Critical Appraisal of Included Publications

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Diagnostic Methods for Neuropathic Pain: A Review of Diagnostic Accuracy [Internet]. Ottawa (ON): Canadian Agency for Drugs and Technologies in Health; 2015 Apr 7.

Cover of Diagnostic Methods for Neuropathic Pain: A Review of Diagnostic Accuracy

Diagnostic Methods for Neuropathic Pain: A Review of Diagnostic Accuracy [Internet].

Show details

Contents

< Prev Next >

APPENDIX 3Critical Appraisal of Included Publications

Table A2. QUADAS-2 Ratings of Included Studies.

Figure A1

Graphical Display of QUADAS-2 Results by Proportion of Low, High, or Unclear Bias or Applicability.

1. Risk of Bias

2. Applicability

Table A3Specific Strengths and Limitations of Included Studies using QUADAS-2

Strengths	Limitations
Abdallah³⁵
Risk of Bias Consecutive sample of participants enrolled Case-control design avoided Threshold of tests was pre-specified Reference test is regarded as the clinical standard Reference assessor was blinded Index and reference tests were conducted during the same period All patients received the same reference standard Minimal losses to follow up Applicability Patient population appropriate for research question Index and reference tests specific for measuring NP	Risk of Bias Included individuals who had undergone paravertebral blocks and those who had not Single assessor completed both the reference and index tests Index test may have been completed with knowledge of reference test results Reference test was missing diagnostic component, limiting utility Assessment tests were conducted by a research assistant trained in application of the tools but with unclear clinical credentials (i.e., not conducted by pain specialist) Blinding of index assessor unclear One practitioner conducted all the assessments Applicability Patients only representative of individuals undergoing breast tumor resection
Markman³⁶
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided No inappropriate patient exclusion All patients received the same reference standard Blinded investigator conducted reference tests All patients included in the analysis Minimal loss to follow up Applicability Patients and index test appropriate for research question	Risk of Bias Patient flow unclear Patient information was retrieved retrospectively, possible recall bias Pre-specification of test thresholds was unclear Diagnostic test administrator unclear for both index and reference tests Patients only representative of individuals with failed back surgery syndrome Applicability Appropriateness of reference standard for NP unclear
Mick²⁴
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided No inappropriate patient exclusion Good generalizability of tool accuracy given chronic pain population was unselected – no exclusions for pain of assumed mixed origin Index test performed prior to reference standard Threshold for index test pre-specified (algorithm of 4 questions) Reference standard appropriate for target condition Reference test assessor blinded Same reference standard assessed in all patients referred for assessment by pain specialist Statistical corrections were made for incomplete verification and dropouts Applicability Patients and setting match those established in the review question Conduct of index and reference tests appropriate for the review question	Risk of Bias Up to 14 day lag between index and reference test Reference test not completed on all individuals who underwent index test (selected for more probable NP) Tests completed by multiple practitioners (31 for index, 3 for reference) Tests completed in different settings (primary care versus hospital) Only every 10^th patient without LNP was referred for reference test – all patients with LNP referred Method of recruitment unclear Diagnostic accuracy was a secondary outcome A proportion of referred patients did not visit the specialist and 1/10 patients with diagnosis of nLNP or nNP were allocated to specialist Not all patients were included in the analysis – filtered out to select for more patients with probable NP diagnosis – reduced generalizability to wider chronic pain population, especially those with possible mixed pain origin
Bryce²³
Risk of Bias Consecutive patients enrolled in multicentre RCT for treatment of depression with venlafaxine hydrochloride Case-control design avoided Only individuals undergoing alternate treatment for depression, with suicidal intent or plan, substance dependence, or on non-stable doses of psychoactive medications were excluded Clinical assessors were blinded to index test results Reference test appropriate for target condition Applicability: Index test representative of intended assessment	Risk of Bias Difficult to diagnose patients were excluded (only patients with 4–5 certainty rating in clinical assessment included) All patients had at least moderate depression – limited generalizability Clinical classifications were only of high confidence – those with low confidence excluded – diagnostic accuracy of SCIPI may be lower in all assessed patients Assessments conducted by multiple clinicians and research assistants Threshold was not pre-specified – thresholds to optimize sensitivity and specificity were proposed post-hoc Only highly certain reference standard assessments were considered plausible Patient flow unclear Multiple pain sites assessed independently for single patients Applicability: Patients may not be representative of all patients with spinal cord injury as only those with certain clinical diagnoses of NP were included in analysis
Perez³¹
Risk of Bias Case-control design avoided No inappropriate patient exclusions Index test assessors blind to reference test results Thresholds for index tests were pre-specified Index and reference test assessments occurred during the same visit Applicability Included patients, index and tests match review question	Risk of Bias Recruitment was ad hoc based on availability of investigator Single-center study Only some patients had diagnostic imaging or neurophysiologic tests in their medical files Applicability Reference test did not include comprehensive diagnostic tests (excluded quantitative sensory testing) resulting in lacking confidence in classification of individuals
Sadler²⁷
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided Thresholds were established a priori Reference assessor was blinded for pain service patients Index and reference tests were completed on the same day All patients were included in analysis Applicability Included patients, reference, and index test match review question	Risk of Bias Method of classification for reference did not follow most recent guidelines despite being available at time of publication Both outpatients and surgical populations were included – reason for inclusion was unclear Index test was conducted by different pain specialists depending on the patient group and site Patient flow unclear Multiple pain specialists classified NP (reference test)
Tampin³²
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided Reference standard appropriate for target condition Reference standard conducted without knowledge of index test results All assessments were completed on the same day Applicability Target condition and reference standard match research question	Risk of Bias One index test (PD-Q) performed before clinical evaluation and one performed after (LANSS) Clinical assessor was blinded to PD-Q results LANSS assessor was not blinded LANSS was assessed by the same practitioner who completed the clinical assessment (reference standard) Cut-off scores were based on ROC curve analysis – cut-offs may only be appropriate for this study population Index tests were not conducted under the same circumstances (one before clinical exam, one after) Patients had been referred by general practitioner LANSS index test conducted with knowledge of clinical findings
Gauffin²⁸
Risk of Bias A consecutive sample of patients was enrolled (targeted retrospective sample of FM patients) Case control design avoided Index test appropriate for review question Applicability Due to exclusion of patients previously diagnosed with NP patient population may have less prevalent and severe NP Reference standard and index tests were appropriate for research question	Risk of Bias Blinding of test assessors was unclear Threshold was optimized for sensitivity and specificity (although pre-specified thresholds were assessed) Self-administration of index test may introduce subjectivity Reference standard (clinical assessment) did not follow the most up to date guidelines Duration between index and reference tests were unclear Applicability Patient population excluded patients who had been previously diagnosed with NP
Rayment⁶
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided Index and reference test data collected during the same time period Applicability Index and tests were appropriate for review question	Risk of Bias Standardization of reference standard across sites was unclear Unclear differences in patient characteristics for sub-population for which PD-Q data was available Blinding of reference and index assessors was unclear Reference test was not up to date with most recent clinical guidelines Applicability Population may not represent those patients with more advanced disease who would be unfit to complete assessments (excluded those with more advanced disease)
Haroun²⁵
Risk of Bias Consecutive sample of patients recruited Case-control design avoided Test thresholds were pre-specified Same reference standard applied to all patients No loss to follow up Applicability Patients and setting appropriate for review question Index test appropriate for review question	Risk of Bias Patients with comorbidities that could contribute to NP (e.g., diabetes) excluded (severely ill patients excluded) Blinding of investigators to index and reference results unclear Interval between index and reference tests unclear Applicability Reference test did not follow most recent clinical guidelines for assessing NP
Smart²⁶
Risk of Bias Case-control design avoided Criteria for NP All patients received the same reference standard Applicability Patients appropriate for review question	Risk of Bias Convenience sample Index method determined using logistic regression to identify clusters of signs and symptoms Criteria for index test developed through modeling methods Patients with mixed or indiscriminate pain were excluded (hard to diagnose patients) The same clinician completed the reference and index test assessments Time of administration of screening unclear Reference method did not follow most recent guidelines for NP diagnosis Applicability Appropriateness of index and reference methods for NP unclear
Spallone²⁹
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided Single investigators responsible for both reference and index tests Both index and reference test investigators blinded Thresholds for index test were pre-specified Both assessments completed on the same day All patients received the same reference standard Applicability Included patients are appropriate for review question Both index and reference tests were appropriate for review question	Risk of Bias Patients with NP due to non-diabetes related causes were excluded (only representative of patients with pain related to diabetic neuropathy) Reference assessment did not follow most recent guidelines for identification of NP
Lasry-Levy³⁰
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided No loss to follow up Applicability Included patients are appropriate for review question Index test and target condition as defined by the reference test appropriate	Risk of Bias Patients who had not undergone multi-drug therapy were excluded Patients who had other potential pathologies related to NP were not excluded Blinding of assessors was unclear Reference assessment did not follow most recent guidelines for identification of NP (may not have been available at time of assessment) Interval between index and reference tests unclear Credentials of reference test administrator unclear Process of reference test unclear Threshold specification unclear
Hallstrom³³
Risk of Bias Consecutive sample of patients enrolled Case-control design avoided Cut-off thresholds were pre-specified for reference and index tests Interviewers and clinical assessors were blinded for index and reference assessments Questionnaires were administered in a random order All patients received same reference standard Minimal losses to follow up All patient assessments were done on a single day Applicability Included patients, index test, reference test were appropriate for review question	Risk of Bias Two physicians administered the reference test Limited sample size Reference assessment did not follow most recent guidelines for identification of NP (may not have been available)

: FM = fibromyalgia; LANSS = Leeds Assessment of Neuropathic Symptoms and Signs; LNP = localized neuropathic pain; nLNP = non-localized neuropathic pain; NP = neuropathic pain; PD-Q = painDETECT questionnaire; RCT = randomized controlled trial; SCIPI = spinal cord injury pain assessment

Copyright: This report contains CADTH copyright material and may contain material in which a third party owns copyright. This report may be used for the purposes of research or private study only. It may not be copied, posted on a web site, redistributed by email or stored on an electronic system without the prior written permission of CADTH or applicable copyright owner.

Links: This report may contain links to other information available on the websites of third parties on the Internet. CADTH does not have control over the content of such sites. Use of third party sites is governed by the owners’ own terms and conditions.

Except where otherwise noted, this work is distributed under the terms of a Creative Commons Attribution-NonCommercial- NoDerivatives 4.0 International licence (CC BY-NC-ND), a copy of which is available at http://creativecommons.org/licenses/by-nc-nd/4.0/

Bookshelf ID: NBK304869

Contents

< Prev Next >

PubReader
Print View
Cite this Page
Diagnostic Methods for Neuropathic Pain: A Review of Diagnostic Accuracy [Internet]. Ottawa (ON): Canadian Agency for Drugs and Technologies in Health; 2015 Apr 7. APPENDIX 3, Critical Appraisal of Included Publications.
PDF version of this title (709K)

Critical Appraisal of Included Publications - Diagnostic Methods for Neuropathic...
Critical Appraisal of Included Publications - Diagnostic Methods for Neuropathic Pain: A Review of Diagnostic Accuracy

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Bookshelf

Diagnostic Methods for Neuropathic Pain: A Review of Diagnostic Accuracy [Internet].

APPENDIX 3Critical Appraisal of Included Publications

Table A2

Figure A1

1. Risk of Bias

2. Applicability

Table A3Specific Strengths and Limitations of Included Studies using QUADAS-2

Views

In this Page

Other titles in this collection

Recent Activity