Abstract
A literature review of the nine most widely used, condition-specific, self-administered assessment questionnaires for low back pain has been undertaken. General and historic aspects, reliability, responsiveness and minimum clinically important difference, external validity, floor and ceiling effects and available languages were analysed for the nine most-used outcome tools. When considering which condition-specific measure to employ, the present overview on assessment tools should provide the necessary information to define the technical aspects of the nine questionnaires. These criteria, however, are only part of the consideration. In part II the construction of these scales in relationship to the measurement domains will be evaluated.
Keywords: Spine, Outcome, Assessment, Review, Low back pain
Introduction
The evaluation of therapies for low back pain requires consideration of a number of variables. A full evaluation is recommended to include a condition-specific disability measure, a general health measure (e.g. EQ-5D [22], WHODAS II [65], SF36 [67]), a pain measure (e.g. VAS [34]), a satisfaction measure and a measure of employment [7]. The present review is concerned with condition-specific measures.
Among a broad range of available tools, only a limited number of measurement instruments are generally known and frequently used. The application of a frequently used tool allows comparisons to be made between the study group and other populations. All currently available measures have flaws or restrictions regarding their construction, validation or application. In the absence of an ideal instrument [8], the choice of a commonly used measurement tool may be considered reasonable.
Apart from its currency of application, however, the availability, validity and responsiveness of the outcome scale are important criteria in choosing the appropriate measuring tool [45, 52, 58]. Other important, but sometimes underestimated criteria, are the characteristics of the individual questions and answers in the questionnaire [50, 58, 62]. Does the question focus on one domain or several domains? Are the sentences and questions unambiguous? Do they address performance or capacity? Are the offered answers precise and clear? Does the scoring system allow separate assessment of subscales?
The goal of this review is to compare the most common condition-specific assessment instruments for low back pain and to analyse each of these instruments separately, addressing these issues. This first paper addresses internal consistency, reliability, external validation, floor and ceiling effects, responsiveness and availability. In part II, the questionnaires will be discussed in relation to their face and construct validity and domains of measurement. Following this analysis, we aim to provide a basis for choice when considering a self-administered assessment tool in the field of low back problems.
Methods
Using the Internet, various medical search sites were addressed in order to find condition-specific, self-administered outcome measurements used in spine surgery (http://www.BioMedNet.com, http://www.PubMed.com, http://www.aaos.org, http://www.outcomes-trust.org, http://www.qlmed.org, http://www.hoi-stratishealth.org, http://www.isoquol.org, http://www.update-software.com/Cochrane, http://www.mapi-research-inst.com). Out of 82 condition-specific outcome scales, the nine most commonly used in the literature were selected (Table 1).
Table 1.
Characteristics of the nine chosen condition-specific questionnaires
| Characteristic | ODI | RMDQ | LBOS | QBPDS | MVAS | ALBDS | NASS LSO | LBPRS | WDI |
|---|---|---|---|---|---|---|---|---|---|
| Existing since | 1980 | 1982 | 1992 | 1995 | 1982 | 1994 | 1996 | 1995 | 1984 |
| Items | 10 | 24 | 13 | 20 | 15 | 19 | 62 | 21 | 9 |
| Completion time | 5 | 10 | 5 | 10 | 10 | 10 | 21 | 15 | 5 |
| Med Line used | 117 | 103 | 23 | 30 | 29 | 71 | 21 | 14 | 64 |
To quantify the use of each of these outcome tools, the number of studies evaluating or using it were determined. Measures used ten times or less in the literature were excluded from further analysis. General measures such as the SF36 [67] were not included in the search. Isolated pain measures were also excluded.
The nine scales examined were: (1) the Oswestry Disability Index (ODI), (2) the Roland–Morris Disability Questionnaire (RMDQ), (3) the Low Back Outcome Score (LBOS), (4) the Quebec Back Pain Disability Scale (QBPDS), (5) the Million Visual Analogue Scale (MVAS), (6) the Aberdeen Low Back Disability Scale (ALBDS), (7) the NASS Lumbar Spine Outcome Assessment Instrument (NASS LSO), (8) the Low Back Pain Rating Scale (LBPRS), and (9) the Waddell Disability Index (WDI).
The construction of outcome scales requires a number of considerations [58]. The measures were examined for general characteristics, reliability, internal consistency, responsiveness, correlation with other measures, floor and ceiling effects and available languages.
General characteristics
The population from which the score was developed, number of items, items scored, whether the measure produces one single score or is divided into subscales and a brief description of the domains are provided.
Reliability
There are a variety of ways of examining the reproducibility of a measure administered on different occasions. Test–retest reliability is the most important. It is measured best by using tests of agreement such as the kappa test [6, 45, 58]. The Pearson correlation coefficient [6] is a measure of correlation and, although commonly used, is a less reliable measure. Pearson correlation values should exceed 0.8 and kappa values should exceed 0.5. Another measure is the Bland–Altman plot [6]. This describes the spread of the score values within the same individuals between the test and the retest examination and provides a 95% confidence interval.
Internal consistency
Measures of internal consistency are based on a single administration of the outcome measure. If the outcome measure has a relatively large number of items addressing the same dimension, such as measures of physical function, it is reasonable to expect that scores on each item would be correlated with scores on all other items. Thus, if the internal consistency is low, the different items should not be summed, because they measure different domains. Internal consistency is predominantly measured by Cronbach’s alpha correlations [14]. Values above 0.8 are acceptable.
Responsiveness to changes
The minimum clinically important difference (MCID) is the value of the change in the score which equates to the smallest change in the condition of interest the patient can detect. Responsiveness can also be evaluated using the receiver operating characteristic (ROC) curve which is constructed by calculating the sensitivity (true positive rate) and specificity (true negative rate) of the cut-off point for each of the possible score values [58]. An index of the “goodness” of the questionnaire is the area under this curve (AUC), which is usually abbreviated as D’. A poorly discriminating questionnaire has an area of 0.5 and a perfect test has an AUC of 1.0 [58].
External validation
Comparison of a new score with existing scores allows assessment of its performance against known measures, particularly in selection of measurement domains, responsiveness and floor and ceiling effects.
Floor and ceiling effect
Floor and ceiling effects describe the percentage of subjects which have maximal or minimal points in the score [7, 58]. Here the measure is inefficient in discriminating between subjects. A similar problem occurs when the results are skewed in a certain region. Floor and ceiling effects may be observed if a measure developed in one population, e.g. severely disabled subjects in a pain clinic, is used in a very different population, e.g. attenders in primary care.
The nine questionnaires
The Oswestry Disability Index (ODI)
General characteristics
The score was initiated in 1976 in a specialist referral clinic with a large number of chronic low back pain patients. The index was designed as a measure for both assessment and outcome. Version one was published and validated in 1980 [24] using a sample of 25 patients suffering from acute low back pain. A larger validation was published in 1994 by Stratford on a population with musculoskeletal LBP [60]. The ten items can be completed in approximately 5 min and scored in less than 1 min. No sub-scoring is reported. The administration is easy.
The ODI was further developed and validated and is now available in version 2.0 [23]. A modified ODI is used in the NASS [15] questionnaire (see below) and another modification was used in England (computer interview). Version 2.0 is now recommended [1, 53] and no permission for its use is required. The questionnaire focuses on abilities (personal care, lifting, walking, sitting, standing, sleeping, sex life, social life and travelling in combination with pain) and on pain level. However, important items considering the ability to work, need for help or items about environmental factors are not included (Table 2).
Table 2.
Content and question-and-answer characteristics of the chosen evaluation tools using the ICF classification
| Characteristic | ODI | RMDQ | LBOS | QBPDS | MVAS | ALBDS | NASS LSO | LBPRS | WDI |
|---|---|---|---|---|---|---|---|---|---|
| Assessment of | |||||||||
| Pain | 1a | 1a | 1a | – | 1a | 1a | 1a | 1a | – |
| Sleep | 1a | 1a | 1a | 1a | 1a | 1a | 1a | 1a | 1a |
| Self-care | 2 | 2 | – | 2 | – | 2 | – | – | – |
| Walking | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 |
| Sitting | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 |
| Standing | 2 | 2 | – | 2 | 2 | 2 | 2 | – | 2 |
| Lifting | 2 | 2 | – | 2 | – | – | 2 | 2 | 2 |
| Sex life | 2 | – | 2 | – | – | – | 2 | – | 2 |
| Travelling | 2 | – | 2 | – | – | – | 2 | – | 2 |
| Social life | 3 | – | – | – | 3 | – | 3 | 3 | 3 |
| Work | – | 2 | 2 | – | 2 | – | – | 2 | – |
| Dressing | – | 2 | 2 | – | – | – | 2 | 2 | 2 |
| Sport | – | – | 2 | 2 | – | 2 | – | – | – |
| Stairs | – | 2 | – | 2 | – | – | – | 2 | – |
| Housework | – | 2 | 2 | 2 | – | 2 | – | 2 | – |
| Resting | – | 2 | 2 | – | – | 2 | – | – | – |
| Appetite | – | 1a | – | – | – | – | – | – | – |
| Need of help | – | 2 | – | – | – | – | – | – | – |
| Psychological factors | – | – | – | – | – | – | – | 1a | – |
| Need of treatment | – | – | X.9 | – | – | – | – | – | – |
| Need of medicaments | – | – | X.8 | – | X.8 | X.8 | – | – | – |
| Car driving | – | – | – | 2 | – | – | – | 2 | – |
| Throwing | – | – | – | 2 | – | – | – | – | – |
| Stiffness | – | – | – | – | 1a | – | – | – | – |
| Twisting | – | – | – | – | 1a | – | – | – | – |
| Bending | – | – | – | – | – | 1a | – | – | – |
| Loss of feelings | – | – | – | – | – | 1b | 1b | – | – |
| Leg weakness | – | – | – | – | – | 1b | 1b | – | – |
| Special features | – | – | A VAS is used for pain, but not for the rest | – | – | – | SF36 and health survey item 18–28 are included | VAS for pain | – |
| Question’s targets | 1 clear target | 1–2 targets | 1 target | 1–2 targets | 1–2 targets | 1–2 targets | 1–2 targets | 1–2 targets | 1–2 targets |
| Answer levels | 2 or more | Yes/no | 1–2 | 1 | 1 | 1 (some more) | 1–2 | 1 | 1 |
| Answer type | Text | Yes/no | Text | Scale | Scale | Text | Text | Scale | Yes/no |
| Answer scale | Scaled text | Yes/no | Scaled text | 0–5 | 0–11 | Scaled text | Scaled text | 0–3 | Yes/no |
| Scoring points | 0–50 | 0–24 | 0–75 | 100 | 150 | 0–100 | 0–102 | 0–90 | 0–9 |
Reliability
Reliability has been tested in the literature (Table 3) using only test–retest correlations (Pearson correlation) in a small population (n=22) [26, 30, 38]. Kappa values or Bland–Altman plots are not available.
Table 3.
Technical data of the nine condition-specific questionnaires in comparison
| ODI | RMDQ | LBOS | QBPDS | MVAS | ALBDS | NASS LSO | LBPRS | WDI | |
|---|---|---|---|---|---|---|---|---|---|
| Floor effect | Yes [49] | Unknown | 0% [27] | Unknown | Unknown | Unknown | Unknown | Unknown | 16.1% [27] |
| Ceiling effect | Unknown | Present [59] | 4.6% [27] | Unknown | Unknown | Unknown | Unknown | Unknown | 1/0% [27] |
| Retest Pearson correlation | 0.83 [29] | 0.91 [55] | 0.92 [32] | 0.67 [26] | 0.92 [46] | 0.94 [56] | 0.85–0.99 [15] | – | 0.73–0.9 |
| 0.78 [26] | 0.88 [35] | – | – | – | – | – | – | – | |
| 0.91 [38] | 0.83 [17] | – | – | – | – | – | – | – | |
| Kappa values | – | – | 0.51–0.86 [32] | – | – | – | 0.85–0.97 [15] | – | 0.27–1.0 |
| Bland–Altman plot | – | – | 11.6 [32] | – | – | 26.6 [56] | – | – | – |
| ICC | 0.91/0.9 [26] | 0.53 [16] | – | 0.92/0.55 [26] | – | – | – | – | 0.74 [16] |
| 0.84 [16] | 0.86 [59] | – | 0.84 [16] | – | – | – | – | – | |
| 0.89 [1] | 0.91 [38] | – | 0.92 [38] | – | – | – | – | – | |
| Cronbach’s alpha | 0.76 [25] | 0.89/0.92 [33] | 0.85 [32] | 0.96 [38] | 0.93 [46] | 0.8 [56] | 0.88 [15] | 0.98 [44] | – |
| 0.81 [38] | 0.91 [61] | – | – | – | – | – | – | – | |
| 0.77/93 [33] | – | – | – | – | – | – | – | – | |
| Spearman rank analysis | 0.59 [60] | 0.6 [60] | – | – | – | – | – | – | – |
| ROC with an AUC value of | 0.76 [4] | 0.78/0.74 | – | 0.87/0.74 | – | – | – | – | 0.76 [16] |
| 0.94 [26] | 0.77 [16] | – | 0.74 [16] | – | – | – | – | 0.87 [26] | |
| 0.78 [16] | – | – | 0.93 [38] | – | – | – | – | – | |
| Norman–Steiner sensitivity analysis | 0.15 [38] | 0.2 [38] | – | 0.26 [38] | – | – | – | – | – |
| Standardized response mean | 0.36 [59] | 0.5 [59] | – | – | 0.49 [16] | – | – | – | 0.35 [16] |
| 0.52 [16] | 0.55 [16] | – | – | – | – | – | – | – | |
| MCID | 16 [24] | 5.2 [53] | 7.5 [32] | 14 [38] | – | – | – | – | 2.8 [16] |
| 15 [16] | 8.6–9.5 [16] | – | 19 [16] | – | – | – | – | – | |
| 6 [26] | 4 [59] | – | 15 [26] | – | – | – | – | – | |
| Compared with | RMDQ [23, 25, 45] | ODI [16, 33, 42] | ODI [27, 63] | RMDQ [16, 38] | ODI [46, 70] | SF36 [56] | SF36 [15] | ODI [11] | RMDQ [16] |
| QBPDS [16, 26, 38] | QBPDS [16, 38] | WDI [27] | ODI [16, 26, 38] | RMDQ [16] | – | – | – | QVPDS [16] | |
| WDI [16, 63] | WDI [16] | SF36 [63] | SF36 [16, 38] | WDI [16] | – | – | – | ODI [16] | |
| LBOS [27] | SF36 [16] | – | – | – | – | – | – | LBOS [27] | |
| SF36 [16] | – | – | – | – | – | – | – | SF36 [16] | |
| MVAS [70] | – | – | – | – | – | – | – | MVAS [16] | |
| LBPRS [11] | – | – | – | – | – | – | – | – |
Internal consistency
Version 2.0 of the ODI shows an acceptable internal consistency, with Cronbach’s alpha between 0.76 and 0.8716 [25, 33, 38].
Responsiveness
The ROC was shown to be between 0.76 and 0.78 [4, 16, 60], which is acceptable (but not as good as the Roland–Morris score [16, 60]). Several studies show a good correlation between the severity of pain and disability on the one hand and patients satisfaction on the other [26, 60, 63], even in patients with cervical problems [70]. The ODI is able to detect clinical change 3 weeks after surgery [56] and the MCID varies between 6 and 16 points [4, 16, 26, 60]. Using mean values, Beurkens [4] adopted the method advocated by Cohen [12] to determine the effect size and obtained values of 0.8 for patients who had improved and 0.04 for patients whose condition had not changed. However, Taylor found that the ODI was more sensitive to patients who had improved and less sensitive for patients whose condition had remained unchanged [63].
External validation
The ODI was compared with the RMDQ [16, 33, 42], the LBOS [27], the QBPDS [16, 26, 37, 38], the LBPRS [16], the MVAS [70] and the SF36 [16] (see below).
Floor and ceiling effects
Floor effects have been demonstrated in non-surgical patients [49], demonstrating that the ODI is less sensitive in less disabled patients.
Languages
The ODI is validated in English [24], German [2], French [21], Finnish [29] and Greek [9] (Table 4). Translations in several other languages do not appear to be validated.
Table 4.
Available languages for all the nine condition-specific outcome scores
| ODI | RMDQ | LBOS | QBPDS | MVAS | ALBDS | NASS LSO | LBPRS | WDI | |
|---|---|---|---|---|---|---|---|---|---|
| Original language | English [24] | English [54] | English [27] | English [38] | English [46] | English [56] | English [15] | Danish [44] | English [66] |
| Validated languages | Finnish [29, 30] | French [13] | – | Dutch [57] | – | Chinese [43] | German [51] | – | – |
| French [21] | German [68] | – | French [38] | – | – | Italian [48] | – | – | |
| German [2] | Greek [10] | – | – | – | – | – | – | – | |
| Greek [9] | Portuguese [47] | – | – | – | – | – | – | – | |
| – | Spanish [40] | – | – | – | – | – | – | – | |
| – | Swedish [35] | – | – | – | – | – | – | – | |
| – | Turkish [41] | – | – | – | – | – | – | – | |
| Unvalidated languages | Danish | Flemish | German | – | – | – | – | – | French |
| Dutch | Czech | Spanish | – | – | – | – | – | – | |
| Norwegian | Italian | – | – | – | – | – | – | – | |
| Spanish | Polish | – | – | – | – | – | – | – | |
| Swedish | Romanian | – | – | – | – | – | – | – |
The Roland–Morris Disability Questionnaire (RMDQ)
General characteristics
The RMDQ was derived from the Sickness Impact Profile, of which 24 out of 136 items were selected. The questionnaire was designed as a self-reporting measure for both assessment and outcome, and was published in 1983 [54]. First validation was done using a LBP population in a general practice treated with pain medication. In the first version a six-point pain rating scale was included. Recently, the authors recommended the use of the pain scale of the SF36 [53]. Despite several published modification proposals, [20, 49, 61, 64] the original version of the RMDQ is favoured by an international expert group [18].
The questions in the RMDQ focus consistently on disabilities related to the back and the answers are dichotomous: yes/no. This might cause subtle changes in the functional abilities of patients to remain undetected. The questionnaire can be completed in a maximum of 5 min and an un-weighted score can be calculated in less than 1 min. No sub-scoring is reported and administration is very easy. The questions deal with body functions (pain, sleeping and appetite) as well as activities (self care, walking, sitting, standing, lifting, work, dressing, stairs, housework and resting), but no environmental questions are asked (social life, need of help etc.) (Table 2).
Reliability
The test–retest reliability using Pearson correlation lies between 0.81, 0.88 and 0.91 [17, 35, 54]. Kappa values or Bland–Altman plots are not available.
Internal consistency
The RMDQ seems to be a reliable evaluation tool with an internal consistency between 0.77 and 0.93, [33, 61] and with equal [42, 52] or even slightly superior [33] consistency in comparison with the Oswestry questionnaire (Table 3).
Responsiveness
The RMDQ seems to detect changes over time slightly more sensitively than the Oswestry scale [16, 42], provided that the initial score is in the range between 4 and 20 [59]. Beaton suggests that a change of 5 points in the RMDQ can be interpreted as a significant improvement or decline in patients’ outcome [3]. This value comes close to the figure of Davidson [16], who defined the MCID to be 5.2 points.
In the French version, Dionne finds a poor relation between the RMDQ and the work status of patients [19]. When comparing the ODI with the RMDQ, Baker et al. [1] found that the RMDQ is more sensitive in patients with mild disabilities and the ODI discriminates better in more severe disabilities. This means that the RMDQ and the ODI are not linearly related.
In contrast, Yang [69] and Leclaire [42] found a linear but moderate Pearson correlation coefficient of r=0.72 and 0.64, respectively. Hsieh [33] compared the two measures in four treatment groups and found a poor correlation (r=0.51–0.61) at the initial visit and a high correlation after treatment (r=0.85–0.92). The latter suspected that both instruments may have a different performance at different levels of pain. Davidson [16] found similar responsiveness when comparing the RMDQ with the SF36, the ODI, the QBPDS and the WDI—but the RMDQ showed an insufficient reliability (Table 3).
Floor and ceiling effects
A ceiling effect is reported [59].
Languages
The RMDQ is translated into several languages (Table 4). Many of these translations do not appear to be validated.
The Low Back Outcome Score (LBOS)
General characteristics
The questionnaire was first published in 1992 and tested using 274 consecutive patients with low back pain [27]. The score was designed as a self-reporting measure for both assessment and outcome. The 13 items can be completed in approximately 5 min and scored in about 1 min. The answering possibilities of each item are scaled. For pain, an eleven-point scale ranging from “no pain” to “maximum pain possible” is used. The other items use a four-point scaled text. The total score gives different weights to different questions. For example, normal abilities in sport score 9 points, normal abilities in sex life score 6 points and normal abilities in dressing score 3 points. Although this weighting has to be conducted with care, the administration is easy. All aspects of the ICF classification [28] are considered, which is the strength of this questionnaire (for the questionnaire’s content see Table 2). Slight extensions to the questionnaire have been reported [32].
Reliability
The test–retest reliability is high, with a Pearson correlation of 0.92 and kappa values between 0.51 and 0.8645. The Bland–Altman plot shows that test–retest scores can change by up to 11.6 points [32]. This scale is the only one to have been subjected to this more stringent criterion.
Internal consistency
Holt reported a good internal reliability with a Cronbach’s alpha of 0.85 in an English population and 0.79 in an Australian population [32].
Responsiveness
The MCID is reported to be 7.5 points [36].
External validation
The LBOS correlates well with the ODI (r=0.87), the WDI (r=0.74) and the Waddell physical impairment rating (r=0.63) [32]. Taylor et al. [63] made a comparison between the LBOS, the SF36 and the ODI, and found a Spearman rank correlation of 0.64 for the ODI, 0.62 for the LBOS, 0.56 for the SF36 PF and 0.54 for the SF36 PCS. They demonstrated that the condition-specific questionnaires (ODI and LBOS) showed a greater overall responsiveness, but the SF36 PF was more sensitive to changes than both of the specific questionnaires.
Floor and ceiling effects
Values for floor and ceiling effects have been defined in a non-surgical population by Greenough [27], and are 0%/0.8% respectively in a compensated population and 0%/8.3% in a non-compensated low back pain population.
Languages
The LBOS has been validated in English [27, 32] and translated into German and Spanish.
The Quebec Back Pain Disability Scale (QBPDS)
General characteristics
The QBPDS was validated on a back pain population and published in 1995 [38]. The questions were designed using a conceptual model. Item selection was done using factor analysis for 46 disability items. Twenty items were selected and tested for reliability (Table 2).
The QBPDS measures only functional disability (self care, walking, sitting, standing, lifting, sport, stairs and housework) and sleep. Pain has to be evaluated with other tools. Items about social life, sex life and need for help are not included. Nevertheless, the items give a comprehensive view of the patient’s disabilities because questions about easy as well as more difficult functional abilities are asked. The questionnaire can be filled out in about 5–10 min and scored in about 2 min. Sub-scores are not reported and administration is easy.
Reliability
Test–retest reliability was analysed using an intraclass correlation coefficient (ICC) (0.92) [38] (Table 3) but kappa values or Bland–Altman plots are not available.
Internal consistency
Internal consistency was measured using the Cronbach’s alpha (0.96) [38] and is satisfactory.
Responsiveness
The MCID lies between 14 and 19 points [16, 37]. The Norman-Steiner sensitivity coefficient [38] was 0.19 after a six month period, and the sensitivity in relation to a self-rated change of disability was good.
External validation
Kopec [38] correlated the questionnaire in French and English with the RMDQ (r=0.81 and 0.72), ODI (r= 0.83 and 0.77), and SF36 physical functioning (r=0.77 and 0.67), and against a single-item pain scale with a seven-point Likert scale (r=0.54 and 0.51). Fitz [26] found that the modified ODI showed higher levels of test–retest reliability and responsiveness compared with the QBPDS. The MCID was approximately 15 points for the QBPDS (and approximately 6 points for the modified ODI). Davidson [16] found similar responsiveness when comparing the QBPDS with the SF36, the ODI, the RMDQ and the WDI, but the QBPDS showed a better reliability than the RMDQ and the WDI (Table 3).
Floor and ceiling effects
Values for floor and ceiling effects are not available in the literature.
Languages
The QBPDS is validated in English, French [38] and Dutch [57] (Table 4).
The Million Visual Analogue Scale (MVAS)
General characteristics
The questionnaire was first published in 1982 and validated on patients with chronic low back pain [46]. The 15 items focus on body functions (pain, sleep, stiffness and twisting), on activities (walking, sitting, standing and work) and on social life. Questions on self care, sex life lifting, housework etc. are not included. The questionnaire offers a 100 mm visual analogue scale as the answering method. Its administration is easy, the completion time is around 5–10 min and scoring requires 2–3 min.
Reliability
The Pearson correlation is 0.84–0.94 [46] (Table 3). Kappa values or a Bland–Altman plot are not available.
Internal consistency
Internal consistency shows a high Cronbach’s alpha of 0.93 [46].
Responsiveness
The MCID or ROC are not available.
External validation
Zoega found a good correlation and a similar reliability between the MVAS and the ODI [70].
Floor and ceiling effects
Values for floor and ceiling effects are not available in the literature.
Languages
The Aberdeen Low Back Disability Scale (ALBDS)
General characteristics
The ALBDS was first published by Ruta in 1994 [56] and validated using a random sample of the general population. The 19 questions focus on body functions (pain, sleep and bending) and activities (self care, walking, sitting, standing, sport, housework and resting), and include questions about body structures (loss of feeling, leg weakness). Answer categories vary between three and six categories. Questions on environmental factors, lifting, sex life, work, dressing etc. are not included. The questionnaire is easy to administer, can be completed within 5–10 min and scoring can be done within 3 min. Because of the various answering scales, the ALBDS gives variable weights to the different questions.
Reliability
The ALBDS showed an acceptable test–retest correlation of 0.94 (Pearson correlation) [56]. Kappa values or Bland–Altman plots are not available.
Internal consistency
Internal consistency is acceptable with a Cronbach’s alpha value of 0.8 (Table 3) [56].
Responsiveness
The MCID or ROC values are not available. The ALBDS is able to detect clinical change 2 weeks after surgery [56]. Ruta reports that the ALBDS is more sensitive than the SF36 regarding the measurement of the local outcome [56].
External validation
Ruta compared his ALBDS with the SF36 and found a rather low correlation between these two instruments (e.g. r=0.56 for SF36 physical functioning) [56]
Floor and ceiling effects
Values for floor and ceiling effects are not available in the literature.
Languages
The ALBDS is validated in English [56] and Chinese [43]. (Table 4).
The NASS Lumbar Spine Outcome Assessment Instrument (NASS LSO)
General characteristics
The NASS LSO was first published by Daltroy in 1996 [15] and is based on a consensus of the North American Spine Society. Validation was done in a cross-sectional study of a convenience sample of lumbar spine patients (n=136). The NASS LSO questionnaire is made for assessment and outcome measurements. It considers all aspects of the ICF classification, e.g. demographic data (age, sex, race, education and insurance information), medical history (co-morbidities, past surgeries etc.), body functions (pain, neurogenic symptoms etc.) and employment history.
The questionnaire construct includes the SF36, a modified ODI and a modified employment assessment published by Bigos [5]. For follow-up assessment the questionnaire is slightly modified. It takes 20–25 min to fill in the form and the scoring is complex. To score the SF36, a special algorhythm is used. Sub-scores are extractable (modified ODI, SF36, pain and disability scale, neurogenic symptoms scale, job exertion scale, expectation and satisfaction scale). Although the NASS LSO is not a condition-specific questionnaire in a strict sense, we included this tool in our analysis because it contains condition-specific measures.
Reliability
The reliability tests for the condition-specific parts of the NASS LSO show a test–retest reliability (Pearson correlation) of 0.96 for pain and disability and 0.81 for neurogenic symptoms. The ICC is 0.97/0.85 for the two subscales [15].
Internal consistency
Cronbach’s alpha is 0.93 in both measures [15].
Responsiveness
The MCID was not found in the literature.
External validation
The NASS LSO pain and limitation scale correlates with other measures of the same phenomena (pain VAS r=0.84, SF36 pain subscale r= 0.66 and SF36 physical limitation subscale r=0.75). The neurogenic symptoms correlated on a lower level with the three mentioned variables (r=0.60, 0.43 and 0.49 respectively) [15]. Even though the ODI is included, a correlation between the NASS LSO disability and pain scale in comparison with the original ODI could not be found in the literature.
Floor and ceiling effects
Values for floor and ceiling effects are not available in the literature.
Languages
The NASS LSO is validated in English [15], German [51] and Italian [48] (Table 4).
The Low Back Pain Rating Scale (LBPRS)
General characteristics
The LBPRS was first published and validated on different populations by Manniche in 1994 [44]. It consists of a disability index with 15 items (walking, sitting, lifting, work, dressing and car riding), and a pain assessment index with six questions. Questions about self care, lifting, standing, sex life and sport are not included. It can be filled out in about 15 min and scored within 3–5 min. Sub-scores for low back pain, leg pain, disability and physical impairment are extractable.
Reliability
Data on test–retest reliability are not available.
Internal consistency
Internal consistency was calculated using Cronbach’s alpha with values between 0.89 and 0.95 for the sub-scores and 0.98 for the entire LBPRS [44].
Responsiveness
The MCID was not found in the literature.
External validation
Christensen et al. [11] could show a correlation with r=0.82 between the ODI and the LBPRS 18 months after therapy. Despite detailed clinical assessment of the patients involved, the clinical results were not correlated with the values of the two outcome measures.
Floor and ceiling effects
Values for floor and ceiling effects are not available in the literature.
Languages
The LBPRS is available in English and validated in Danish [44] (Table 4).
The Waddell Disability Index (WDI)
General characteristics
The WDI was first published in 1984 [66] and validated on a chronic low back pain population. It is a brief nine-item scale focussing on disabilities (walking, sitting, standing, lifting, sex life, travelling and dressing), on body functions (pain, sleep) and on social life. Questions about work, self care and sports are not included. The questionnaire is easy to administer, can be completed in 5 min and scored in about 1 min.
Reliability
The test–retest reliability of the different items show kappa values between 0.27 (unacceptable low) and 1.0 (very high) [27].
Internal consistency
Internal consistency seems to be rather low with an ICC of 0.74 [16], and Cronbach’s alpha values are not available.
Responsiveness
The ROC is acceptable [16, 26] and the MCID is 2 [16]. The WDI is able to detect clinical change 4 weeks after surgery [56].
External validation
Davidson analysed the WDI against the RMDQ, ODI and the QBPDS [16] and found a lower ICC in the WDI (0.74) than in the ODI (0.84) but a higher ICC in the WDI than in the RMDQ (0.53). The ROC was similar in all four scores but the standardized response mean was lowest in the WDI (0.35). The MCID was similar in all four scores.
Floor and ceiling effects
Values for floor and ceiling effects have been defined in conservatively treated patients by Greenough [27] and are 2.2%/5.9% in a compensation group and 0% and 27% in a non-compensation group.
Languages
The WDI is validated in English [66] and Spanish. A French translation is available, but has not been validated [31] (Table 4).
Discussion
If a specific outcome questionnaire has to be chosen, five main aspects should be considered: First: Is the questionnaire reliable? Second: Is it responsive? Third: Can the results be compared with the literature? Fourth: Is the questionnaire available in the target language? And fifth: Is the questionnaire easy to complete and to score?
1. Reliability
The reliability of a questionnaire is the most important quality a questionnaire has to fulfil. All the nine mentioned questionnaires went through a validation process. The LBOS [27] was particularly stringently validated and the QBPDS [26, 38] and the ALBDS [56] have also passed a serious validation process. The ODI and the RMDQ went through several validation processes and have been used for comparison in many studies. Although Davidson found an insufficient ICC for the RMDQ [16], this was in contradiction to other studies [37], and all the nine questionnaires reviewed may be considered to have passed the minimum acceptable validation standard.
2. Responsiveness in the study population
If it is important to know how subjects have improved following an intervention, then this figure is crucial. Figures are available for the ODI [4, 16, 21, 26], the RMDQ [16, 53, 59], the LBOS [32], the QBPDS [16, 26, 38], and the WDI [16]. For the other questionnaires the MCID was not available.
3. Comparability
Comparing the results with other studies is one of the important goals in quality assessment. In an optimal study design, patients are randomised in two groups (treatment group/control group). This procedure is sometimes not possible and the subjects have to be matched with a control group from the literature. In this case, the same questionnaire as used in the literature must be chosen. A broad comparability is assured if the Oswestry [24] or Roland–Morris scale [54, 55] is used. Nevertheless, this comparability should not be overestimated because of differences in culture, patients’ collectives etc.. If randomisation within the study group is possible, then comparability of the results with the literature is less crucial. In this situation, the LBOS offers a short and comprehensive measure.
4. Availability
For non English speaking countries the availability of a questionnaire is an important factor. The RMDQ followed by the ODI have the largest number of translations. Target languages are defined in Table 4. A more scientific approach is to define the optimal questionnaire using the criteria 1–3. If the optimal questionnaire is not available in the target language, it should be translated and validated before use.
5. Ease of administration
This issue is of some importance in self-administered questionnaires, to avoid question fatigue and increase compliance. Easy and short questionnaires like the RMDQ, LBOS, QBPDS, MVAS or the WDI, give fewer opportunities for errors and are easier for data analysis.
Conclusion
There is no gold standard for quality or outcome assessment in low back therapies. In order to define the optimal questionnaire, its reliability and responsiveness in the study population must be considered. Only the ODI, the RMDQ, the LBOS, the QBPDS and the WDI provide the crucial data on the MCID.
The present overview on the reliability and responsiveness of condition-specific assessment tools should provide the necessary information to define the optimal questionnaire to be used in any proposed study. However, the circumstances of use, domains of interest, construct validity of the instrument as well as possible score bias are of crucial importance and will be considered in part II.
References
- 1.Baker D, Pynsent P, Fairbank J (1989) The Oswestry Index revisited. In: Roland M, Jenner J (eds) Back pain: new approaches to rehabilitation and education. Manchester University Press, Manchester, UK, pp 174–86
- 2.Basler Patient Educ Counsel. 1997;31:113. doi: 10.1016/S0738-3991(97)00996-8. [DOI] [PubMed] [Google Scholar]
- 3.Beaton Spine. 2000;25:3192. doi: 10.1097/00007632-200012150-00015. [DOI] [PubMed] [Google Scholar]
- 4.Beurskens Pain. 1996;65:71. doi: 10.1016/0304-3959(95)00149-2. [DOI] [PubMed] [Google Scholar]
- 5.Bigos Spine. 1991;16:1. doi: 10.1097/00007632-199101000-00001. [DOI] [PubMed] [Google Scholar]
- 6.BlandLancet 198683072868172 [Google Scholar]
- 7.Bombardier Spine. 2000;25:3100. doi: 10.1097/00007632-200012150-00003. [DOI] [PubMed] [Google Scholar]
- 8.Bombardier Spine. 2000;25:3097. doi: 10.1097/00007632-200012150-00002. [DOI] [PubMed] [Google Scholar]
- 9.Boscainos P, Sapkas G, Stilianesi E (1999) Clinical relevance of specific parameters isolated within the Oswestry and Roland Morris functional disability index. J Bone Joint Surg Br 81 [Suppl]:239
- 10.Boscainos PJ, Sapkas G, Stilianessi E, Prouskas K, Papadakis SA (2003) Greek versions of the Oswestry and Roland–Morris disability questionnaires. Clin Orthop 40–53 [DOI] [PubMed]
- 11.Christensen Scand J Rheumatol. 1993;22:25. doi: 10.3109/03009749309095107. [DOI] [PubMed] [Google Scholar]
- 12.Cohen Statistical. 1977;power:analysis. [Google Scholar]
- 13.Coste J, Le Parc JM, Berge E, Delecoeuillerie G, Paolaggi JB (1993) [French validation of a disability rating scale for the evaluation of low back pain (EIFEL questionnaire)]. 60:335–341 [PubMed]
- 14.Cronbach Psychometrika. 1951;16:297. doi: 10.1007/BF02289113. [DOI] [PubMed] [Google Scholar]
- 15.Daltroy Spine. 1996;21:741. doi: 10.1097/00007632-199603150-00017. [DOI] [PubMed] [Google Scholar]
- 16.DavidsonPhys Ther 200282811784274 [Google Scholar]
- 17.Deyo J Chronic Dis. 1986;39:897. doi: 10.1016/0021-9681(86)90038-x. [DOI] [PubMed] [Google Scholar]
- 18.Deyo Spine. 1998;23:2003. doi: 10.1097/00007632-199809150-00018. [DOI] [PubMed] [Google Scholar]
- 19.Dionne J Clin Epidemiol. 1997;50:31. doi: 10.1016/s0895-4356(96)00313-7. [DOI] [PubMed] [Google Scholar]
- 20.Dionne Spine. 1999;24:2339. doi: 10.1097/00007632-199911150-00009. [DOI] [PubMed] [Google Scholar]
- 21.Dropsy Rev Rhum. 1994;61:S44. [Google Scholar]
- 22.EuroquolHealth Pol 19901619910109801 [Google Scholar]
- 23.Fairbank Spine. 1995;20:1535. [PubMed] [Google Scholar]
- 24.Fairbank Physiotherapy. 1980;66:271. [PubMed] [Google Scholar]
- 25.Fisher Physiother Theory Pract. 1997;22:61. [Google Scholar]
- 26.Fritz Phys Ther. 2001;81:776. doi: 10.1093/ptj/81.2.776. [DOI] [PubMed] [Google Scholar]
- 27.GreenoughSpine 199217361531553 [Google Scholar]
- 28.Grimby J Rehabil Med. 2001;33:193. doi: 10.1080/165019701750419545. [DOI] [PubMed] [Google Scholar]
- 29.Gronblad Clin J Pain. 1993;9:189. doi: 10.1097/00002508-199309000-00006. [DOI] [PubMed] [Google Scholar]
- 30.Gronblad Clin J Pain. 1994;10:197. doi: 10.1097/00002508-199409000-00005. [DOI] [PubMed] [Google Scholar]
- 31.Guillemin Br J Rheumatol. 1994;33:148. doi: 10.1093/rheumatology/33.2.148. [DOI] [PubMed] [Google Scholar]
- 32.Holt Spine. 2002;27:206. doi: 10.1097/00007632-200201150-00017. [DOI] [PubMed] [Google Scholar]
- 33.HsiehJ Manipulative Physiol Ther 19921541531488 [Google Scholar]
- 34.Huskisson Lancet. 1974;2:1127. doi: 10.1016/s0140-6736(74)90884-8. [DOI] [PubMed] [Google Scholar]
- 35.Johansson Scand J Rehabil Med. 1998;30:139. doi: 10.1080/003655098444066. [DOI] [PubMed] [Google Scholar]
- 36.Kahtri H, Greenough CG (2002) Minimum clinically important difference in the low back outcome score. Society of Back Pain Research, Leeds, UK
- 37.Kopec Spine. 1995;20:1943. doi: 10.1097/00007632-199509000-00021. [DOI] [PubMed] [Google Scholar]
- 38.Kopec Spine. 1995;20:341. doi: 10.1097/00007632-199502000-00016. [DOI] [PubMed] [Google Scholar]
- 39.Kosinski Arthritis Rheum. 2000;43:1478. doi: 10.1002/1529-0131(200007)43:7<1478::AID-ANR10>3.0.CO;2-M. [DOI] [PubMed] [Google Scholar]
- 40.Kovacs Spine. 2002;27:538. doi: 10.1097/00007632-200203010-00016. [DOI] [PubMed] [Google Scholar]
- 41.Kucukdeveci Spine. 2001;26:2738. doi: 10.1097/00007632-200112150-00024. [DOI] [PubMed] [Google Scholar]
- 42.Leclaire Spine. 1997;22:68. doi: 10.1097/00007632-199701010-00011. [DOI] [PubMed] [Google Scholar]
- 43.Leung Spine. 1999;24:961. doi: 10.1097/00007632-199905150-00006. [DOI] [PubMed] [Google Scholar]
- 44.Manniche Pain. 1994;57:317. doi: 10.1016/0304-3959(94)90007-8. [DOI] [PubMed] [Google Scholar]
- 45.McDowell I, Newell C (1996) The theoretical and technical foundations of health measurement. Measuring health. A guide to rating scales and questionnaires. Oxford University Press, pp 10–42
- 46.MillionSpine 198172046214028 [Google Scholar]
- 47.Nusbaum L, Natour J, Ferraz MB, Goldenberg J (2001) Translation, adaptation and validation of the Roland–Morris questionnaire — Brazil Roland–Morris. Braz J Med Biol Res = Revista brasileira de pesquisas medicas e biologicas / Sociedade Brasileira de Biofisica ... [et al.] 34:203–210 [DOI] [PubMed]
- 48.Padua Spine. 2001;26:E344. doi: 10.1097/00007632-200108010-00012. [DOI] [PubMed] [Google Scholar]
- 49.Patrick Spine. 1995;20:1899. doi: 10.1097/00007632-199509000-00011. [DOI] [PubMed] [Google Scholar]
- 50.Porst R (2000) Question wording—Zur Formulierung von Fragebogen-Fragen. ZUMA, Mannheim 2:1–11
- 51.Pose Z Orthop Ihre Grenzgeb. 1999;137:437. doi: 10.1055/s-2008-1037387. [DOI] [PubMed] [Google Scholar]
- 52.RoeseRehabilitation 1996351038767540 [Google Scholar]
- 53.RolandSpine 200025311511124727 [Google Scholar]
- 54.Roland Spine. 1983;8:141. doi: 10.1097/00007632-198303000-00004. [DOI] [PubMed] [Google Scholar]
- 55.Roland Spine. 1983;8:145. doi: 10.1097/00007632-198303000-00005. [DOI] [PubMed] [Google Scholar]
- 56.Ruta Spine. 1994;19:1887. doi: 10.1097/00007632-199409000-00004. [DOI] [PubMed] [Google Scholar]
- 57.Schoppink Phys Ther. 1996;76:268. doi: 10.1093/ptj/76.3.268. [DOI] [PubMed] [Google Scholar]
- 58.Steiner D, Normann G (1995) Health measurement scales. A practical guide to their development and use. Oxford Medical, 2nd edn.
- 59.Stratford Phys Ther. 1996;76:359. doi: 10.1093/ptj/76.4.359. [DOI] [PubMed] [Google Scholar]
- 60.Stratford Phys Ther. 1994;74:528. doi: 10.1093/ptj/74.6.528. [DOI] [PubMed] [Google Scholar]
- 61.Stratford Spine. 1997;22:2416. doi: 10.1097/00007632-199710150-00018. [DOI] [PubMed] [Google Scholar]
- 62.Sudman S, Bradburn N, Schwarz N (1996) Thinking about answers. The application of cognitive processes to survey methodology. Jossey-Bass, San Francisco
- 63.TaylorSpine 199924180510488511 [Google Scholar]
- 64.Tubergen Ann Rheum Dis. 2003;62:140. doi: 10.1136/ard.62.2.140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Underwood Spine. 1999;24:1104. doi: 10.1097/00007632-199906010-00010. [DOI] [PubMed] [Google Scholar]
- 66.WaddellSpine 198492046233713 [Google Scholar]
- 67.Ware Med Care. 1992;30:473. [PubMed] [Google Scholar]
- 68.Wiesinger Spine. 1999;24:1099. doi: 10.1097/00007632-199906010-00009. [DOI] [PubMed] [Google Scholar]
- 69.Yang J Manipulative Physiol Ther. 1983;16:14. [Google Scholar]
- 70.ZoegaEur Spine J 2000913710823430 [Google Scholar]