the communication of these challenging issues requires specific skills that not all healthcare professionals have, and supported the provision of training in advanced communication skills to all heart failure staff. Specific information on heart failure issues could be incorporated into broader communication training for chronic or life-limiting conditions. The importance of consistency was discussed. It was agreed that the patient should receive consistent information from all MDT members. The committee recognised that this was closely linked to the separate evidence review on ‘Transition between heart failure care settings’ and the recommendations developed out of that review. The need for printed leaflets on heart failure to be available within the NHS was unanimously agreed by the committee as patients value the provision of information that they can take away with them to read and absorb in their own time. The committee also acknowledged that patients need to be directed to reliable sources of information outside the NHS – for example, British Heart Foundation, Pumping Marvellous, Cardiomyopathy UK.
Trade-off between net effects and costs	No previously published economic evaluations were identified for this question. The committee considered that overall the recommendations made above in conjunction with the patient experience guideline would improve patient satisfaction and help reduce anxiety about their condition and as a result lead to an improvement in quality of life. The majority of the recommendations made above will not incur any cost as they simply specify what information should be given to patients and the manner in which it should be delivered. However, the committee acknowledged that a couple of the recommendations could have cost implications. Firstly, an extended first consultation will incur an additional cost due to the required extra time from the clinician to fully explain the person’s diagnosis and prognosis. However, the committee considered that a standard consultation appointment is unlikely to be long enough to allow sufficient time to explain the diagnosis and prognosis as recommended in the patient experience guideline and thought this was necessary for good clinical practice. Furthermore, this type of recommendation is endorsed by the patient experience guideline which specifies that when discussing patient views and preferences that adequate time should be allowed so that discussion does not seemed rushed. In many centres, extended consultations are likely to already be occurring. The committee also acknowledged that further communication training would incur a cost. The committee were aware of training for other chronic conditions that could also be extended to heart failure specialists. Furthermore, such training is likely be undertaken in a group setting to minimise cost, or alternatively could be provided through e-learning, or at medical school in future. The committee acknowledged that the benefits of improved communication training would be advantageous across the entire treatment pathway for chronic heart failure patients as well as other conditions outside of heart failure, and therefore overall, the long-term cost effectiveness of training in effective communication is likely to be negligible.
Other considerations	The committee acknowledged ‘The Second Conversation Project’ which was currently piloting an educational intervention, to enable junior doctors to build their skills and confidence in navigating end of life discussions which they agreed would provide insight into the benefit of this type of training and may be extrapolated to the heart failure population.