Content and psychometric evaluations of questionnaires for assessing physical function in people with neck disorders: a systematic review of the literature

Abstract

Purpose: The purpose was to investigate how physical function is assessed in people with musculoskeletal disorders (MSD) in the neck. Specifically, we aimed to determine: (1) Which questionnaires are used to assess physical function in people with MSD in the neck? (2) What do those questionnaires measure? (3) What are the measurement properties of the questionnaires?

Materials and methods: A systematic review was performed to identify questionnaires and psychometric evaluations. The content of the questionnaires was categorized according to the International Classification of Function, Disability and Health, and the psychometric properties were quality-rated using the COnsensus-based Standards for the selection of health Measurement INstruments checklist.

Results: Ten questionnaires and 32 articles evaluating measurement properties were analyzed. Most questionnaires covered only the components body functions and activity and participation, more often activity participation than body function. Internal consistency was adequate in most questionnaires, whereas responsiveness was generally low. Neck Disability Index was most evaluated, but the evaluations of all questionnaires tended to cover most properties in the checklist.

Conclusions: The questionnaires differed substantially in items and extent to which their psychometric properties had been evaluated. Focus of measurement was on activities in daily life rather than physical function as such.

Implications for Rehabilitation

• To provide early diagnostics and effective treatment for patients with neck disorders, valid and reliable instruments that measure relevant aspects of the disorders are needed.

• This paper presents an overview of content and quality of questionnaires used to assess physical function in neck disorders, which may facilitate informed decisions about which measurement instruments to use when evaluating the course of neck disorders.

• Most of the questionnaires need more testing to judge the quality, however the NDI was the most frequently tested questionnaire.

• The COnsensus-based Standards for the selection of health Measurement INstruments checklist is a useful tool in relation to psychometric testing of questionnaires, but clear definitions of interpretation of the quality criteria in each study would enhance comparability of results.

Keywords:

• Questionnaire
• measure
• neck
• ICF
• psychometrics

Introduction

Musculoskeletal disorders (MSD) are common, and cause a burden to both individuals and society. World-wide prevalence rates of 33% for shoulder pain, 50% for low back pain, and 20% for widespread pain in a 1-month period have been reported [1]. Similar numbers are presented for neck disorders in, e.g., the Netherlands (31% 1-year prevalence) [2]. To reduce these disorders, we need early diagnostics and effective treatment/rehabilitation. In order to achieve this, valid and reliable instruments that measure relevant aspects of the disorders are required. This is of importance both for correctly diagnosing a patient's condition prior to treatment, and for evaluating intervention effects.

A cost-efficient way of assessing a patient’s condition is to use patient rated outcome measure scales, as they provide the necessary information without involving a lot of personnel to perform health examinations or interviews. One problem with using patient rated outcome measure scales, however, is the lack of consistency in selecting outcome measures, both in general pain research [3,4], and in research focusing on MSD [5]. The Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) [4] has recommended that the following four domains should be included: pain intensity, physical function, emotional function, and general improvement.

When assessing physical function, IMMPACT recommends the Brief Pain Inventory [6] or the Multidimensional Pain Inventory [7,8]. A closer examination of the items included in the Multidimensional Pain Inventory shows that they tend to measure the consequences of disability in daily life rather than the disability itself, and it is not clear which components and aspects of function and disability are being measured according to the International Classification of Functioning, Disability and Health (ICF) [9]. It may be components related to body functions and body structures, i.e., impairment level, or components related to activities and participation. In order to correctly diagnose and provide specific, targeted interventions to patients, we need to distinguish physical function per se (e.g., ability to turn the head) from the consequences of daily life (e.g., ability to dress).

Vast numbers of instruments are available for measuring physical function in MSD. Reasons for this may be differences in definitions of MSD, and that many instruments are specific to certain body regions. In this study, we focus on questionnaires intended and evaluated for measuring physical function among people with nonspecific musculoskeletal conditions in the neck, hence we exclude studies where the questionnaires have been used in populations with symptoms due to trauma, infection, systemic diseases and congenital or acquired deformities. This limitation was made to assess the utility of the questionnaires in a homogeneous sample. Besides what is being captured by the instrument, the quality of the instrument is also important when deciding which instrument to use. A useful tool for determining which properties have been evaluated for an instrument, and subsequently the validity and reliability of the instrument, is the COSMIN checklist (COnsensus-based Standards for the selection of health Measurement INstruments) [10–13].

The aim of this systematic review was to investigate how physical function is assessed in people with MSD in the neck. Specifically, we aimed to determine:

• Which questionnaires are used to assess physical function in people with MSD in the neck?

• What do those questionnaires measure?

• What are the measurement properties of the questionnaires?

Materials and methods

Search for questionnaires

The study was performed as a systematic literature review [14,15] searching the databases PubMed, Cinahl, Web of Science, and PsycInfo for studies published until 22 November 2016. Table 1 contains the keywords used in the search, which were defined based on the ideas behind the PICO model [16]. We selected suitable keywords for P (population), I (intervention), O (outcome), while C (comparison) was excluded as comparative studies were not the focus of this study.

Table 1. Keywords sorted in word groups.

O (outcome)	I (intervention)	P (population)	P (population)
physical funct*	measure*	pain (not acute)	musculoskeletal
funct*	scale		neck
disability	questionnaire*		low back
	outcome measure*		back
	inventory
	evaluation
	psychometric*
	reliability
	validity

Excluded: rheumatoid arthr*, gonarthros*, gout, cancer, foot, ankle, hip, knee, elbow, wrist, clinical trial, effect*, jaw, head*, trauma*, fear avoidance, kinesiophobia.

The systematic literature review process is illustrated in a flowchart (Figure 1). The literature search in PubMed resulted in 1333 articles, in Web of Science 457 articles, in PsycInfo 91 articles, and in Cinahl 41 articles. After excluding duplicates, the selection of relevant articles to include was performed by reviewing the title and abstracts first, for a crude culling of irrelevant articles. At this stage, 284 articles were included from PubMed, 139 articles from Web of Science, 0 article from Cinahl, and 0 article from PsycInfo. Thereafter, the abstracts, and when needed the full text, of the articles were reviewed to find articles presenting the original version of questionnaires for measurement of physical function in MSD. Both authors performed the review separately and then discussed the selection, to reach consensus. Excluded were questionnaires regarding general health, fibromyalgia, spinal stenosis, scoliosis, ankylosing spondylitis, ratings of functional capacity, work disability, not patient rated outcome measure scales, and questionnaires for use among specific occupational groups or adolescents/children.

Figure 1. Flowchart; systematic literature review process.

A total of 98 different questionnaires were identified. Despite the keywords and exclusion criteria used in the search, some questionnaires had to be excluded because they were measuring pain beliefs or were physical tests to be performed by clinicians. One questionnaire, the Patient Specific Functional Scale [17] was excluded at this stage, as it did not have fixed items but asked the patient to specify two examples of their disability, which made further analysis impossible. For the purpose of this review, only the questionnaires used to measure physical function among people with neck pain were selected for further analysis. They were classified according to the ICF [9] to determine their focus of measurement and coverage of items. The ICF-classification was performed independently by both authors, followed by a comparative discussion to reach consensus.

Search for quality assessment of questionnaires

To identify the measurement properties that have been evaluated for the questionnaires, a second literature search in the previous databases was performed using the key words “validity OR reliability OR responsiveness AND neck AND name of questionnaire” to find articles that had quality-assessed the questionnaires.

Only articles which analyzed the original version of the questionnaire were included. Studies concerning translated versions of the questionnaires and studies where ≥50% of the sample met our previously used exclusion criteria were excluded.

Quality assessment of the questionnaires

The finally selected articles were examined with respect to relevant quality indicators of validity and reliability using the COSMIN checklist [11,13]. This examination was performed by both authors, first separately, then together to reach consensus regarding what properties of methodological quality had been assessed for each questionnaire. For each of the studies assessing the quality of the questionnaires, we rated the findings of the measurement properties as “adequate” or “not adequate” according to guidelines in the COSMIN checklist. Our quality criteria were inspired by the work of Schellingerhout et al. [18], and are listed below.

• Internal consistency (adequate = unidimensional (sub)scale or Cronbach’s alpha(s) ≥ 0.70; not adequate = otherwise or cannot be determined)

• Reliability (adequate = intraclass correlation/weighted Kappa ≥0.70 or Pearson’s r ≥ 0.80; not adequate = otherwise or cannot be determined)

• Measurement error (adequate = standard error of measurement, smallest detectable change or limits of agreement presented; not adequate = otherwise or cannot be determined)

• Content validity (adequate = the target population was involved in the development of the questionnaire; not adequate = otherwise or cannot be determined)

• Construct validity (adequate = factors explained at least 50% of the variance; not adequate = otherwise or cannot be determined)

• Criterion validity (adequate = correlation with criterion instrument(s) ≥ 0.50; not adequate = otherwise or cannot be determined)

• Responsiveness (adequate = correlation with anchor instrument ≥0.50 or area under receiver operator curve ≥0.70; not adequate = otherwise or cannot be determined)

• Interpretability was marked “all” if the distribution of scores, floor/ceiling effects, scores for relevant groups and minimal important change or minimal important difference were reported, and “some” otherwise.

For criterion validity, only questionnaires (i.e., not single scales such as visual analog scales) were considered as criterion instruments. If an instrument (either evaluated instrument or criterion instrument) contained several scales, the methodological quality was determined by taking the lowest rating of any of the scales.

Results

From the literature search, we identified 10 questionnaires that are used to assess physical function among people with neck pain [19–28]. They are listed in Table 2. When each questionnaire’s items were classified according to the ICF, we found that most questionnaires covered only the components body functions and activity and participation (Tables 3 and 4). The Cervical Spine Outcomes Questionnaire differed by also including items about environmental factors.

Table 2. Included neck questionnaires.

Cervical Spine Outcomes Questionnaire [19]

Copenhagen Neck Functional Disability Scale [20]

Neck Bournemouth Questionnaire [21]

Neck Disability Index [22]

Neck Pain and Disability Scale [23]

Neck Pain and Functional Limitation Scale [24]

Neck Functional Status Questionnaire [25]

Neck OutcOme Score [26]

Northwick Park Neck Pain Questionnaire [27]

ProFitMap Neck [28]

Table 3. ICF classification of the included neck questionnaire.

	ICF-categories
	Function and Impairment																Contextual factors
	Body functions (b)							Activity and participation (d)									Environmental factors (e)
Questionnaire	1	2	3	4	5	6	7	1	2	3	4	5	6	7	8	9	1	2	3	4	5
Cervical Spine Outcomes Questionnaire	134 152	265 280			510	640	730 760				440	520 540 550	649		845 850	910 920 930	110 130				570 580
Copenhagen Neck Functional Disability Scale	134 152 160	280						166	230		410 420 430	530 540		750 760		920 999
Neck Bournemouth Questionnaire	152 160 164	280						166			430 475	540	640		850	910 920
Neck Disability Index	134 160	280						166			430 475	599	640		859	920
Neck Pain and Disability Scale	134 152 160	280					780				410 429 450 470 475	510 540 550		750 760 770	859	920 999
Neck Pain And Functional Limitation Scale	134 152 160 164	280						110 166			410 430 489	540	620	760	859	920
Neck Functional Status Questionnaire							710				410 415 430 445	510 520	640 650			920
Neck OutcOme Score	134 152 160	240 280					710 780				415		620 630 640	750 760 770	850	920
Northwick Park Neck Pain Questionnaire	134	265 280						110 166			430 475		640		859	999
ProFitMap-neck	134 152 160	210 230 235 240 265 280		440	510 535		710 720 730 735 740 760 780				410 415 430 445 450 455	540			845

Body functions (b): b1 = mental functions, b2 = sensory functions and pain, b3 = voice and speech, b4 = cardio-vascular-, hematological-, immunological- and respiratory functions, b5 = digestive-, metabolic- and endocrine functions, b6 = gender-, urinary- and reproductive functions, b7 = neuromusculoskeletal and movement related functions. Activity and participation (d): d1 = learning and applying knowledge, d2 = general tasks and demands, d3 = communication, d4 = mobility, d5 = personal care, d6 = domestic life, d7 = Interpersonal interactions and relationships, d8 = Important areas of life, d9 = community, social and civic life. Environmental factors (e): e1 = products and technology, e2 = natural environment and human-made environmental changes, e3 = support and relationships, e4 = attitudes, e5 = services, systems and policies.

Note: If several items of a questionnaire concerned the same aspect, it was only categorized once. If an item included several aspects, all aspects were listed in the table.

Table 4. ICF codes with explanation.

b: Body functions	d: Activities and participation	e: Environmental factors
1 Mental functions 134 Sleep functions 152 Emotional functions 160 Thought functions 164 Higher-level cognitive functions2 Sensory functions and pain 210 Seeing functions 230 Hearing functions 235 Vestibular functions 240 Sensations associated with hearing and vestibular function 265 Touch function 280 Sensation of pain 4 Functions of the cardiovascular, hematological, immunological and respiratory systems 440 Respiration functions 5 Functions of the digestive, metabolic and endocrine systems 510 Ingestion functions 535 Sensations associated with the digestive system 6 Genitourinary and reproductive functions 640 Sexual functions 7 Neuromusculoskeletal and movement-related functions 710 Mobility of joint functions 720 Mobility of bone functions 730 Muscle power functions 735 Muscle tone functions 740 Muscle endurance functions 760 Control of voluntary movement functions 780 Sensations related to muscles and movement functions	1 Learning and applying knowledge 110 Watching 166 Reading 2 General tasks and demands 230 Carrying out daily routine 4 Mobility 410 Changing basic body position 415 Maintaining a body position 420 Transferring oneself 429 Changing and maintaining body position, other specified and unspecified 430 Lifting and carrying objects 440 Fine hand use 445 Hand and arm use 450 Walking 455 Moving around 470 Using transportation 475 Driving 489 Moving around using transportation, other specified and unspecified 5 Self-care 510 Washing oneself 520 Caring for body parts 530 Toileting 540 Dressing 550 Eating 599 Self-care, unspecified 6 Domestic life 620 Acquisition of goods and services 630 Preparing meals 640 Doing housework 649 Household tasks, other specified and unspecified 650 Caring for household objects 7 Interpersonal interactions and relationships 750 Informal social relationships 760 Family relationships 770 Intimate relationships 8 Major life areas 845 Acquiring, keeping, and terminating a job 850 Remunerative employment 859 Work and employment, other specified and unspecified 9 Community, social and civic life 910 Community life 920 Recreation and leisure 930 Religion and spirituality 999 Community, social, and civic life, unspecified	1 Products and technology 110 Products or substances for personal consumption 130 Products and technology for education 5 Services, systems and policies 570 Social security services, systems and policies 580 Health services, systems, and policies

Within the component body functions, items tended to cover “mental functions” and “sensory functions and pain”. For mental functions, the most common items assessed were “sleep functions” (b134), “thought functions” (b160) and “emotional functions” (b152). For sensory functions and pain, item “pain” (b280) was most commonly assessed. The widest coverage was found for the ProFitMap-neck questionnaire and Cervical Spine Outcomes Questionnaire, each covering five of seven aspects of body functions (Tables 3 and 4). The Neck Functional Status Questionnaire did not include any items on sensory functions such as pain.

Within the component activity and participation, the questionnaires generally covered more aspects than in other components, the exception thus being the ProFitMap-neck questionnaire and Cervical Spine Outcomes questionnaire. Most items in activity and participation addressed “mobility”, while “communication” was not considered in any of the questionnaires. Items concerning the ability “to lift and carry objects” (d430), “dressing” (d540), and “recreation and leisure” (d920) were most frequently included in the questionnaires.

In the search for studies evaluating measurement properties of the 10 questionnaires, we identified 32 articles that fulfilled the inclusion criteria [19–50]. From these, we extracted and categorized the properties of methodological quality that was assessed for each of the questionnaires according to the COSMIN checklist (Table 5). The results show that Neck Disability Index has been evaluated in many studies, and it also has been assessed for all analyzed properties in the checklist. Neck OutcOme Score has also been assessed for all analyzed properties in the check list, but only in few studies. While remaining questionnaires have been evaluated in fewer studies than the Neck Disability Index, the evaluations tend to cover most properties in the checklist. The exception is the Neck Pain and Functional Limitation Scale that have been assessed for three properties only. Notably, internal consistency has been assessed for each of the 10 questionnaires, but measurement error has been assessed for only five of the 10 questionnaires.

Table 5. Mapping of psychometric assessment of the included questionnaires.

	Reliability			Validity
	Internal consistency	Reliability	Measurement error	Content validity	Construct validity	Criterion validity	Responsiveness	Interpretability
CSOQ	[19,29,30]	[19,30]	–	[19]	[30]	[19,29]	[19,30]	[19,29,30]
CNFDS	[20]	[20]	–	–	[31]	–	[20]	[20]
NBQ	[21,32]	[21]	[21,32]	–	[33,34]	[21,32]	[21,32,34,35]	[21,32,33]
NDI	[22,32,36–39]	[22,37,40–43]	[32,37,40–44]	[45]	[31,36,46]	[22,32,37,44,45]	[32,40–43]	[22,32,36–46]
NFSQ	[25]	–	–	[25]	[25]	–	–	[25]
NOOS	[47]	[47]	[47]	[26]	[26]	[47]	[47]	[26,47]
NPDS	[23]	[48]	–	[23,45]	[23,31]	[23,45,48]	[48]	[23,45,48]
NPFLS	[24]	–	–	[24]	–	[24]	–	–
NPNPQ	[27,49]	[27]	[49]	–	[31]	–	[27,49]	[27,49]
PFM	[28]	[28]	[28]	[28]	–	[28]	[50]	[28,50]

CSOQ: Cervical Spine Outcomes Questionnaire; CNFDS: Copenhagen Neck Functional Disability Scale; NBQ: Neck Bournemouth Questionnaire; NDI: Neck Disability Index; NFSQ: Neck Functional Status Questionnaire; NOOS: Neck OutcOme Score; NPDS: Neck Pain and Disability Scale; NPFLS: Neck Pain and Functional Limitation Scale; NPNPQ: Northwick Park Neck Pain Questionnaire; PFM: ProFitMap-neck.

The methodological quality of the questionnaires is summarized in Table 6. Internal consistency of the questionnaires clearly varied between studies. For most questionnaires, however, the internal consistency was found to be adequate in the majority of studies investigating the property. Test–retest reliability, on the other hand, was generally only adequate in questionnaires where it had been evaluated in a single study. While measurement error was reported adequately in few studies, most of these studies evaluated Neck Disability Index. Thus, several studies report on the measurement error of Neck Disability Index.

Table 6. Methodological quality of the included questionnaires.

		Reliability			Validity
	Article	Internal consistency	Reliability	Measurement error	Content validity	Construct validity	Criterion validity	Responsiveness	Interpretability
CSOQ	BenDebba et al. (2002)	+	+		+		ODI – SF36 –	–	Some
	Skolasky et al. (2007)	–					NDI + SF36– ODI–		Some
	Skolasky et al. (2009)		–			+		–	Some
CNFDS	Jordan et al. (1998)	+	+					–	Some
	Pickering et al. (2011)					+
NBQ	Bolton et al. (2002)	+	–	+			NDI– CNFDS + SF36–	–	Some
	Gay et al. (2007)	+		–			NDI +	+	Some
	Bolton et al. (2004)							–
	Geri et al. (2015)	–				+			Some
	Hurst et al. (2004)							–
NDI	Vernon and Mior (1991)	+	+				MPQ +		Some
	Hains et al. (1998)	+				+			Some
	Gabel et al. (2014)					–			Some
	Riddle and Stratford (1998)			–			SF36–		Some
	Cleland et al. (2006)		–	+				–	Some
	Gay et al. (2007)	+		–			NBQ+	+	Some
	McCarthy et al. (2007)	+	+	+			SF36–		Some
	Cleland et al. (2008)		–	+				+	Some
	En et al. (2009)				–		NPDS+		Some
	van der Velde et al. (2009)	–							Some
	Young et al. (2009)		–	+				+	Some
	Young et al. (2010)		+	+				+	Some
	Hung et al. (2014)	–							Some
	Pickering et al. (2011)					+
NFSQ	Wang et al. (2015)	–			–	+			Some
NOOS	Juul et al. (2015)				+	+			Some
	Juul et al. (2016)	+	+	+			NDI + SF36+	–	Some
NPDS	Wheeler et al. (1999)	+			–	+	ODI + PDI+		Some
	Goolkasian et al. (2002)		+				NDI– PDI– ODI–	+	Some
	En et al. (2009)				–		NDI +		Some
	Pickering et al. (2011)					+
NPFLS	Leonard et al. (2009)	+			+		NBQ+
NPNPQ	Leak et al. (1994)	–	+					–	Some
	Sim et al. (2006)	+		+				–	Some
	Pickering et al. (2011)					+
PFM	Björklund et al. (2012)	+	+	+	+		NDI + SF36– SES–		Some
	Björklund et al. (2016)							+	All

+: adequate; −: not adequate. For further information, see “Quality assessment of the questionnaires” section.

CSOQ: Cervical Spine Outcomes Questionnaire; CNFDS: Copenhagen Neck Functional Disability Scale; NBQ: Neck Bournemouth Questionnaire; NDI: Neck Disability Index; NFSQ: Neck Functional Status Questionnaire; NOOS: Neck OutcOme Score; NPDS: Neck Pain and Disability Scale; NPFLS: Neck Pain and Functional Limitation Scale; NPNPQ: Northwick Park Neck Pain Questionnaire; PFM: ProFitMap-neck. Criterion questionnaires; MPQ: McGills Pain Questionnaire; ODI: Oswestry Disability Index; PDI: Pain Disability Index; SF-36: Short Form 36 Health Survey; SES: Self Efficacy Scale.

Content validity was found to be adequate in four questionnaires only: Cervical Spine Outcomes Questionnaire, Neck OutcOme Score, Neck Pain and Functional Limitation Scale, and ProFitMap-neck. When assessing construct validity, we found that factor analysis of the construct to be measured had been performed for most questionnaires, and the majority of the studies reported adequate validity. It should be noted, however, that the findings are based on few studies. For criterion validity, the findings were rated for each of the criterion instruments listed in Table 6. The two most commonly used criterion instruments were Neck Disability Index and the Short Form 36 Item Health Survey (SF-36). For NDI the results were inconsistent. Overall, the criterion validity toward SF-36 was low, except for Neck OutcOme Score which showed adequate criterion validity toward the bodily pain subscale of SF-36, according to the authors’ hypotheses [47]. The Oswestry Disability Index was used as criterion instrument in four studies. Its correlation with Cervical Spine Outcomes Questionnaire was low, and the findings with respect to Neck Pain and Disability Scale were inconsistent. Responsiveness has been evaluated for eight of the 10 questionnaires, and most results were negative. Only for Neck Disability Index, Neck Pain and Disability Scale, and ProFitMap-neck, the majority of the findings were positive. For each of the latter two questionnaires, responsiveness was only evaluated in one study.

In addition to the ratings of quality, we assessed the degree to which important characteristics relating to the instrument’s interpretability were reported in the studies. Almost all studies reported some characteristics of the instrument, but only one study evaluating ProFitMap-neck reported all characteristics considered important.

Discussion

Most questionnaires included questions about body functions as well as activity and participation. However, the questionnaires tended to have a larger focus (more items) on activity and participation than on body function. In other words, they tended to measure disability or the ability to cope in everyday life, rather than physical function as such. Two questionnaires constituted an exception; Cervical Spine Outcomes Questionnaire and ProFitMap-neck questionnaire. While the former was nearly equally balanced in items between the two components, the latter had an overweight on body function.

Reviewing the literature, we found four articles [51–54] that classified the content of MSD questionnaires according to ICF. Only one of these [54] investigated neck questionnaires. In that study, five of the questionnaires included in the present study (Copenhagen Neck Functional Disability Scale, Neck Bournemouth Questionnaire, Neck Disability Index, Neck Pain and Disability Scale, Northwick Park Neck Pain Questionnaire) were analyzed. Similarly to our findings, Ferreira et al. [54] showed that the questionnaires contained more items about activity and participation than about body function. They concluded that the Neck Bournemouth Questionnaire, Neck Disability Index, and Neck Pain and Disability Scale have a well-balanced distribution of items over the ICF-components. In the present study, however, we found the Cervical Spine Outcomes Questionnaire to be even more equally balanced than the other questionnaires included.

For activity and participation, items regarding mobility were most commonly assessed, followed by personal care and community, social and civic life. This seems reasonable, since being able to move, take care of oneself and participate in social life are basic abilities to be self-sufficient for everyone. For body functions, mental functions and sensory functions and pain were covered by almost all questionnaires, which is relevant as depression, difficulty concentrating, and pain are common symptoms in this disorder [5]. Neuromusculoskeletal function, on the other hand, was assessed in Neck Pain and Disability Scale and Neck Functional Status Questionnaire (by one item), in Cervical Spine Outcomes Questionnaire and Neck OutcOme Score (by two items), and in ProFitMap-neck (by seven items). It is somewhat surprising that items on neuromusculoskeletal functions were not included in all questionnaires, as it would appear to be a highly relevant aspect to measure in patients with neck MSD where radiating pain, sensory loss and muscle weakness are common [55].

According to a previous review of psychometric evaluations performed on neck specific questionnaires, the evidence for validity and reliability of the questionnaires was mostly limited, and at least 50% of the information about measurement properties per questionnaire was lacking [18]. Our results showed that most properties of the COSMIN checklist have been assessed for each of the included questionnaires, although the evaluations have generally been performed in few studies. An obvious exception is Neck Disability Index, which has been evaluated in more than twice as many studies as the other questionnaires. This is consistent with the previous review, where Neck Disability Index also was reported to be the most frequently evaluated questionnaire. Among the measurement properties investigated in the questionnaires, measurement error was the least assessed, and should be considered in future studies.

Unfortunately, the instruments in present study that covered most aspects of body function were only evaluated in few studies. Three studies were found for the Cervical Spine Outcomes Questionnaire [19,29,30] and two for the ProFitMap-neck questionnaire [28,50]. However, as psychometric evaluations continue to be made, more studies on the included questionnaires may already be underway.

Making comparisons to previous reviews on methodological assessment of questionnaires comes with some practical problems. We found two reviews that appeared to be comparable; one by Schellingerhout et al. [18] and another by Leahy et al. [56]. Methodologically these studies are comparable to the present, as all are based on the COSMIN checklist. However, the interpretation of the included quality criteria slightly differs. Furthermore, they used different criteria for inclusion and exclusion of samples and questionnaires. For example, Schellingerhout et al. [18] also included people with whiplash and traumatic neck injuries in their sample, and Leahy et al. [56] included questionnaires concerning pain anywhere in the spine region. As a consequence, comparisons can only be made to Schellingerhout et al. [18], and only for some of the included questionnaires. The comparable questionnaires included in both studies are Neck Disability Index, Neck Pain and Disability Scale, Neck Bournemouth Questionnaire, and Copenhagen Neck Functional Disability Scale.

Similar to Schellingerhout et al. [18], we found Neck Disability Index to be the most frequently assessed questionnaire. Its internal consistency was confirmed in slightly more than half of the studies evaluating it. Our findings of questionable reliability and adequate responsiveness are in agreement with Schellingerhout et al. [18], whereas we did not find evidence of adequate content validity in the included studies. Also, the measurement error of Neck Disability Index was reported in several studies, albeit not with minimal important change which was required in Schellingerhout et al. [18].

Compared to Neck Disability Index, few studies have evaluated Neck Pain and Disability Scale, Neck Bournemouth Questionnaire, and Copenhagen Neck Functional Disability Scale. Previous reports of limited positive evidence of responsiveness for Neck Pain and Disability Scale [18] were confirmed in this study. In addition, our results show poor content validity and adequate construct validity in the studies reviewed. For Neck Bournemouth Questionnaire, the findings of internal consistency were inconsistent. Contrary to Schellingerhout et al. [18], we found poor responsiveness of the questionnaire. Due to few studies evaluating Copenhagen Neck Functional Disability Scale, no firm conclusions can be drawn about its reliability or validity.

As indicated previously, it is difficult to interpret and compare results from different reviews, depending on differences in inclusion/exclusion criteria, differences in ways of determining the quality criteria in the COSMIN checklist, and differences in the ways the results are presented. Reasons for low inter-rater reliability may be differences in interpretation of checklist items [57]. This was evident in the review by Leahy et al. [56] where content validity was judged excellent for the Extended Aberdeen Back Pain Scale, based on the results of a study using qualitative content analysis to compare interviews with people with neck pain and the content of questionnaires [5]. That study does not make any psychometric assessment of the included questionnaires. Moreover, it only notes that only Extended Aberdeen Back Pain Scale and ProFitMap-neck have used the experiences of the affected in development, and the correspondence between experiences and questionnaires was mostly low. In the present study, we were methodologically stringent by including a homogenous sample, only neck questionnaires, and clearly defining and describing our use of the quality criteria.

To conclude, we found that questionnaires for assessing physical function in people with neck disorders tend to focus more on the ability to engage in activities of daily life than on physical function per se. There are some exceptions, but the measurement properties of these questionnaires have been evaluated in few studies.

As we have experienced, the results from quality assessments differ between studies, even though the same checklist has been used. The COSMIN checklist [11,13] is a useful contribution in quality assessment of patient rated outcome measure scales, but clear descriptions of the interpretation of these quality criteria in each study would enhance the comparability between studies. A homogenous sample would also facilitate comparisons.

When investigating neck disorders, it is vital to measure aspects that are of importance for the individual wellbeing and ability to function in work and daily life. Likewise, including the most relevant aspects of physical function in research studies will increase the validity of the results. Therefore, we recommend that more psychometric testing is performed of the questionnaires, especially with respect to content validity and measurement error, and particularly of questionnaires that cover important aspects of the ICF categories.

Disclosure statement

The authors declare that they have no conflict of interest.