Measurement in low back pain: a brief introduction

Measurement is at the core of science and essential in clinical practice. In health sciences, this typically corresponds to the measurement of health- and/or disease-related outcome measurement instruments. These can include measures of pathophysiological variables (e.g. radiography, magnetic resonance imaging (MRI) or clinical chemistry measurement in blood samples), physical tests for measuring constructs such as muscle strength or range of motion, and patient-reported outcome measures (PROMs) aiming to measure health-related quality of life (HRQoL). The results of these measurements are often the basis on which the clinical management is (or is not) altered. In addition, decisions about the reimbursement of health care interventions are (at least partly) based on measurements such as the EuroQol-5D (EQ-5D) questionnaire. This means that outcome measurement instruments need to be valid, reliable and responsive, otherwise there is a serious risk of imprecise or biased results. In clinical trials on low back pain (LBP), PROMs are the most frequently used type of measurement instruments , and the same is likely in LBP clinical practice. They are efficient and do not require advanced technologies or high costs for administration. In this manuscript, we focus on PROMs, although the fundamental issues also apply to other types of instruments.

The number of available PROMs has dramatically increased over the past few decades; consequently, the choice of which PROM to use is becoming more difficult. There are often multiple instruments available for measuring the same health construct in the same patient population. For example, a systematic review published in 2005 identified 36 PROMs for measuring back-specific functional status in patients with LBP . This means that there is a high risk that poor quality instruments are being used, which can introduce information bias into research or practice.

Given the number of PROMs available, it is not surprising that there is an inconsistency in outcome assessment across clinical trials; this hampers the comparability of results and makes conducting meta-analyses difficult . The lack of large meta-analyses means that estimates of intervention effectiveness are not precise and that research is less informative for clinical practice. Another problem is that researchers tend to selectively report their outcomes, choosing only those for which there were more favourable results . Problems of outcome inconsistency and selective reporting can be addressed by the development of a core outcome set (COS) . A COS is an agreed minimum set of outcomes to be measured and reported in all clinical trials in a specific health condition . COSs are usually developed for clinical research, but since they represent the most relevant outcomes, they are often applicable to clinical practice as well .

Different stakeholders (e.g. clinicians, researchers, patients, policy makers, health insurance and industry representatives) with relevant expertise should be involved in establishing a COS. The development of a COS is a two-step process: first, determine which outcome domains should be included (i.e. ‘what’ to measure) and second, select measurement instruments for the core outcome domains (i.e. ‘how’ to measure) . The outcome domain is the construct or aspect of interest to be measured, and it is sometimes represented by a latent variable that cannot be directly observed (e.g. physical functioning, pain interference or fatigue). The measurement instrument is the means used to quantify the construct . A detailed description of the methodology to develop a COS can be found in the recent work summarising the topic .

How to select an outcome measurement instrument

The first step in selection of an instrument is definition of the outcome domain and the target population . Defining specifically ‘what’ to measure is crucial because domains with the same name may be defined in different ways . For instance, ‘disability’ is defined by the World Health Organization as ‘problems an individual may experience in functioning, namely impairments, activity limitations and participation restrictions’ ; however, Garrad and Bennett defined ‘disability’ as ‘limitation of the performance of an individual when compared to a fit person’ . The target population also needs to be carefully defined because aspects of the same domain may be differently important in different populations. For example, self-care activities relevant to very disabled patients with LBP may not be very relevant for high-functioning patients, such as long-distance runners who experience LBP only after having run a certain distance.

The second step concerns the identification of suitable measurement instruments for the outcome domain . These can be identified through up-to-date systematic literature reviews or other sources such as databases specific to outcome instruments. An example is the systematic review of Miles et al. who aimed to identify all PROMs used in research to measure the domain of pain self-efficacy in patients with chronic pain. Widely used databases to retrieve PROMs are the Rehabilitation Measures Database, for the rehabilitation field , or the eProvide Mapi Research website .

Researchers and clinicians who wish to select a few instruments from a long list may consider the OMERACT ‘eye-ball’ test . This test includes two simple questions: (1) ‘is the instrument a good match with the domain?’, and (2) ‘is the instrument feasible in the setting of the core set?’ . The first question refers mainly to the face validity of an instrument with regard to the population and setting of interest; the second question requires an evaluation of the costs, patient and responder burden, equipment needs and other practical aspects. Having pre-selected some instruments, the next step is to extensively assess their quality . It should be noted that although this is an attractive and pragmatic strategy, it should be preceded by a systematic literature search for all available instruments to avoid missing suitable instruments.

Judging the quality of an outcome measurement instrument requires the evaluation of the evidence regarding its measurement properties in the target population . The Consensus-based Standards for the selection of health Measurement INstruments (COSMIN) initiative has developed methods for this . Nine measurement properties were identified and defined and subdivided into three major groups: validity, reliability and responsiveness ( Table 1 ) . The COSMIN checklist has been developed to rate the methodological quality of studies with regard to the nine measurement properties (Mokkink LB, de Vet HC, Prinsen CA, Patrick DL, Alonso J, Bouter LM et al. COSMIN Risk of Bias checklist for systematic reviews of Patient-Reported Outcome Measures. Under review) and proposes criteria to judge if the results for each measurement property are satisfactory ( Table 2 ) . The COSMIN initiative has also developed a system to rate the quality of the body of evidence on each property (Prinsen CA, Mokkink LB, Bouter LM, Alonso J, Patrick DL, de Vet HCW et al. COSMIN guideline for systematic reviews of Patient-Reported Outcome Measures. Under review) by adapting the GRADE approach for systematic reviews of clinical trials . These tools can be used in systematic reviews of measurement properties. Existing systematic reviews on measurement properties of measurement instruments for various health conditions can be retrieved in the COSMIN database ( http://database.cosmin.nl/ ).

These steps provide the framework for the recommendations in this paper regarding outcome domains and measurement instruments for research and clinical practice in patients with LBP.

How to select an outcome measurement instrument

The first step in selection of an instrument is definition of the outcome domain and the target population . Defining specifically ‘what’ to measure is crucial because domains with the same name may be defined in different ways . For instance, ‘disability’ is defined by the World Health Organization as ‘problems an individual may experience in functioning, namely impairments, activity limitations and participation restrictions’ ; however, Garrad and Bennett defined ‘disability’ as ‘limitation of the performance of an individual when compared to a fit person’ . The target population also needs to be carefully defined because aspects of the same domain may be differently important in different populations. For example, self-care activities relevant to very disabled patients with LBP may not be very relevant for high-functioning patients, such as long-distance runners who experience LBP only after having run a certain distance.

The second step concerns the identification of suitable measurement instruments for the outcome domain . These can be identified through up-to-date systematic literature reviews or other sources such as databases specific to outcome instruments. An example is the systematic review of Miles et al. who aimed to identify all PROMs used in research to measure the domain of pain self-efficacy in patients with chronic pain. Widely used databases to retrieve PROMs are the Rehabilitation Measures Database, for the rehabilitation field , or the eProvide Mapi Research website .

Researchers and clinicians who wish to select a few instruments from a long list may consider the OMERACT ‘eye-ball’ test . This test includes two simple questions: (1) ‘is the instrument a good match with the domain?’, and (2) ‘is the instrument feasible in the setting of the core set?’ . The first question refers mainly to the face validity of an instrument with regard to the population and setting of interest; the second question requires an evaluation of the costs, patient and responder burden, equipment needs and other practical aspects. Having pre-selected some instruments, the next step is to extensively assess their quality . It should be noted that although this is an attractive and pragmatic strategy, it should be preceded by a systematic literature search for all available instruments to avoid missing suitable instruments.

Judging the quality of an outcome measurement instrument requires the evaluation of the evidence regarding its measurement properties in the target population . The Consensus-based Standards for the selection of health Measurement INstruments (COSMIN) initiative has developed methods for this . Nine measurement properties were identified and defined and subdivided into three major groups: validity, reliability and responsiveness ( Table 1 ) . The COSMIN checklist has been developed to rate the methodological quality of studies with regard to the nine measurement properties (Mokkink LB, de Vet HC, Prinsen CA, Patrick DL, Alonso J, Bouter LM et al. COSMIN Risk of Bias checklist for systematic reviews of Patient-Reported Outcome Measures. Under review) and proposes criteria to judge if the results for each measurement property are satisfactory ( Table 2 ) . The COSMIN initiative has also developed a system to rate the quality of the body of evidence on each property (Prinsen CA, Mokkink LB, Bouter LM, Alonso J, Patrick DL, de Vet HCW et al. COSMIN guideline for systematic reviews of Patient-Reported Outcome Measures. Under review) by adapting the GRADE approach for systematic reviews of clinical trials . These tools can be used in systematic reviews of measurement properties. Existing systematic reviews on measurement properties of measurement instruments for various health conditions can be retrieved in the COSMIN database ( http://database.cosmin.nl/ ).

These steps provide the framework for the recommendations in this paper regarding outcome domains and measurement instruments for research and clinical practice in patients with LBP.

Outcome domains and measurement instruments for low back pain

Studies published in 1998 and 2000 proposed a set of outcome domains and instruments for clinical research into LBP. Five domains were recommended: back-specific function, pain symptoms, generic health status, work disability and satisfaction with care . These recommendations were endorsed by the Cochrane Back and Neck Review Group for inclusion in systematic reviews of LBP . More recently, an international, multidisciplinary group of researchers, clinicians and patient representatives updated the recommendations . In a Delphi survey, agreement was reached on the measurement of three core outcome domains in all LBP clinical trials: physical functioning, pain intensity and HRQoL . The same three outcome domains (although with different names) were also recommended by the International Consortium for Health Outcomes Measurement (ICHOM) for LBP, which aimed to provide a standard set of outcomes for routine clinical monitoring .

In addition to the three core outcome domains, 10 other domains reached a higher level of consensus than others (among 41 in total): work ability, psychological functioning, pain interference, health care services use, self-rated health, recreation and leisure, temporal aspects of pain, social functioning, work productivity and sleep functioning . Satisfaction with care was recommended in the original core set , but the level of consensus in the last Delphi was very low, and substantial arguments were presented against its inclusion (e.g. it tells relatively little about the effectiveness of an intervention as it can be influenced by several factors other than the intervention itself) .

In this manuscript, we present measurement instruments for the three core outcome domains (i.e. physical functioning, pain intensity and HRQoL). We also provide recommendations for three other important domains (work, psychological functioning and pain interference) that came close to reaching consensus in the Delphi study and recommended by other initiatives aimed at standardising measurement in LBP .

Physical functioning

Physical functioning, defined as ‘a patient’s ability to carry out daily physical activities required to meet basic needs, ranging from self-care to more complex activities that require a combination of skills’, achieved the highest level of consensus . The same domain was endorsed by the original core set for LBP , the ICHOM for LBP routine clinical practice , the NIH Task Force report on research standards for chronic LBP , and the Initiative on Methods, Measurement and Pain Assessment in Clinical Trials (IMMPACT) for chronic pain clinical trials . A recent systematic review of 185 trials of rehabilitation interventions in LBP reported that this domain was measured in approximately 64% of trials .

Characteristics of frequently used measurement instruments

The Roland Morris Disability Questionnaire (RMDQ) and the Oswestry Disability Index (ODI) are the instruments that have been recommended in the past and most frequently used to measure this domain . At least six different versions of the RMDQ and four different versions of the ODI have been developed and used . The 24-item original English version (RMDQ-24) and version 2.1a of the ODI (ODI 2.1a) are probably most commonly used. The RMDQ-24 was developed from the Sickness Impact Profile (SIP), and it consists of 24 statements representing ‘physical functions that were likely to be affected by LBP’, each statement is checked if it applies to the patient that day . The ODI 2.1a consists of 10 items: six items represent different physical functioning activities, while the other four represent other health constructs: pain intensity, sleep and social functioning .

The RMDQ-24 has been cross-culturally adapted into several languages and countries (Chiarotto A, Ostelo RW, Boers M, Terwee CB. A systematic review highlights the need to investigate the content validity of patient-reported outcome measures for physical functioning in low back pain. Under review). Froud et al. performed a qualitative assessment of the comprehensibility of this tool in patients with LBP, showing that its structure and content are understood as intended. However, its relevance and comprehensiveness (other aspects of content validity) have been questioned . More precisely, its time frame (i.e. ‘today’) may be inappropriate because LBP is often not a static condition, rather a fluctuating one . Further, some RMDQ items may be irrelevant for patients with chronic LBP, and some key aspects of functioning (e.g. driving, housework, leisure and exercise) are not covered . In addition, the unidimensionality (whether all items measure one single domain) has been questioned in all studies using factor analytic approaches or item response theory (IRT) analysis (Chiarotto A, Ostelo RW, Boers M, Terwee CB. A systematic review highlights the need to investigate the content validity of patient-reported outcome measures for physical functioning in low back pain. Under review). Other measurement properties such as test-retest reliability, construct validity and responsiveness appear to be satisfactory , but measurement error was relatively large when compared to the scale range .

The ODI 2.1a has also been cross-culturally adapted in various languages and countries. It appeared comprehensible in all languages, but other aspects of content validity (e.g. item relevance and comprehensiveness) have not yet been adequately investigated (Chiarotto A, Ostelo RW, Boers M, Terwee CB. A systematic review highlights the need to investigate the content validity of patient-reported outcome measures for physical functioning in low back pain. Under review). Conflicting evidence has been found on dimensionality, with some good quality studies indicating unidimensionality, and others not (Chiarotto A, Ostelo RW, Boers M, Terwee CB. A systematic review highlights the need to investigate the content validity of patient-reported outcome measures for physical functioning in low back pain. Under review). The other measurement properties (i.e. test-retest reliability, measurement error, construct validity and responsiveness) appear to be satisfactory .

According to a recent systematic review of head-to-head comparisons of the measurement properties in patients with non-specific LBP, the test-retest reliability and measurement error of ODI 2.1a may be better, while the construct validity of the RMDQ-24 may be better . It was concluded that there are no strong reasons to prefer one instrument over the other, but future studies should compare content validity and unidimensionality of these tools in the same sample as these properties have never been directly compared .

The third most frequently used PROM for physical functioning is the Quebec Back Pain Disability Scale (QBPDS) . The QBPDS was developed in English and French by Kopec et al. who interviewed experts and patients about which activities were most difficult or not possible because of LBP . The QBPDS includes 20 items and has been successfully cross-culturally adapted in several languages and countries, albeit fewer than RMDQ-24 and ODI 2.1a (Chiarotto A, Ostelo RW, Boers M, Terwee CB. A systematic review highlights the need to investigate the content validity of patient-reported outcome measures for physical functioning in low back pain. Under review). The content validity of the QBPDS has been assessed in Greece and Palestine and showed to adequately reflect relevant problems for patients with LBP . The dimensionality of the QBPDS has been investigated in a few studies, exhibiting conflicting findings . A recent systematic review on the QBPDS measurement properties did not find high-quality evidence for any measurement property but showed satisfactory results for all properties in different language versions, with the exception of unsatisfactory results for measurement error .

Suggested outcome measurement instruments

We suggest clinicians and researchers use one of RMDQ-24, ODI 2.1a or QBPS to measure physical functioning in patients with LBP ( Table 3 ). Although some measurement properties of these instruments may be suboptimal, these are the most thoroughly investigated tools, and the results are satisfactory for most properties. They are feasible for research and practice, requiring little burden to assessors and respondents, with fees applying only in certain circumstances for ODI 2.1a and QBPS ( Table 3 ). Of these, the ODI 2.1a was endorsed in preference to the others in a recent Delphi study (Chiarotto A, Boers M, Deyo RA, Buchbinder R, Corbin TP, Costa LO et al. Core outcome measurement instruments for clinical trials in non-specific low back pain. Submitted), in line with ICHOM recommendations for routine clinical practice . Nevertheless, we believe that there are no strong reasons to prefer one of these three PROMs as evidence directly comparing their measurement properties has exhibited that none of them are clearly and consistently better . To learn how to interpret their change scores, readers are referred to guidance published on this topic .

Table 3

Outcome domains and outcome measurement instruments for low back pain.

Outcome domains	Sub-domains	Outcome measurement instruments	Number of items or Questions	Response options	Scoring	Copyright, Fees	Online availability
Physical functioning		24-item Roland Morris Disability Questionnaire (RMDQ-24)	24	0–1	0–24	No, no fees	http://www.rmdq.org/
		Oswestry Disability Index version 2.1a (ODI 2.1a)	10	0–6	0–100	Yes, fees may apply	https://eprovide.mapi-trust.org/instruments/oswestry-disability-index
		Quebec Back Pain Disability Scale (QBPDS)	20	0–5	0–80	Yes, fees may apply	https://eprovide.mapi-trust.org/instruments/quebec-back-pain-disability-scale
Pain intensity		Numeric Rating Scale (NRS)	How would you rate your average low back pain intensity over the last week?	0–10	0–10	No, no fees
Health-related quality of life		EuroQol-5D-3L (EQ-5D-3L)	5 items 1 VAS	1-3 (items) 0-100 (VAS)	0-1 (items) 0-100 (VAS)	Yes, fees may apply	http://www.euroqol.org/
		Short Form Health Survey 12 (SF-12)	12	Varying number from 2 to 6	0-100 (physical and mental component summary scores)	Yes, fees may apply	https://campaign.optum.com/optum-outcomes/what-we-do/health-surveys/sf-12v2-health-survey.html
Work	Work ability	A single-item question	Are you working at a physically less demanding job now because of your back and/or leg pain?	Yes, No, NA		No, no fees
	Work productivity	Two single-item questions	What is your current work status?	Working full-time, working part-time, seeking employment not working by choice, unable to work because of problems other than my back pain, unable to work due to back pain		No, no fees
			How long after you received treatment for low back pain did you return to work?	<3 months, 3–6 months, 6–9 months, 9–12 months, 1–2 years, >2 years, NA		No, no fees
Psychological functioning	Depression	Hospital Anxiety and Depression Scale (HADS) subscale	7	0–3	0–21	Yes, fees may apply	https://eprovide.mapi-trust.org/instruments/hospital-anxiety-and-depression-scale
	Anxiety	Hospital Anxiety and Depression Scale (HADS) subscale	7	0–3	0–21	Yes, fees may apply
Pain interference		Pain Interference subscale of the Brief Pain Inventory (BPI-PI)	7	0–10	0–10	Yes, fees may apply	https://www.mdanderson.org/research/departments-labs-institutes/departments-divisions/symptom-research/symptom-assessment-tools/brief-pain-inventory.html
		Pain Interference items of the Multidimensional Pain Inventory (MPI-PI)	9	0–6	0–54	Yes, no fees	http://gpsupport.workcover.wa.gov.au/content/uploads/sites/2/2015/07/west_haven_yale_multidimensional_pain_inventory.pdf

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Evidence and methods in back pain research Clinical guidelines for low back pain: A critical review of consensus and inconsistencies across three major guidelines The prevalence, risk factors, prognosis and treatment for back pain in children and adolescents: An overview of systematic reviews Causal mechanisms in the clinical course and treatment of back pain Developing implementation science to improve the translation of research to address low back pain: A critical review Decision support tools in low back pain

Stay updated, free articles. Join our Telegram channel

Join