Neuropathic pain screening questionnaires have limited measurement properties. A systematic review

Accepted Manuscript Neuropathic pain screening questionnaires have limited measurement properties. A systematic review Stephanie Mathieson, Christopher G. Maher, Caroline B. Terwee, Tarcisio Folly de Campos, Chung-Wei Christine Lin PII:

S0895-4356(15)00155-9

DOI:

10.1016/j.jclinepi.2015.03.010

Reference:

JCE 8849

To appear in:

Journal of Clinical Epidemiology

Received Date: 30 May 2014 Revised Date:

3 February 2015

Accepted Date: 18 March 2015

Please cite this article as: Mathieson S, Maher CG, Terwee CB, Folly de Campos T, Lin C-WC, Neuropathic pain screening questionnaires have limited measurement properties. A systematic review, Journal of Clinical Epidemiology (2015), doi: 10.1016/j.jclinepi.2015.03.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Stephanie Mathieson a

([email protected])

Christopher G Maher a

([email protected])

Caroline B Terwee b

([email protected])

Tarcisio Folly de Camposa

([email protected])

Chung-Wei Christine Lin a

([email protected])

RI PT

ACCEPTEDhave MANUSCRIPT Neuropathic pain screening questionnaires limited measurement properties. A systematic review

SC

a. The George Institute for Global Health, Sydney Medical School, University of Sydney, Sydney, NSW, 2000, Australia.

M AN U

b. Department of Epidemiology and Biostatistics and the EMGO+ Institute for Health and Care Research, VU University Medical Center, Amsterdam, The Netherlands.

Contact details

Stephanie Mathieson

TE D

Corresponding author:

The George Institute for Global Health Level 13, 321 Kent St,

AC C

Ph +612 9657 0412

EP

Sydney, NSW 2000, Australia.

Fx +612 9657 0301

Email: [email protected]

Number of figures: 1

Number of tables: 2

Web appendix: A-F

1

ACCEPTED MANUSCRIPT

Abstract

Objective: The Douleur Neuropathique (DN4), ID Pain, Leeds Assessment of Neuropathic Symptoms and Signs (LANSS), PainDETECT and Neuropathic Pain Questionnaire have been recommended as screening questionnaires for neuropathic pain. This systematic review aimed to evaluate the measurement properties (e.g. criterion validity and reliability) of these questionnaires.

RI PT

Study Design and Setting: Online database searches were conducted and two independent reviewers screened studies and extracted data. Methodological quality of included studies and the measurement properties were assessed against established criteria. A modified Grading of

SC

Recommendations Assessment, Development and Evaluation (GRADE) approach was used to summarise the level of evidence.

M AN U

Results: 37 studies were included. Most studies recruited participants from pain clinics. The original version of the DN4 (French) and Neuropathic Pain Questionnaire (English) had the most number of satisfactory measurement properties. The ID Pain (English) demonstrated satisfactory hypothesis testing and reliability but all other properties tested were unsatisfactory. The LANSS (English) was unsatisfactory for all properties, except specificity. The PainDETECT (English)

TE D

demonstrated satisfactory hypothesis testing and criterion validity. In general, the cross-cultural adaptations had less evidence than the original versions. Conclusion: Overall, the DN4 and Neuropathic Pain Questionnaire were most suitable for clinical

EP

use. These screening questionnaires should not replace a thorough clinical assessment.

AC C

Keywords: Neuropathic pain; Questionnaire. Word Count: 202

Running title: A systematic review of the measurement properties of neuropathic pain screening questionnaires.

What’s new? •

Key findings: The original language versions of the DN4 (French) and Neuropathic Pain Questionnaire (English) had the most evidence for their measurement properties.

•

Overall the language versions had less evidence than the original questionnaires. 2

•

MANUSCRIPT What this adds to whatACCEPTED is known: None of the screening questionnaires for neuropathic pain had all measurement properties assessed and the overall level of evidence was often low to very low.

•

What is the implication, what should change now: Clinicians and researchers should not to solely rely on these screening questionnaires to identify neuropathic pain as they

RI PT

cannot replace clinical assessment.

1. Introduction

SC

Neuropathic pain arises from a direct lesion or disease affecting the somatosensory system [1]. It has no single diagnostic feature and is often described as shooting, burning or tingling [2].

M AN U

This reflects the various maladaptive neural mechanisms and physiological responses of nerve pathway damage in the central or peripheral nervous system. Neuropathic pain encompasses a wide scope of clinical disorders, such as multiple sclerosis, painful peripheral neuropathy, trigeminal neuralgia and spinal radiculopathy [3] and depending on the setting, the prevalence rates vary. Neuropathic pain characteristics in chronic pain patients ranges between 3 to 17%

be as high as 51.9% [4].

TE D

in the general population [3], while prevalence rates within patients attending pain clinics can

EP

The clinical diagnosis of neuropathic pain should use a sequential approach of a thorough clinical history, clinical examination including a neurological examination and detailed

AC C

sensory testing to identify any relevant sensory alterations, plus if available, appropriate diagnostic tests (e.g. electromyography, MRI or laboratory findings) to confirm the presence of a disease or lesion of the somatosensory system [5]. In addition, methods such as quantitative sensory testing can be used to assess and quantify altered sensory nerve function, but these have not gained widespread acceptance in clinical practice because of the lack of unified standards [6]. Full examination for neuropathic pain can be time consuming, so questionnaires that screen for neuropathic pain have been developed [2].

3

ACCEPTED MANUSCRIPT The Neuropathic Pain Special Interest Group (NeuPSIG) of the International Association for the Study of Pain (IASP) has recommended five questionnaires to screen for neuropathic pain [5]; these are the Douleur Neuropathique 4 (DN4) [7], ID Pain [8], the Leeds Assessment of Neuropathic Symptoms and Signs (LANSS) [9], the PainDETECT questionnaire [10] and the Neuropathic Pain Questionnaire [11]. These questionnaires have been developed [7-11], modified

RI PT

[e.g.12] and culturally adapted to other languages [e.g.13] and used in various neuropathic pain disorders, with individual studies testing their measurement properties. Critically appraising the measurement properties of these questionnaires may guide clinicians and researchers on the

SC

relevance and clinical value of these questionnaires. However, to date there are no systematic critical appraisals of these measurement properties. The aim of this systematic review was to

M AN U

evaluate the measurement properties of these questionnaires.

2. Methods 2.1 Search strategy

This systematic review was registered with International Prospective Register of Systematic

TE D

Reviews (PROSPERO) in April 2013 (CRD42013004449). Electronic database searches were conducted from inception to 3rd May 2013 on Medline, EMBASE and CINAHL (SM). Published search filters of measurement properties [14] were used and combined with the name and

EP

abbreviations of each neuropathic pain screening questionnaires. Grey literature search was conducted by reviewing references lists of included studies plus abstracts to identify any additional

AC C

studies (SM). An example search strategy is provided in appendix A.

2.2 Study selection criteria Eligible studies assessed any measurement property (e.g. validity) of the five recommended neuropathic pain screening questionnaires including those modified, shortened, translated or cross culturally adapted from the original language version. There was no restriction in study population, as long as the study used the questionnaire to screen for neuropathic pain in that population, study design or publication status. Only studies published in English were included.

4

ACCEPTED MANUSCRIPT

2.3 Screening

All articles were independently screened by two authors (SM, TF), first by title and abstract and then by full text. Differences were resolved by discussion and if necessary arbitrated by an independent third reviewer (CL). If a screened abstract was deemed eligible and the full text was

RI PT

not available, study authors were contacted for publication details or relevant study information.

2.4 Data extraction

Data extraction was conducted by two independent reviewers (SM, TF, CL) on piloted electronic

SC

forms. Differences were resolved by discussion and included a third reviewer (CL, CT) when necessary. The following data was extracted: study setting, participant’s characteristics (including

M AN U

age, gender, pain intensity, other important disease characteristics), neuropathic pain assessment methods, distribution of scores and measurement property outcomes.

2.5 Assessment of measurement properties

Measurement properties were assessed according to definitions by Mokkink (2010) [15] and

TE D

criteria by Terwee (2007) [16] as satisfactory, unclear, unsatisfactory or no information available (appendix B). Two reviewers independently assessed each property, with disagreements resolved by discussion or by a third reviewer. The measurement properties were content validity (the

EP

degree which the content of the questionnaire adequately reflection of the construct to be measured and includes face validity); criterion validity (how well the questionnaire reflects the

AC C

gold standard); construct validity (the degree the questionnaire scores are consistent with hypotheses of the concept being measured); reliability (the extent to which the questionnaire is free from error, measuring people who have not changed for repeated measures under the same conditions including test-retest, intra-rater and inter-rater reliability); internal consistency (the extent the items in the questionnaire are correlated, hence measuring a single concept); responsiveness (the ability of the questionnaire to detect changes over time in the construct being measured) and interpretability (the degree to which one can assign a qualitative meaning to quantitative scores or changes in scores). After consultation with an expert in clinimetrics (CT), assessment of sensitivity and specificity was included in criterion validity; and convergent and 5

ACCEPTED divergent validity included as part of constructMANUSCRIPT validity.

Methodological quality of included studies was assessed using Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) [17] for criterion validity (i.e. sensitivity and specificity), rated as high, low or unclear risk of bias and the COSMIN checklist for all other properties, rated as

property was the lowest score for that measurement property.

SC

2.6 Results synthesis

RI PT

excellent, good, fair or poor [18]. The overall methodological quality rating for each measurement

All results of study characteristics, measurement properties and methodological quality were

M AN U

electronically tabulated. Data were grouped by questionnaire and presented descriptively. In an amendment to the protocol, the level of evidence supporting each measurement property was assessed by modifying the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria [19] (appendix C). The level of evidence supporting all measurement properties started as high and could be downgraded to moderate, low or very low evidence. We

TE D

operationalized the GRADE factors for this review; inconsistency (where level of evidence was downgraded one point if <75% of the measurement property results across studies were consistent), risk of bias (downgraded one point if the majority of COSMIN scores across studies

EP

were “fair” or downgraded two points if the majority of COSMIN scores across studies were “poor” or there was no majority of scores. For criterion validity, downgraded one point for an overall

AC C

unclear QUADAS-2 score or downgraded two points for a overall high QUADAS-2 score) and imprecision (downgraded one point if <400 participants across studies). We considered that a minimum of 400 participants per measurement property and questionnaire language version was a suitable sample size. This was determined by using a statistical calculator [20] to establish a value to have enough precision across each measurement property while maintaining a reasonable and practical sample size.

6

ACCEPTED MANUSCRIPT

3. Results 3.1 Search results

A total of 314 unique records were identified through searches, of which 37 studies were eligible for this review (figure 1).

RI PT

3.2 Study Characteristics Studies recruited participants with a variety of pain conditions (n=24), for example post herpetic neuralgia, trigeminal neuralgia and radiculopathy, or with a specific pain condition (spinal pain

SC

(n=3), leprosy (n=2), cancer (n=3), peripheral diseases (n=2), fibromyalgia (n=1), spinal cord injury (n=1) and knee osteoarthritis (n=1)). The study characteristics are shown in appendix D.

M AN U

Most studies administered the questionnaires face to face in a clinical setting, most commonly by a pain physician or specialist [7-10,12,29-32,34,35,37,40,42,43,45,48] or neurologist [7,21,24,29,31,32,41] (table 1). Five studies either mailed or collected the questionnaire over the phone [33,43,44,47,49]. Twenty-six studies used an adequate method to diagnose neuropathic pain (i,e. clinical history and examination that included sensory testing) [7-10,12,21-23,26-

TE D

32,34,37-43,45,47,48] of which sixteen studies included additional testing, such as laboratory tests and imaging when indicated [9,10,12,21,23,28-30,32,34,38-42,45]. Five studies used 2 clinicians to diagnose neuropathic pain [7,10,24,35,45], however only three studies used adequate methods

EP

[7,10,45]. Eleven studies used inadequate methods to diagnose neuropathic pain, such as medical chart review, self reported pain or the method was not reported

AC C

[11,13,24,25,33,35,36,44,46,49,50].

3.3 Measurement properties The summary of findings for each measurement property, per questionnaire, incorporating individual measurement property results and the level of evidence are reported in table 2. No studies assessed all measurement properties. Criterion validity, which includes sensitivity and specificity, was the most frequently reported property with individual scores summarised in table 1. Reported sensitivity ranged from 22% [38] to 100% [22,29] and specificity from 42% [22] to 100% [26,31,37]. Individual measurement property results and methodological quality are reported in 7

ACCEPTED MANUSCRIPT

appendix E and F.

3.3.1 DN4 The DN4 was originally developed in French in France [7]. The French version [7,21] demonstrated very low-level evidence for satisfactory criterion validity and low level evidence for

RI PT

satisfactory hypothesis testing (part of construct validity) and reliability (table 2). Content validity and structural validity (another aspect of construct validity) was unsatisfactory with very low level and low level of evidence respectively. These properties were evaluated in people with a variety of

SC

chronic pain conditions [7] and specifically in low back pain [21].

M AN U

The DN4 has since been cross-culturally adapted or translated into Amharic, Arabic, Dutch, English, Hindi, Marathi, Italian, Portuguese, Spanish, Swedish and Turkish [22-32] and demonstrated unsatisfactory cross cultural validity with low level evidence [23,30-32] in all but one study [29]. The non-French language versions had satisfactory criterion validity (sensitivity and specificity), except specificity in Amharic [22] and majority of studies had low level of evidence

TE D

supporting this. The level of evidence for construct validity, internal consistency and reliability was low or very low with a mixture of scores. The Portuguese version [29] had the most satisfactory evidence out of all the DN4 non-French language versions. Content validity was not assessed in

EP

any of the non-French language versions (table 2). These properties were evaluated in people with a variety of pain [32] or chronic pain [23,24,29,30] conditions and specifically in leprosy

AC C

[22,26], low back-leg pain [25], peripheral nerve disease [28], painful diabetic polyneuropathy [27] and spinal cord injury [31].

3.3.2 ID Pain

The ID Pain questionnaire was originally developed in English in the United States [8]. The English ID Pain [8,33] had moderate level evidence for satisfactory hypothesis testing and unsatisfactory structural and content validity. There was low level evidence for satisfactory reliability and very low level evidence for unsatisfactory criterion validity (table 2). These properties were evaluated in people with a variety of chronic pain conditions [8] and specifically in breast cancer survivors with 8

possible neuropathic pain [33]. ACCEPTED

MANUSCRIPT

The ID Pain had several non-English language versions: Italian, Chinese and Thai [28,34,35,36]; the latter two language versions were unsatisfactory for cross-cultural validity supported by very low level evidence. Criterion validity was satisfactory for all language versions [28,34,36] except

RI PT

specificity in Chinese [35] with the same level of evidence as the original English version. The language versions demonstrated very low level evidence for unsatisfactory internal consistency and reliability when assessed. Hypothesis testing, only assessed in the Italian language version

SC

[28], had similar results to the original but with less evidence. Content validity was not assessed in any non-English language versions (table 2). These properties were evaluated in people with a

[28].

3.3.3 LANSS and S-LANSS

M AN U

variety of pain [34,36] or chronic pain [35] conditions and specifically in peripheral nerve disease

The LANSS was originally developed in English in the United Kingdom [9]. The English version of

TE D

the LANSS [9,37,38] was unsatisfactory for all properties, except specificity, supported by very low level evidence (table 2). These properties were evaluated in people with a variety of chronic pain conditions [9] and specifically in head and neck cancer pain [37] and neck and upper limb pain

EP

[38].

AC C

The non-English language versions of LANSS in Amharic, Chinese, Italian, Portuguese, Swedish and Turkish [35,39,40,41,31,42], all demonstrated unsatisfactory cross-cultural validity when assessed [31,35,40-42]. Criterion validity was satisfactory for most non-English language versions [35,40-42], with very low level evidence. The Chinese [35] and Portuguese versions [40,41] overall had unsatisfactory internal consistency with very low evidence. Reliability was assessed in the Portuguese [40,41] and Swedish [31] versions, both with satisfactory low level evidence. Hypothesis testing, structural validity and content validity were not assessed in any of the non-English LANSS versions (table 2). These properties were evaluated in people with a variety of pain [42] or chronic pain [35,40,41] conditions and specifically in chronic cancer pain [39], 9

ACCEPTED spinal cord injury [31] and leprosy [22].

MANUSCRIPT

The authors of the LANSS have modified the questionnaire to be used as a self-reported questionnaire, titled S-LANSS [12]. The English version [12,25,43] had moderate evidence supporting satisfactory hypothesis testing and low level evidence for satisfactory criterion validity

RI PT

and unsatisfactory internal consistency and reliability. The overall level of evidence was less for the phone and mail formats in English [25] (table 2). These properties were evaluated in people

SC

with a variety of chronic pain conditions [12,43] and specifically in low back-leg pain [25].

The S-LANSS had two non-English language versions, Arabic and Turkish [44,45], both

M AN U

demonstrating low level evidence for unsatisfactory cross-cultural validity. The Arabic version [44] had very low level evidence for unsatisfactory internal consistency and satisfactory reliability. The Turkish version [45] had satisfactory criterion validity but with much stronger evidence than the original English version and similar evidence for the other properties. Content validity was not assessed in any of the non-English language versions (table 2). These properties were evaluated

TE D

in people with a variety of chronic pain conditions [44,45].

3.3.4 PainDETECT and m-PainDETECT

EP

The PainDETECT questionnaire was designed in Germany specifically for use in low back pain [10]. There was very low level evidence for satisfactory criterion validity and internal consistency

AC C

with content validity determined as unsatisfactory (also very low level evidence) (table 2).

The PainDETECT had been translated or cross culturally adapted to Dutch, English, Finnish, Japanese, Spanish and Swedish [46,38,47,48,13,31]. Cross cultural validity was unsatisfactory in Dutch, Japanese, Spanish and Swedish all with low-level evidence. The English version [38] was unsatisfactory in criterion validity and hypothesis testing. The Finnish version [47] reported criterion validity only, in which low level evidence for satisfactory sensitivity and unsatisfactory specificity was shown. The Japanese [48] and Spanish versions [13] had a mix of scores with low level of evidence or less. The Swedish version [31] had satisfactory specificity and unsatisfactory 10

ACCEPTED MANUSCRIPT reliability, both with low level evidence. Content validity was not assessed in any of the nonGerman language versions (table 2). These properties were evaluated in people with a variety of pain [46,48] or chronic pain [13] conditions and specifically in fibromyalgia [47], spinal cord injury [31,48] and neck and upper limb pain [38].

RI PT

One Canadian study modified the PainDETECT to the m-PainDETECT questionnaire by improving the questions specifically to suit for the knee osteoarthritis population, while maintaining the scoring scheme [49]. Construct validity was the only measurement property assessed and was

SC

determined to be satisfactory based on low level evidence (table 2).

M AN U

3.3.5 Neuropathic Pain Questionnaire and Neuropathic Pain Questionnaire-short version The Neuropathic Pain Questionnaire was originally developed in English in the United States [11]. It demonstrated moderate level evidence for satisfactory hypothesis testing and internal consistency, low level evidence for satisfactory structural validity, unsatisfactory reliability and content validity, while there was very low level evidence for satisfactory criterion validity (table 2).

TE D

These properties were evaluated in people with a variety of chronic pain conditions [11].

The Neuropathic Pain Questionnaire had a Chinese, Italian and Swedish language version

EP

[35,39,31] that had less evidence than the English version. The modified version, Neuropathic Pain Questionnaire–short form [50] was found to have satisfactory hypothesis testing, mixed

AC C

criterion validity, both with a very low level of evidence. The only language version, the Italian version [39], only assessed criterion validity, which was found to be unsatisfactory with very lowlevel evidence [39] (table 2). These properties were evaluated in people with a variety of chronic pain conditions [35,50] and specifically in chronic cancer pain [39] and spinal cord injury [31].

4. Discussion This review systematically appraised the measurement properties of the DN4, ID Pain, LANSS, PainDETECT, and Neuropathic Pain Questionnaire. Of the 37 studies, many measurement properties frequently met the assessment criteria satisfactorily, but the methodological quality was 11

ACCEPTED MANUSCRIPT often fair or poor which had a major influence on the overall level of evidence determined for each questionnaire and language version. The original French DN4 version and original English Neuropathic Pain Questionnaire had the most positive supporting evidence. Overall the language versions had less evidence than the original questionnaires. Considering cross-cultural validity was poor across all studies, the language versions should be used with caution. This is the first

RI PT

time a comprehensive comparison has evaluated the measurement properties of these five screening neuropathic pain screening questionnaires. This novel contribution to research advances clinical knowledge but means literature comparisons are limited. An additional first is

SC

the use of the COSMIN checklist in this field of health related tools. Clinically our results provide important information to non-pain specialists that the presence of neuropathic pain is best

M AN U

diagnosed using a stepwise approach and diagnostic tests if available [1] rather than using a screening questionnaire. While the validity and reliability of these screening tools varies, the role of these questionnaires remains as an indicative tool.

The conclusions drawn in this review are based on the overall evidence for each questionnaire

TE D

and language version across different pain populations. For example, the neuropathic pain subgroups included a range of conditions like spinal cord injury [7,23,30,31,36,48], multiple sclerosis [7,23,29,30], post herpetic neuralgia [7-10,12,13,23,29,30,32,34,35,41,42,45] and

EP

diabetic neuropathy [8,10,12,13,23,27,28,32,34,35,41,42,45,48] while non-neuropathic groups may have included osteoarthritis [7,10,13,23,24,30,32,35,41,42,45,48,49], an arthropathies

AC C

[7,9,10,12,13,23,27,34,36,41,45] and mechanical low back pain [7-10,12,13,21,23-25,30,32,41]. The diagnosis of neuropathic pain in different pain populations can vary, based on differences in clinician diagnosis and the availability of diagnostic testing. Mixed pain conditions, where both neuropathic and nociceptive components are present, can be a diagnostic challenge. Many of our studies excluded participants with mixed pain conditions [7,9,10,29,34,35,40,42,45,48], which may have influenced the sensitivity and specificity of the questionnaires, and hence limit generalisability when applying our results to clinical practice. Of the studies that did include mixed pain in their analysis [8,12,13,23,25,30,32,36,38,41,43,44] there was no consistent trend in increased or decreased sensitivity and/or specificity when compared to the original version. 12

ACCEPTED Several questionnaires were developed beforeMANUSCRIPT the redefinition of neuropathic pain in 2008 [1]. This did not affect our assessment of the criterion validity of these questionnaires, as the methods that were recommended as the reference standard to diagnose neuropathic pain remained similar. Many studies used adequate diagnostic methods, of which three studies used two clinicians to diagnose neuropathic pain [7,10,45]. This quality provides confidence when assessing criterion

RI PT

validity. Since 2008, a diagnostics grading system has been proposed to indicate the likelihood of neuropathic pan as either no, possible, probably or definite [1]. Five of the more recent studies have adopted this grading criteria into their research [21,32,38,39,47]. Thirty percent of studies

SC

used inadequate diagnostic methods [11,13,24,25,33,35,36,44,46,49,50] such as self reported diagnosis of neuropathic pain [33,44] or not reporting the diagnostic methods [46] while other

M AN U

studies administered the questionnaire(s) over the phone or by mail [33,43,44,47,49] with several months delay between administering the questionnaire and clinical examination. Some studies omitted to report the time between the initial clinical assessment and administration of the questionnaire(s) [9-13,22,23,26-29,31-37,40-44,46-50]. In these cases patients may have changed over time, thus reducing the methodological quality and the robustness of the results on

TE D

criterion validity.

The overall level of evidence of the included studies was low or very low due to the

EP

methodological quality and the small sample sizes. Future studies should aim to improve the study design, analysis and reporting of measurement properties in neuropathic pain. Most

AC C

measurement properties lacked major constituents, for example clearly defined domains in content validity and the lack of factor analysis in internal consistency. Since the publication of many of the included studies, the knowledge and assessment of measurement properties has advanced. With this in mind, and our results, other areas of improvement for future study designs and reporting can be achieved with some simple changes. For example, methodological quality could be improved by reporting the percentage of missing items and how they were handled. This was frequently not reported and a high number of missing items may indicate bias in the results particularly if the data missing was not at random. Defining hypotheses a priori for assessing construct validity, their direction, magnitude and factor analysis should also be performed. Studies 13

ACCEPTED MANUSCRIPT pertaining to cross-cultural validity should include multiple independent forward and backwards translations, pre-testing, developmental committees and include multiple group factor analysis. These elements were frequently not described, incomplete or omitted. Future research in neuropathic pain could re-evaluate these screening questionnaires using the new definition of neuropathic pain, promote the grading system, which indicates the likelihood of neuropathic pain

RI PT

[1] and encourage consistent assessment of measurement properties.

5. Conclusion

SC

This systematic review evaluated the measurement properties of IASP recommended neuropathic pain screening questionnaires and their modifications. Measurement properties indicate the

M AN U

clinical value of clinical tools such a questionnaires. The original DN4 and Neuropathic Pain Questionnaire had the most evidence for their measurement properties, particularly in patients with chronic pain. However, none of the questionnaires had all the measurement properties assessed. For the measurement properties that were assessed, not all of them were satisfactory and most of the findings were supported by low or very low level of evidence. While these

TE D

screening questionnaires may provide an indication of the presence of neuropathic pain, they cannot replace a clinical assessment.

EP

Conflicts of interest

AC C

The authors report no conflicts of interest.

Acknowledgements

SM has an Australian National Health and Medical Research Council (NHMRC) postgraduate scholarship. CM has an Australian Research Council Future Fellowship. CL has a NHMRC Career Development Fellowship.

Author contributions SM, CM and CL conceived the review and drafted the protocol. CT contributed clinimetric and COMSIN expertise. SM conducted the search. SM, TF and CL screened for eligibility and 14

ACCEPTED MANUSCRIPT extracted data. All authors contributed to the analysis of results. SM drafted the manuscript. All authors contributed to the manuscript and approved the final version for publication.

References Treede RD, Jensen TS, Campbell JN, Cruccu G, Dostrovsky JO, Griffin JW,

1.

RI PT

Hansson P, Hughes R, Nurmikko T, Serra J. Neuropathic pain; Redefinition and a grading system for clinical and research purposes. Neurology 2008;70:1630-1635.

Haanpaa M, Treede RD. Diagnosis and classification of neuropathic pain:

SC

2.

3.

M AN U

epidemiology and impact of neuropathic pain. Pain Clinical Updates 2010;XVIII:1-6. Van Hecke O, Austin SK, Khan RA, Smith BH, Torrance N. Neuropathic pain in the general population: a systematic review of epidemiological studies. Pain 2014;155:654-662. 4.

Perez C, Ribera MV, Galvez R, Mico JA, Barutell C, Failde I, Sanchez-Magro I,

TE D

Stern A. High prevalence of confirmed, but also of potential and believed, neuropathic pain in pain clinics. Eur J Pain 2013;17:347-356. 5.

Haanpaa M, Attal N, Backonja M, Baron R, Bennett M, Bouhassira D, Cruccu G,

EP

Hansson P, Haythornthwaite JA, Iannetti GD, Jensen TS, Kauppila T, Nurmikko TJ, Rice AS, Rowbotham M, Serra J, Sommer C, Smith BH, Treede RD.

AC C

NeuPSIG guidelines on neuropathic pain assessment. Pain 2011;152:14-27.

6.

Backonja M, Attal N, Baron R, Bouhassira D, Drangholt M, Dyck PJ, Edwards RR, Freeman R, Gracely R, Haanpaa MH, Hannsson P, Hatem SM, Krumova EK, Jensen TS, Maier C, Mick G, Rice AS, Rolke R, Treede RD, Serra J, Toelle TR, Tugnoli V, Walk D, Walalce M, Ware M, Yarnitsky D, Ziegler D. Value of quantitative sensory testing in neurological and pain disorders: NeuPSIG consensus. Pain 2013;154:1807-1819.

7.

Bouhassira D, Attal N, Alchaar H, Boureau F, Brochet B, Bruxelle J, Cunin G, 15

MANUSCRIPT Fermanian J, GiniesACCEPTED P, Grun-Overdyking A, Jafari-Schluep H, Lanteri-Minet M, Laurent B, Mick G, Serrie A, Valade D, Vicaut E. Comparison of pain syndromes associated with nervous or somatic lesions and development of a new neuropathic pain diagnostic questionnaire (DN4). Pain 2005;114:29-36. Portenoy R. Development and testing of a neuropathic pain screening questionnaire: ID

8.

RI PT

Pain. Curr Med Res Opin 2006;22:1555-1565.

Bennett M. LANSS Pain Scale: The Leeds Assessment of Neuropathic Symptoms

9.

and Signs. Pain 2001;92:147-157.

Freynhagen R, Baron R, Gockel U, Tolle TR. PainDETECT: a new screening

SC

10.

questionnaire to identify neuropathic components in patients with back pain. Curr

11.

Krause SJ, Backonja MM. Development of a Neuropathic Pain Questionnaire. Clin J Pain 2003;19:306-314.

12.

M AN U

Med Res Opin 2006;22:1911-1920.

Bennett M, Smith BH, Torrance N, Potter J. The S-LANSS score for identifying pain of

TE D

predominantly neuropathic origin: validation for use in clinical and postal research. J Pain 2005;6:149-158. 13.

De Andrés J, Pérez-Cajaraville J, Lopez-Alarcón MD, López-Millán JM, Margarit

EP

C, Rodrigo-Royo MD, Franco-Gay ML, Abejón D, Ruiz MA, López-Gomez V,

AC C

Pérez M. Cultural adaptation and validation of the PainDETECT scale into Spanish. Clin J Pain 2012;28:243-253.

14.

Terwee CB, Jansma EP, Riphagen II, de Vet HC. Development of a methodological PubMed search filter for finding studies on measurement properties of measurement instruments. Qual Life Res 2009;18:1115-23.

15.

Mokkink LB, Terwee CB, Patrick DL, Alonso J, Stratford PW, Knol DL, Bouter LM, de Vet HCW. The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health-related patientreported outcomes. J Clin Epidemiol 2010;63:737-45. 16

16.

MANUSCRIPT Terwee CB, Bot SD, ACCEPTED de Boer MR, van der Windt DA, Knol DL, Dekker J, Bouter, LM, de Vet, HCW. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol 2007;60:34-42.

17.

Whiting PF, Rutjes AWS, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, Leeflang MMG, Sterne JAC, Bossuyt PMM, QUADAS-2 Group. QUADAS-2: a revised tool for

RI PT

the quality assessment of diagnostics accuracy studies. Ann Intern Med 2011;155:529-536. 18.

Terwee CB, Mokkink LB, Knol DL, Ostelo RWJG, Bouter LM, de Vet HCW. Rating

scoring system for the COSMIN checklist. Qual Life Res

2012;21:651-657. 19.

M AN U

properties: a

SC

the methodological quality in systematic reviews of studies on measurement

Guyatt GH, Oxman AD, Akl EA, Kunz R, Vist GE, Brozek J, Norris S, Falck-Ytter Y, Glaser K, de Beer H, Jaeschke H, Rind D, Meerpohl J, Dahm P, Schunemann HJ. GRADE Guidelines: 1. Introduction GRADE evidence profiles and summary of

20.

TE D

findings tables. J Clin Epidemiol 2011;64:383-394.

Statistical Solutions, Power and sample size calculator, Statistical Solutions LLC, http://www.statisticalsolutions.net/pss_calc.php, accessed 9/12/13. Attal N, Perrot S, Fermanian J, Bouhassira D. The neuropathic components of

EP

21.

chronic low back pain: a prospective multicenter study using the DN4

AC C

questionnaire. J Pain 2011;12:1080-1087.

22.

Haroun OM, Hietaharju A, Bizuneh E, Tesfaye F, Brandsma JW, Haanpaa M, Rice AS, Lockwood DN. Investigation of neuropathic pain in treated leprosy patients in Ethiopia: a cross-sectional study. Pain 2012;153:1620-1624.

23.

Harifi G, Ouilki I, El Bouchti I, Ait Ouazar M, Belkhou A, Younsi R, Amine M, Tazi I, Abouqal R, Niamane R, El Hassani S. Validity and reliability of the Arabic adapted version of the DN4 questionnaire (Douleur Neuropathique 4 questions) for differential diagnosis of pain syndromes with a neuropathic or somatic component. Pain Pract

17

2011;11:139-147. 24.

ACCEPTED MANUSCRIPT

van Seventer R, Vos C, Giezeman M, Meerding WJ, Arnould B, Regnault A, van Eerd M, Martin C, Huygen F. Validation of the Dutch version of the DN4 diagnostic questionnaire for neuropathic pain. Pain Pract 2012;13:390-398. Walsh J, Rabey MI, Hall TM. Agreement and correlation between the Self-report

RI PT

25.

Leeds Assessment of Neuropathic Symptoms and Signs and Douleur

Neuropathique 4 questions neuropathic pain screening tools in subjects with low back–related leg pain. J Manip Physiol Ther 2012;35:196-202.

Lasry-Levy E, Hietaharju A, Pai V, Ganapati R, Rice AS, Haanpaa M,

SC

26.

Lockwood DN. Neuropathic pain and psychological morbidity in patients with

Neglect Trop D 2011;5:e981. 27.

M AN U

treated leprosy: a cross-sectional prevalence study in Mumbai. PLOS

Spallone V, Morganti R, D'Amato C, Greco C, Cacciotti L, Marfia GA. Validation of DN4 as a screening tool for neuropathic pain in painful diabetic

28.

TE D

polyneuropathy. Diabetic Med 2012;29:578-585.

Padua L, Briani C, Truini A, Aprile I, Bouhassira D, Cruccu G, Jann S, NobileOrazio E, Pazzaglia C, Morini A, Mondelli M, Ciaramitaro P, Cavaletti G, Cocito

EP

D, Fazio R, Santoro L, Galeotti F, Carpo M, Plasmati R, Benedetti L, Schenone A. Consistence and discrepancy of neuropathic pain screening tools DN4 and

AC C

ID-Pain. Neurol Sci 2013;34:373-377.

29.

Santos JG, Brito JO, de Andra DC, Kaziyama VM, Ferreira KA, Souza I, Teixeira MJ, Bouhassira D, Baptista AF. Translation to Portuguese and validation of the Douleur Neuropathique 4 questionnaire. J Pain 2010;11:484-490.

30.

Perez C, Galvez R, Huelbes S, Insausti J, Bouhassira D, Diaz S, Rejas, J. Validity and reliability of the Spanish version of the DN4 (Douleur Neuropathique 4 questions) questionnaire for differential diagnosis of pain

18

ACCEPTED MANUSCRIPT syndromes associated to a neuropathic or somatic component. Health Qual Life Out 2007;5:66-76. 31.

Hallström H, Norrbrink C. Screening tools for neuropathic pain: can they be of use in individuals with spinal cord injury? Pain 2011;152:772-779. Unal-Cevik I, Sarioglu-Ay S, Evcik DA. Comparison of the DN4 and LANSS

RI PT

32.

questionnaires in the assessment of neuropathic pain: validity and reliability of the Turkish version of DN4. J Pain 2010;11:1129-1135. 33.

Reyes-Gibby C, Morrow PK, Bennett MI, Jensen MP, Shete S. Neuropathic pain

34.

M AN U

Symptom Manag 2010;39:882-889.

SC

in breast cancer survivors: using the ID pain as a screening tool. J Pain

Chan A, Wong S, Chen PP, Tsoi TH, Lam J, Ip J, Wong CP, Wong L, Mok V. Validation study of the Chinese Identification Pain Questionnaire for neuropathic pain. Hong Kong Med J 2011;17:297-300.

35.

Li J, Feng Y, Han J, Fan B, Wu D, Zhang D, Du D, Liu H, Lin J, Wang J, Jin Y, Fu

TE D

Z. Linguistic adaptation, validation and comparison of 3 routinely used neuropathic pain questionnaires. Pain Physician 2012;15:179-186. 36.

Kitisomprayoonkul W. Validation study of the Thai ID Pain Scale. J Med Assoc

Potter J, Higginson IJ, Scadding JW, Quigley C. Identifying neuropathic pain in

AC C

37.

EP

Thai 2011;94:610-615.

patients with head and neck cancer: use of the Leeds Assessment of Neuropathic Symptoms and Signs Scale. J Roy Soc Med 2003;96:379-383.

38.

Tampin B, Briffa NK, Goucke RC, Slater H. Identification of neuropathic pain in patients with neck/upper limb pain: application of a grading system and screening tools. Pain 2013;154:2813-22.

39.

Mercadante S, Gebbia V, David F, Aielli F, Verna L, Casuccio A, Porzio G, Mangione S, Ferrera P. Tools for Identifying Cancer Pain of Predominantly Neuropathic Origin and Opioid Responsiveness in Cancer Patients, J Pain 19

2009;10:594-600. 40.

ACCEPTED MANUSCRIPT

Barbosa M, Bennett MI, Verissimo R, Cavalho D. Cross cultural psychometric assessment of the Leeds Assessment of Neuropathic Symptoms and Signs (LANSS) pain scale in the Portuguese population. Pain Prac 2013;14:620-624. Schestatsky P, Félix-Torres V, Chaves MLF, Câmara-Ehlers B, Mucenic T, Caumo W,

RI PT

41.

Nascimento O, Bennett MI. Brazilian Portuguese validation of the Leeds Assessment of Neuropathic Symptoms and Signs for Patients with chronic pain. Pain Med

42.

SC

2011;12:1544-1550.

Yucel A, Senocak M, Orhan EK, Cimen A, Ertas M. Results of the Leeds

M AN U

Assessment of Neuropathic Symptoms and Signs pain scale in Turkey: a validation study. J Pain 2004;5:427-432. 43.

Weingarten TN, Watson JC, Hooten WM, Wollan PC, Melton LJ3rd, Locketz AJ, Wong GY, Yawn BP. Validation of the S-LANSS in the community setting. Pain

44.

TE D

2007;132:189-194.

Elzahaf RA, Tashani OA, Unsworth BA, Johnson MI. Translation and linguistic validation of the Self-Completed Leeds Assessment of Neuropathic Symptoms and Signs (S-LANSS) Scale for use in a Libyan population. Pain Pract

Koc R, Erdemoglu AK. Validity and reliability of the Turkish Self-administered

AC C

45.

EP

2013;13:198-205.

Leeds Assessment of Neuropathic Symptoms and Signs (S-LANSS) Questionnaire. Pain Med 2010;11:1107-1114.

46.

Timmerman H, Wolff AP, Schreyer T, Outermans J, Evers AW, Freynhagen R, Wilder- Smith OH, van Zundert J, Vissers KC. Cross-cultural adaptation to the Dutch language of the PainDETECT-Questionnaire. Pain Pract 2013;13:206-214.

47.

Gauffin J, Hankama T, Kautiainen H, Hannonen P, Haanpaa M. Neuropathic pain and use of PainDETECT in patients with fibromyalgia: a cohort study. BMC Neurol 2013;13:21-27. 20

48.

ACCEPTED MANUSCRIPT Matsubayashi Y, Takeshita K, Sumitani, M, Oshima, Y, Tonosu, J, Kato, S, Ohya, J, Oichi, T, Okamoto, N, Tanaka S. Validity and reliability of the Japanese version of the PainDETECT questionnaire: a multicentre observational study. PLOS One 2013;8:e68013.

49.

Hochman JR, Gagliese L, Davis AM, Hawker GA. Neuropathic pain

RI PT

symptoms in a community knee OA cohort. Osteoarthr Cartilage 2011;19:647-654. 50.

Backonja M, Krause SJ. Neuropathic Pain Questionnaire-short form. Clin J Pain

AC C

EP

TE D

M AN U

SC

2003;19:315-316.

21

ACCEPTED MANUSCRIPT

Legends Figure 1 PRISMA Flow chart

Table 1 Diagnostic methods of included studies

Appendix A Search filter Appendix B Criteria used to assess measurement properties of included studies

RI PT

Table 2 Summary of findings of included studies

M AN U

Appendix D Patient characteristics of included studies

SC

Appendix C Criteria used to assess the level of evidence for each measurement property of included studies

Appendix E Measurement properties of neuropathic pain screening questionnaires, grouped by questionnaire then by language version. The properties were assessed using criteria by Terwee [16] (appendix B)

AC C

EP

TE D

Appendix F Methodological quality of included studies assessed using QUADAS-2 (for criterion validity) [17] and the COSMIN checklist (for all other properties) [18]

22

ACCEPTED MANUSCRIPT

Table 1 Diagnostic methods of included studies. Studies marked in bold indicate that two clinicians were used to diagnose neuropathic pain.

DN4 7 Bouhassira et al 2005 21

Clinical history, sensory testing, additional tests 22

Harourn et al 2012 23 Harifi et al 2011 24 Van Seventer et al 2012 25

Walsh et al 2012 26 Lasry-Levy et al 2011 27 Spallone et al 2012 28 Pauda et al 2013 29 Santos et al 2010

Clinical history and sensory testing Clinical history, sensory testing, additional tests Inadequate methods

Inadequate methods Clinical history and sensory testing Clinical history and sensory testing Clinical history, sensory testing, additional tests

30

Unal Cevik et al 2010

TE D

Clinical history, sensory testing, additional tests Perez et al 2007 31 Hallström & Norrbrink 2011

Clinical history, sensory testing, additional tests Clinical history and sensory testing

32

LANSS 9 a,b Bennett 2001 37 Potter et al 2003

EP

Clinical history, sensory testing, additional tests

Clinical history and sensory testing Inadequate methods Clinical history, sensory testing, additional tests Inadequate methods Clinical history, sensory testing, additional tests

AC C

ID Pain 8 a,b Portenoy 2006 33 e Reyes-Gibby et al 2010 34 Chan et al 2011 35 Li et al 2012 28 Pauda et al 2013 36 Kitisomprayoonkul 2011

Sensitivity d (%)

Specificity d (%)

Pain specialist, Neurologist Neurologist, Rheumatologist Investigator Clinician Anesthetists, Neurologist Physiotherapist Clinician Investigator Clinician Pain specialist, Neurologist, orthopedist Pain specialist Neurologist, Pain physician Neurologist, pain specialist, physiatrists

82.9

89.9

80

92

100 89.4 74

45 72.4 79

NA 78.6 80 82 100

NA 100 92 81 93.2

79.8 92.9

78 75.0

95

96.6

Pain specialist Self reported Pain specialist Pain specialist Not described Physiatrists off site

NA 50 81 78 78 83

NA 86 65 74 74 80

Pain clinician Pain specialist

83 79

87 100

M AN U

Attal et al 2011

Clinical history and sensory testing

Diagnosing clinician

RI PT

Method of neuropathic pain diagnosis

SC

Study

Inadequate methods Clinical history, sensory testing, additional tests Clinical history and sensory testing

1

ACCEPTED MANUSCRIPT

88 42 52.9 91.4 74 NA 100

Clinical history, sensory testing, additional tests

89.9

94.2

Clinical history, sensory testing, additional tests Inadequate methods Clinical history and sensory testing Inadequate methods Clinical history, sensory testing, additional tests

Pain specialist Physiotherapist Pain specialist Self reported Pain specialist

74 NA 52 NA 72.3

76 NA 78 NA 80.4

Clinical history, sensory testing, additional tests Inadequate methods Clinical history, sensory testing, additional tests Clinical history and sensory testing Clinical history and sensory testing

Pain specialist Physician Physiotherapist Rheumatologist Pain specialist Clinicians at pain clinic Neurologist, Pain physician

85 NA 64 79 NA 75

80 NA 62 53 NA 84

67.9

83.0

Not described

NA

NA

Hospital physician Pain specialist Physician Neurologist, Pain physician

74.7 52.9 50.8 50

77.6 91.4 64.8 100

Hospital physician Physician

64.5 50.8

78.6 60.6

Clinical history and sensory testing

31

Clinical history and sensory testing Inadequate methods

EP

Modified PainDETECT 49 e Hochman et al 2011 NPQ 11 Krause & Backonja 2003 35 Li et al 2012 39 Mecandante et al 2009 31 Hallström & Norrbrink 2011

TE D

Inadequate methods

Inadequate methods Inadequate methods Clinical history, sensory testing, additional tests

AC C

Hallström & Norrbrink 2011

22 85 80 29.5 89 NA 35.7

M AN U

Yucel et al 2004 S-LANSS 12 Bennett et al 2005 25 Walsh et al 2012 43 e,f Weingarten et al 2007 44 b,c,f Elzahaf et al 2013 45 Koc & Erdemoglu 2010 PainDETECT 10 Freynhagen et al 2006 46 Timmerman et al 2013 38 Tampin et al 2013 47 e Gauffin et al 2013 48 Matsubayashi et al 2013 13 DeAndres et al 2012

Physiotherapist Investigator Pain specialist Physician Pain clinician Neurologist Neurologist, Pain physician Pain specialist

Clinical history and sensory testing

NPQ-short form 50 Backonja & Krause 2003 39 Mecandante et al 2009

Inadequate methods Clinical history, sensory testing, additional tests

a

c

b

RI PT

42

Clinical history, sensory testing, additional tests Clinical history and sensory testing Inadequate methods Clinical history, sensory testing, additional tests Clinical history, sensory testing, additional tests Clinical history, sensory testing, additional tests

SC

38

Tampin et al 2013 22 Harourn et al 2012 35 Li et al 2012 39 Mecandante et al 2009 40 Barbosa et al 2010 41 Schestatsky et al 2011 31 Hallström & Norrbrink 2011

d

scale development. evaluation study. test-retest study. questionnaire cut off values used to determine sensitivity and specificity: DN4 value was 4 (except Harifi et al [23] which a value of 3 was used). ID pain was 2 (except Chan [34] which a value of 3 was used). LANSS and S-LANSS was 12. e f PainDETECT value was19 (except Gauffin [47] which a value of 17 was used). NPQ and short form was 0. NA= not reported. phone administration. mail administration. 2

ACCEPTED MANUSCRIPT

Table 2 Summary of findings of included studies

+ + + + + -+ -+ + 0 +++

RI PT

+ + + 0 ++ + + + ++ +

Reliability

Hypothesis testing

Cross cultural validity

0 0 0 0 0 0 0 0 0 0

-0 0 0 0 0 0 ++ 0 0 0

++ 0 0 -++ 0 ++ 0 0 0 +

0 0 0 0 0 0 --

0 0 0 0 0 0 ++ -

++ 0 ++ 0 0 0 a ++ ++ ++

+ +

--0 0 0

--0 0 0

+++ 0 ++ 0

0 0 -

0 0 -

++ 0 0

+ + ++ + ++ +

0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0

0 0 0 --

0 0 0 0

0 0 0 ++ ++ 0

0 0 0 0 0

+++ --0 +

0 0 0 -

0 0 -

0 0 + --

+ + 0 +++

0 0 0 0 0

SC

+ + + + 0 ++ + + + ++ +

Internal Consistency

Structural validity

M AN U

Specificity

Construct validity

TE D

Sensitivity

Content validity

EP

DN4 7,21 French (original) 22 Amharic 23 Arabic 24 Dutch 25 English 26 Hindi & Marathi 27,28 Italian 29 Portuguese 30 Spanish 31 Swedish 32 Turkish ID Pain 8,33b English (original) 34,35 Chinese 28 Italian 36 Thai LANSS 9,37,38 English (original) 22 Amharic 35 Chinese 39 Italian 40,41 Portuguese 31 Swedish 42 Turkish S-LANSS 12a,25,43b,c English (original) 43b English 43c English 44c Arabic 45 Turkish

Criterion validity

AC C

QUESTIONNAIRE

1

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

PainDETECT 10 German (original) + + 0 0 0 + 0 46 Dutch 0 0 0 0 0 0 0 38 English 0 0 -0 0 0 47b Finnish + 0 0 0 0 0 0 48 Japanese 0 0 0 ++ + + 13 Spanish + + 0 -++ -31 Swedish -++ 0 0 0 0 -Modified PainDETECT 49b English (original) 0 0 0 0 ++ 0 0 0 NPQ 11 English (original) + + -++ +++ 0 +++ -33 Chinese + 0 0 0 0 39 Italian --0 0 0 0 0 0 31 Swedish -++ 0 0 0 0 ++ NPQ-Short form 50 English (original) + 0 0 + 0 0 0 39 Italian --0 0 0 0 0 0 a b c All studies the clinician administered the questionnaire, except for those self-reported, mail administration or phone administered. NPQ= Neuropathic Pain Questionnaire. Scoring: ++++ = satisfactory measurement property, high level of evidence, +++ = satisfactory measurement property, moderate level of evidence, ++ = satisfactory measurement property, low level of evidence, + = satisfactory measurement property, very low level of evidence, 0 = no evidence (not reported), ---- = unsatisfactory measurement property, high level of evidence, --- = unsatisfactory measurement property, moderate level of evidence, -- = unsatisfactory measurement property, low level of evidence, - = unsatisfactory measurement property, very low level of evidence (Scoring criteria defined in appendix C).

2

ACCEPTED MANUSCRIPT

Additional records identified through other sources (n = 4)

RI PT

Records identified through database searching (n = 424)

Records after duplicates removed (n = 314)

SC

Screening

Identification

Figure 1 PRISMA Flow chart

Full-text articles excluded (n = 14) [No clinimetrics n = 6 Repeat study n = 5 Data not published n = 1 Not published in English n = 2]

TE D

Full-text articles assessed for eligibility (n = 51)

Records excluded (n = 263)

EP

Studies included in qualitative synthesis (n = 37)

AC C

Included

Eligibility

M AN U

Records screened (n = 314)

1

ACCEPTED MANUSCRIPT

Appendix A Search filter

A published sensitive search filter was used [14] against the neuropathic pain screening questions. Similar strategies were applied for Medline and CINHAL.

Search strategy used in EMBASE “PainDETECT” OR “PDQ” AND sensitive search filter (see below)

2.

“Douleur Neuropathique 4” OR “DN4” AND sensitive search filter

3.

“ID Pain” AND sensitive search filter

4.

“Neuropathic Pain Questionnaire” OR “NPQ” AND sensitive search filter

5.

“Leeds Assessment of Neuropathic Symptoms and Signs” OR “LANSS” OR “SLANSS” OR “S-LANSS”

SC

RI PT

1.

M AN U

OR “Self Leeds Assessment of Neuropathic Symptoms and Signs” AND sensitive search filter

Sensitive search filter

Instrumentation OR methods OR Validation Studies OR Comparative Study OR psychometrics OR psychometr* OR clinimetr* OR clinometr* OR outcome assessment (health care) OR outcome assessment OR outcome measure* OR observer variation OR observer variation OR Health Status Indicators OR reproducibility of results OR reproducib* OR discriminant analysis OR reliab* OR unreliab* OR valid* OR coefficient OR homogeneity OR homogeneous OR internal consistency OR (cronbach* AND (alpha OR alphas)) OR (item AND (correlation*

TE D

OR selection* OR reduction*)) OR agreement OR precision OR imprecision OR precise values OR test–retest OR (test AND retest) OR (reliab* AND (test OR retest)) OR stability OR interrater OR inter-rater OR intrarater OR intra-rater OR intertester OR inter-tester OR intratester OR intra-tester OR interobserver OR inter-observer OR intraobserver OR intra-observer OR intertechnician OR inter-technician OR intratechnician OR intra-

EP

technician OR interexaminer OR inter-examiner OR intraexaminer OR intra-examiner OR interassay OR interassay OR intraassay OR intra-assay OR interindividual OR inter-individual OR intraindividual OR intra-individual OR interparticipant OR inter-participant OR intraparticipant OR intra-participant OR kappa OR kappas OR

AC C

repeatab* OR ((replicab* OR repeated) AND (measure OR measures OR findings OR result OR results OR test OR tests)) OR generaliza* OR generalisa* OR concordance OR (intraclass AND correlation*) OR discriminative OR known group OR factor analysis OR factor analyses OR dimension* OR subscale* OR (multitrait AND scaling AND (analysis OR analyses)) OR item discriminant OR interscale correlation* OR error OR errors OR individual variability OR (variability AND (analysis OR values)) OR (uncertainty AND (measurement OR measuring)) OR standard error of measurement OR sensitiv* OR responsive* OR ((minimal OR minimally OR clinical OR clinically) AND (important OR significant OR detectable) AND (change OR difference)) OR (small* AND (real OR detectable) AND (change OR difference)) OR meaningful change OR ceiling effect OR floor effect OR Item response model OR IRT OR Rasch OR Differential item functioning OR DIF OR computer adaptive testing OR item bank OR cross cultural equivalence.

1

ACCEPTED MANUSCRIPT

Appendix B Criteria used to assess measurement properties of included studies PROPERTY Validity Criterion validity

Content validity

SCORE

+

Convincing argument for comparison to gold standard AND correlation (or sensitivity and specificity for dichotomous gold standard and questionnaire) with gold standard ≥0.70. No convincing argument that gold standard is ‘gold’. Correlation with gold standard <0.70, despite good experimental design. No information on criterion validity. Clear description of measurement aim, population, concepts measured, item selection and those involved in item selection Not a clear description of above characters OR only target population described OR poor design. No target population involved. No information on target population. Factors explain ≥50% of variance Explained factors not mentioned Factors explain ≤50% of variance No information on structural validity. Specific hypothesis formulated AND >75% of results fulfilling hypotheses or convergent validity ≥0.5 or divergent validity ≤0.5. Poor design or method. Less than 75% hypotheses confirmed or convergent validity <0.5 or divergent validity >0.5, despite good experimental design. No information on construct validity. No differences in factor structure OR no important differential item functioning between language versions. Multiple group factor analysis not applied and differential item functioning not assessed Differences in factor structure OR important differential item functioning between language versions. No information on cross cultural validity.

? 0 +

0 + ? 0 +

SC

Construct validity - Hypothesis testing

RI PT

?

Construct validity - Structural Validity

M AN U

? -

Construct validity - Cross cultural validity

0 + ? 0

Agreement

Responsiveness

? 0 + ? 0 + ? 0

Adequate sample size (≥100) for factor analysis AND Cronbach’s alpha calculated between 0.70 and 0.95. No factor analysis OR poor design. Cronbach’s alpha <0.70 or >0.95, despite good experimental design. No information on internal consistency. ICC or weighted Kappa ≥0.70. Poor design or method. ICC or weighted Kappa <0.70, despite good experimental design. No information on reliability. SDC>MIC (with sample size ≤50). Poor design or method. MIC≥SDC, despite good experimental design. No information on agreement.

TE D

Reliability

+

EP

Reliability Internal consistency

a

ASSESSMENT CRITERIA [16]

AC C

+ AUC ≥0.70 OR SDC
1

ACCEPTED MANUSCRIPT

Appendix C Criteria used to assess the level of evidence for each measurement property of included studies a

No downgrade Inconsistency

Risk of Bias

Imprecision

b

>75% results are the same in a single measurement property Excellent or good COSMIN score OR low QUADAS-2 score >400 participants

SCORE Serious (downgrade 1 level)

Very serious (downgrade 2 levels)

≤ 75% results inconsistent in a single measurement property

NA

Majority fair COSMIN scores OR unclear QUADAS-2 score

Majority poor or mixture of COSMIN scores OR high QUADAS-2 score

<400 participants

a

RI PT

CRITERIA

Very wide confidence interval (CI) or wide CI plus <400 participants

All scores started at high level of evidence base on the measurement property score (+ as satisfactory, ? or – as unsatisfactory. The worst score for that measurement property was used if more than one study available) and were downgraded according to the criteria including inconsistency, risk of bias and imprecision.

AC C

EP

TE D

M AN U

SC

b Risk of bias for criterion validity was assessed using QUADAS-2 [17]. The overall QUADAS-2 score was the worst score for risk of bias and applicability. Methodological quality for all other properties was assessed using the COSMIN checklist [15]. The overall COSIN score was the worst score for that measurement property.

1

ACCEPTED MANUSCRIPT

Appendix D Patient characteristics of included studies

23

Harifi et al 2011 22 Harourn et al 2012 49 Hochman et al 2011 36

Kitisomprayoonkul 2011

Koc & Erdemoglu 2010

45

Krause & Backonja 2003 26 Lasry-Levy et al 2011 35 Li et al 2012 Matsubayashi et al 2013 Mercadante et al 2009

39

11

48

28/12

Pain Intensity b (VAS /10) (mean ± SD) 6.4 ± 1.8 Not reported 7.2 ± 2.5 [NP=4.9, NoC=5.1] [NP=5.5, NoC=4.2] [NP = 8.0, NoC=8.0] 6.7 ± 1.5 Not reported 6.6 ± 1.7 6.8 ± 2.2 Not reported Not reported [NP=6.3 ± 2.0 NNP=5.6 ± 2.2] [NP=6.0, NNP=3.5]

DN4 NPQ-short form LANSS LANSS LANSS S-LANSS DN4 ID Pain PainDETECT S-LANSS S-LANSS PainDETECT PainDETECT

132 low back pain patients in France 278 chronic pain patients in America 167 chronic pain patients in Portugal 60 chronic pain patients in UK 40 chronic pain patients in UK 200 chronic pain patients in UK 160 chronic pain patients in France 92 pain patients in Hong Kong 221 chronic pain patients in Spain 25 chronic pain patients in Libya 13 chronic pain patients in Libya 392 pain patients in Germany 158 fibromyalgia patients in Finland

DN4, LANSS, PDQ, NPQ DN4 DN4, LANSS Modified PainDETECT ID Pain

40 spinal cord injury patients in Sweden

170 chronic pain patients in Morocco 80 leprosy patients in Ethiopia 171 hip/knee arthritis patients in Canada

49.5 ± 12.4 34.0 76.0

41/129 50/30 39/132

4.7 ± 2.1 Not reported Not reported

100 pain patients in Thailand

[NP=51.3 ± 15.5, NoC=51.3 ± 15.2, Mixed=53.9 ± 13.7] 43.1 ± 11.4

25/75

Not reported

S-LANSS

244 chronic pain patients in Turkey

77/167

47.1 ± 15.8 39.3 [NP=64.3 ± 14.4, NoC=59.5 ± 12.2] [NP=59.0 ± 15.0, NoC=57.0 ± 18.0] [No NP=65.5 ± 12.2 Possible NP=66.5 ± 10.4 Definite NP=65.5 ± 11.2]

192/340 73/28 59/81

[NP=7.3 ± 2.3 NoC=6.8 ± 2.3] Not reported Not reported Not reported

NPQ DN4 ID Pain, LANSS, NPQ PainDETECT LANSS, NPQ, NPQ-SF

SC

RI PT

Sex (M/F)

M AN U

31

Mean Age b (mean years ± SD) 54.7 ± 14.3 47.1 ± 15.8 55.3 ± 13.4 [NP=48.1, NoC=60.8] [NP=52.6, NoC=55.2] 53.8 56.0 ± 16.0 51.0 57.8 ± 14.2 52.0 ± 18.0 35.8 ± 6.7 Not reported [NP=49.0 ± 8.0, NNP=46.0 ± 12.0] 49.0

TE D

Hallström & Norrbrink 2011

Participants

EP

21

Attal et al 2010 50 Backonja & Krause 2003 40 Barbosa et al 2010 9 c Bennett 2001 9 d Bennett 2001 12 Bennett et al 2005 7 Bouhassira et al 2005 34 Chan et al 2011 13 DeAndres et al 2012 44 d Elzahaf et al 2013 44 e Elzahaf et al 2013 10 Freynhagen et al 2006 47 Gauffin et al 2013

Questionnaire

AC C

Study

532 chronic pain patients in America 101 leprosy patients in India 140 chronic pain patients in China

113 pain patients in Japan 167 chronic cancer patients in Italy

52/80 192/340 Not reported 26/34 13/27 73/127 57/103 42/50 91/130 8/17 7/6 Not reported 9/149

70/43 103/64

[NP=6.7 ± 2.0 NoC=4.2 ± 2.3] [No NP=5.4 ± 1.9 Possible NP=4.9 ± 2.2 Definite NP=5.9 ± 2.2] 1

ACCEPTED MANUSCRIPT

37

Potter et al 2003 33 Reyes-Gibby et al 2010 29

Santos et al 2010 41 Schestatsky et al 2011

Spallone et al 2012

Tampin et al 2013

27

38

Timmerman et al 2013 Unal Cevik et al 2010

46

32

Van Seventer et al 2012 25

Walsh et al 2012 43 Weingarten et al 2007 42 Yucel et al 2004 a

ID Pain

308 chronic pain patients in America

LANSS ID Pain

25 head and neck cancer patients in UK 240 breast cancer survivors patients in America 101 chronic pain patients in Brazil 90 chronic pain patients in Brazil

DN4 LANSS

DN4

158 painful diabetic poly neuropathy patients in Italy

LANSS, PainDETECT

152 neck and upper limb pain patients in Australia

PainDETECT

60 pain patients in The Netherlands & Belgium 180 pain patients in Turkey

DN4 24

218/174

Not reported

60.1 ± 15.9 [NP=55.6 ± 12.7, NoC=50.7 ± 14.8, MP=51.2 ± 14.1] [NP=51.9 ±14.0, NoC=50.3 ± 15.3, MP=50.0 ± 14.7] 60.0 58.0 ± 16 .0

65/93 224/362

6.5 ± 2.1 Not reported

111/197

Not reported

14/11 0/240

6.3 Not reported

51.0 ± 12.0 [NP=54.0 ± 12.0, NoC=55.5 ± 14.2, MP=48.7 ± 10.6] [PDPN+= 56.8 ± 12.2, DPN+= 59.6 ± 12.4. DPN=46.3 ± 17.1] [No NP=58.0 ± 14.4, Possible NP=51.2 ± 10.0, Probable NP=51.1 ± 12.0, Definite NP=52.6 ± 11.0] 54.0 ± 15.0

25/76 Not reported

7.0 ± 1.0 [NP=7.6 ± 1.5 NoC=6.7 ± 2.2 MP=6.9 ± 2.5] Not reported

[NP=53.3 ± 14.1, NNP=48.7 ± 13.1] 53.3

68/112

46.0 ± 11.0 56.7 ± 13.0 [NP=53.9 ± 15.6 NoC=49.0 ± 15.7]

22/23 82/123 38/63

RI PT

8 d

DN4 ID Pain

58.8

SC

c

M AN U

30

392 peripheral nerve disease patients in Italy 158 chronic pain patients in Spain 586 chronic pain patients in America

DN4 DN4, S-LANSS S-LANSS LANSS

TE D

Portenoy 2006

DN4, ID Pain

EP

Perez et al 2007 8 Portenoy 2006

28

269 chronic pain patients in The Netherlands 45 low back-leg pain patients in Ireland 205 chronic pain patients in America 101 pain patients in Turkey

AC C

Pauda et al 2013

82/76

80/72

36/24

87/182

[No NP=3.6 ± 2.4, Possible NP= 4.5 ± 2.3, Probable NP=4.8 ± 2.3, Definite NP=4.7 ± 2.2] 5.6 ± 2.5 [NP=5.3 ± 1.6 NNP=4.7 ± 1.5] 5.6 ± 2.2 6.1 ± 2.6 5.0 [NP=6.7 ± 0.9 NoC= 7.1 ± 0.8]

b

Examples of pain and chronic conditions include post herpetic neuralgia, diabetic peripheral neuropathy, tunnel syndromes and radiculopathy. subgroup data presented in square c d e brackets if total sample data not available. scale development. evaluation study. test-retest study. Abbreviations: NP= neuropathic pain, NoC= nociceptive pain, MP= mixed pain, NNP= non-neuropathic pain, VAS = Visual Analogue Scale, NPQ= Neuropathic Pain Questionnaire.

2

ACCEPTED MANUSCRIPT

Appendix E Measurement properties of neuropathic pain screening questionnaires, grouped by questionnaire then by language version. The properties were assessed using criteria by Terwee [16] (appendix B) a

English English English Amharic Chinese

Portenoy 2006 e Reyes-Gibby et al 2010 Chan et al 2001 Li et al 2012 Pauda et al 2013 Kitisomprayoonkul 2001

+ + + + + 0 + + + + + + +

+ + + + 0 + + + + + + +

Bennett 2001 Potter et al 2003 Tampin et al 2013 Haroun et al 2012 Li et al 2012

i

Reliability Internal Reliability consistency

? 0 0 0 0 0 0 0 0 0 0 0 0

? 0 0 0 0 0 0 0 0 + 0 0 0

+ 0 0 0 i g + 0 g,h + g + 0 0 0 g +

0 0 0 ? 0 0 0 0 0 ? ? ? ?

0 0 0 ? 0 0 0 0 0 + ? 0 ?

+ + + +

M AN U

Specificity

TE D

Bouhassira et al 2005 Attal et al 2010 Haroun et al 2012 Harifi et al 2011 Van Seventer et al 2012 Walsh et al 2012 Lasry-Levy et al 2011 Spallone et al 2012 Pauda et al 2013 Santos et al 2010 Perez et al 2007 Hallström & Norrbrink 2011 Unal-Cevik et al

Sensitivity

RI PT

Criterion validity

SC

b

EP

DN4 French French Amharic Arabic Dutch English Hindi & Marathi Italian Italian Portuguese Spanish Swedish Turkish ID Pain English English Chinese Chinese Italian Thai LANSS

STUDY

+ 0 0 + 0 0 0 0 c

0 ? + + + +

0 ? + + +

? 0 0 0 0 0

? 0 0 0 0 0

+ g + 0 0 g + 0

h,i

0 0 ? ? 0 ?

0 0 0 ? 0 ?

+ 0 ? 0 0 0

+ + + +

+ + + +

? 0 0 0 0

0 0 0 0 0

+ i i ? 0 0

i

0 0 0 0 ?

? 0 0 0 ?

0 0 0 0

AC C

QUESTIONNAIRE

MEASUREMENT PROPERTY Validity Content Construct validity validity Structural Hypothesis Cross validity testing cultural validity

1

d

ACCEPTED MANUSCRIPT

+ + 0 + +

0 0 0 0 0

Bennett et al 2005 Walsh et al 2012 e Weingarten et al 2007 f Weingarten et al 2007 f Elzahaf et al 2013 Koc & Erdemoglu 2010

+ 0 0 +

+ 0 + 0 +

0 0 0 0 0 0

Freynhagen et al 2006 Timmerman et al 2013 Tampin et al 2013 e Gauffin et al 2013 Matsubayashi et al 2013 deAndres et al 2012 Hallström & Norrbrink 2011

+ 0 + 0 + -

e

0 0 0 0 0

RI PT

+ 0 +

M AN U + 0 0 + +

? 0 0 0 0 0 0

0 0 0 0 0

0 ? ? ? ?

0 + ? 0 0

0 + + + 0

0 0 0 0 0 0

+ g + i i 0 i +

i

0 0 0 0 ? ?

? 0 0 0 ? ?

? 0 0 0 + ?

0 0 0 0 + ? 0

0 0 i ? 0 h g + 0

0 ? 0 0 ? ? ?

+ 0 0 0 + ? 0

0 0 0 0 + ? -

SC

Mercadante et al 2009 Barbosa et al 2010 Schestatsky et al 2011 Hallström & Norrbrink 2011 Yucel et al 2004

TE D

Italian Portuguese Portuguese Swedish Turkish S-LANSS English English English English Arabic Turkish PainDETECT German Dutch English Finnish Japanese Spanish Swedish Modified PainDETECT

g

AC C

EP

+ 0 0 0 English Hochman et al 2011 0 0 0 0 NPQ i + English Krause & Backonja 2003 + + ? + 0 + ? Chinese Li et al 2012 + 0 0 0 ? ? 0 Italian Mercadante et al 2009 0 0 0 0 0 0 Swedish Hallström & Norrbrink 2011 + 0 0 0 ? 0 + NPQ -Short form i + English Backonja & Krause 2003 + 0 0 0 0 0 Italian Mercadante et al 2009 0 0 0 0 0 0 Scoring: + = positive rating, ? = indeterminate rating, - = negative rating, 0 = no information available (scoring criteria defined in appendix B). a b c d e responsiveness, agreement and interpretability were not reported in any study. original developmental study listed first. practitioner administration. self reported. mail f g h i administration. phone administration. convergent validity assessed. divergent validity assessed. hypothesis testing assessed. NPQ=Neuropathic Pain Questionnaire.

2

ACCEPTED MANUSCRIPT

Appendix F Methodological quality of included studies assessed using QUADAS-2 (for criterion validity) [17] and the COSMIN checklist (for all other properties) [15]

Portenoy 2006 Reyes-Gibby et al 2010 c Chan et al 2001 Li et al 2012 Pauda et al 2013 Kitisomprayoonkul 2001

Poor

Bennett 2001 Potter et al 2003 Tampin et al 2013 Haroun et al 2012 Li et al 2012 Mercadante et al 2009

M AN U

Unclear High Unclear High High Unclear High

High High High Unclear High

Reliability

Fair

Poor

Poor

Fair Poor

Fair Poor Poor e Poor

Fair Poor

Poor Fair Fair Fair

Fair Fair Poor Fair

Poor Fair

Internal Consistency

Fair

Fair

Fair

High High High High High Unclear

Fair

RI PT

High High High High High

MEASUREMENT PROPERTY a Structural Hypothesis Cross-cultural Validity testing validity

SC

Content Validity

TE D

Bouhassira et al 2005 Attal et al 2010 Haroun et al 2012 Harifi et al 2011 Van Seventer et al 2012 Walsh et al 2012 Lasry-Levy et al 2011 Spallone et al 2012 Pauda et al 2013 Santos et al 2010 Perez et al 2007 Hallström & Norrbrink 2011 Unal-Cevik et al

Criterion Validity

EP

DN4 French French Amharic Arabic Dutch English Hindi & Marathi Italian Italian Portuguese Spanish Swedish Turkish ID Pain English English Chinese Chinese Italian Thai LANSS English English English Amharic Chinese Italian

STUDY

AC C

QUESTIONNAIRE

b

Poor

Fair Fair

Fair Poor e Poor

Poor

Poor

Poor

Poor

Fair

Poor

Fair Poor Fair

Poor

Poor

Poor

Poor

1

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Portuguese Barbosa et al 2010 High Fair e Fair Fair Portuguese Schestatsky et al 2011 Poor Poor Fair Swedish Hallström & Norrbrink 2011 Unclear Poor e Fair Turkish Yucel et al 2004 High Poor S-LANSS English Bennett et al 2005 High Fair Poor Poor English Walsh et al 2012 Fair English Weingarten et al 2007 c High Fair d English Weingarten et al 2007 High Fair Arabic Elzahaf et al 2013 d Poor Poor Poor Turkish Koc & Erdemoglu 2010 Low Poor Poor Poor Fair PainDETECT German Freynhagen et al 2006 High Poor Fair Dutch Timmerman et al 2013 Poor English Tampin et al 2013 High Fair Finnish Gauffin et al 2013 c High Japanese Matsubayashi et al 2013 Fair Poor Poor e Poor Poor Spanish deAndres et al 2012 High Fair Fair Poor Poor Fair e Swedish Hallström & Norrbrink 2011 Unclear Poor Fair Modified PainDETECT English Hochman et al 2011 c Fair NPQ English Krause & Backonja 2003 High Poor Fair Fair Fair Poor Chinese Li et al 2012 High Poor Poor Italian Mercadante et al 2009 Unclear Swedish Hallström & Norrbrink 2011 Unclear Poor e Fair NPQ -Short form English Backonja & Krause 2003 High Poor Italian Mercadante et al 2009 Unclear a Measurement error, responsiveness and interpretability were not reported in any study. COSMIN checklist used to assessment methodological quality, except for criterion validity [15]. QUADAS-2 was used to assess methodological quality for criterion validity [17], (scored as low, unclear or high risk of bias). b Original developmental study listed first. c mail administration. d phone administration. e translation only. NPQ= Neuropathic Pain Questionnaire.

2