Validation of the Tswana Versions of the Roland-Morris Disability Questionnaire , Quebec Disability Scale and Waddell Disability Index

INTRODUCTION Low back pain (LBP) has proved to be a considerable challenge in both developed and developing countries and is responsible for a major portion of people staying away from work or visiting a medical practitioner (World Health Organisation, 2003). LBP impacts on the functional status of a patient, interfering with basic activities of daily living such as sitting, standing, walking and other work related activities (Kopec, 2000). According to Beattie & Maher (1997) one of the most important outcomes of physiotherapy management of LBP is to restore normal function. Measurement tools that evaluate functional limitation in patients with LBP, and examine the change in functional status over time are thus important. There are numerous objective tests that can be used by the physiotherapist in the clinical setting. These may give an indication of the outcome of the treatment for LBP. However the functional status of many daily activities may not be directly observed by the physio therapist and will have to form part of the subjective evaluation, assessing them by way of questioning. The only problem with this non standardised way of questioning is that it is unlikely to be very reproducible (reliable) (Streiner & Norman, 1995). Correspondence to: N de Beer Department of Physiotherapy, School of Therapeutic Sciences, Faculty of Health Sciences, 7 York Road Parktown 2193, University of the Witwatersrand, South Africa. Email: debeernic@absamail.co.za A BST R A CT: The use of reliable and valid outcome measures in clinical research as well as clinical practice is very important. Self reported questionnaires are widely used as outcome measures to assess the subjective perception of disability caused by low back pain (LBP). The Roland Morris Disability Questionnaire (RMDQ), Quebec Disability Scale (QDS) and Waddell Disability Index (WDI) have been identified as reliable and valid instruments for assessing disability caused by LPB in English speaking patients. The fact that we do not yet have a validated and published version of the RMDQ, QDS and WDI in a local South African language was the motivation to undertake this study. The three questionnaires were translated, back-translated and tested in a final version for use with Tswana speaking subjects. The questionnaires were tested on one hundred respondents, who met the inclusion criteria, at five hospitals in Tswana speaking areas. Of the one hundred respondents 31 were retested 24 hours later. The Visual Analogue Pain Intensity Scale (VAS-Pain) and Disability Rating Index (DRI) were used as correlation tools. There was moderate correlation between the RMDQ and the DRI (0.74) and the WDI and the DRI (0.63). The correlation between the QDS and DRI was strong (0.85). The RMDQ, QDS and WDI correlated moderately with the VAS-Pain (0.63, 0.68 and 0.74, respectively). The RMDQ, QDS and WDI appeared to be internally consistent scales with Cronbach’s alpha values of 0.92, 0.95 and 0.75, respectively. The RMDQ, QDS and WDI showed excellent test-retest reliability with intra-class correlation coefficient values of 0.93, 0.91 and 0.84, respectively. The results suggest that the Tswana versions of the RMDQ, QDS and WDI validated in this study are easy to understand, valid and reliable instruments for the measurement of functional disability caused by LBP in a Tswana speaking population. Therefore these translated instruments may be useful clinical instruments for collecting standardised data on activity limitations resulting from LBP in a Tswana speaking population.

Validation of the Tswana Versions of the Roland-Morris Disability Questionnaire, Quebec Disability Scale and Waddell Disability Index

INTRODUCTION
Low back pain (LBP) has proved to be a considerable challenge in both developed and developing countries and is responsible for a major portion of people staying away from work or visiting a medical practitioner (World Health Organisation, 2003).LBP impacts on the functional status of a patient, interfering with basic activities of daily living such as sitting, standing, walking and other work related activities (Kopec, 2000).According to Beattie & Maher (1997) one of the most important outcomes of physiotherapy management of LBP is to restore normal function.Measurement tools that evaluate functional limitation in patients with LBP, and examine the change in functional status over time are thus important.
There are numerous objective tests that can be used by the physiotherapist in the clinical setting.These may give an indication of the outcome of the treatment for LBP.However the functional status of many daily activities may not be directly observed by the physiotherapist and will have to form part of the subjective evaluation, assessing them by way of questioning.The only problem with this non standardised way of questioning is that it is unlikely to be very reproducible (reliable) (Streiner & Norman, 1995).

A BST R A CT: The use of reliable and valid outcome measures in clinical research as well as clinical practice is very important. Self reported questionnaires are widely used as outcome measures to assess the subjective perception of disability caused by low back pain (LBP). The Roland Morris Disability Questionnaire (RMDQ), Quebec Disability Scale (QDS) and Waddell Disability Index (W DI) have been identified as reliable and valid instruments for assessing disability caused by LPB in English speaking patients. The fact that we do not yet have a validated and published version of the RMDQ, QDS and W DI in a local South
A frican language was the motivation to undertake this study.The three questionnaires were translated, back-translated and tested in a final version for use with Tswana speaking subjects.The questionnaires were tested on one hundred respondents, who met the inclusion criteria, at five hospitals in Tswana speaking areas.Of the one hundred respondents 31 were retested 24 hours later.The Visual A nalogue Pain Intensity Scale (VA S-Pain) and Disability Rating Index (DRI) were used as correlation tools.There was moderate correlation between the RMDQ and the DRI (0.74) and the W DI and the DRI (0.63).The correlation between the QDS and DRI was strong (0.85).The RMDQ, QDS and W DI correlated moderately with the VA S-Pain (0.63, 0.68 and 0.74, respectively).The RMDQ, QDS and W DI appeared to be internally consistent scales with Cronbach's alpha values of 0.92, 0.95 and 0.75, respectively.The RMDQ, QDS and W DI showed excellent test-retest reliability with intra-class correlation coefficient values of 0.93, 0.91 and 0.84, respectively.The results suggest that the Tswana versions of the RMDQ, QDS and W DI validated in this study are easy to understand, valid and reliable instruments for the measurement of functional disability caused by LBP in a Tswana speaking population.Therefore these translated instruments may be useful clinical instruments for collecting standardised data on activity limitations resulting from LBP in a Tswana speaking population.
the literature three questionnaires were chosen for this study based on the inclusion criteria used by Davidson & Keating (2002).The questionnaires must be able to be patient self-administered, brief and easy to complete, have a clear scoring protocol and measure activity limitation and participation restriction.
The Roland Morris Disability Questionnaire (RMDQ) was selected because it is one of the most widely used low back questionnaires with many studies reporting on its reliability and validity.The Quebec Disability Scale was developed by examining certain measurement properties within a larger group of 48 items and this according to Kopec et al. (1996) produced a 20 item questionnaire with superior measurement properties to those questionnaires developed through pure insight.The Waddell Disability Index is a very brief questionnaire (9 items) and asks specific and direct questions and seemed appropriate for use in this study.
In a recent systematic review done by Grotle et al. (2004) all of the above three questionnaires have been recommended for use, on English speaking patients, without the need for further validation studies.However for researchers and clinicians to use these self reported disability questionnaires in other cultures they will need to be translated and culturally adapted.According to Grotle et al. (2003), the translation must be revalidated to achieve an equivalent questionnaire.The benefit of translating questionnaires is that it allows clinicians and researchers to compare the clinical outcomes of many interventions for LBP by large scale meta-analyses (Davidson & Keating, 2002).The fact that we do not yet have a validated and published version of the RMDQ, QDS and WDI in a local South African language was the motivation to undertake this study.
Reliable and valid Tswana versions of the RMDQ, QDS and WDI, as instruments for the assessment of disability in patients with LBP, have not yet been developed.The purpose of this study then was to translate and culturally adapt the RMDQ, QDS and WDI into Tswana and to establish their validity and reliability for Tswana speaking patients with low back pain.

Research tools
The questionnaires used for translation and cross cultural adaptation in this study were the RMDQ, QDS and WDI (see Appendix A).Translated and crossculturally adapted Tswana versions of the DRI and VAS-Pain were used to test the construct and concurrent validity of the RMDQ, QDS and WDI.

Translation procedure
The translation and cross cultural adaptation of the original English versions of the RMDQ, QDS and WDI into Tswana was carried out in accordance with published guidelines (Beaton et al, 2000) The RMDQ, QDS and WDI were translated into Tswana by two different and independent Tswana speakers.Tswana had to be their home language and they needed to have good English skills.
The translations were compared with one another and with the original English versions.After discussing any discrepancies that may have arisen, consensus was reached and the translated versions were integrated into one common Tswana version for each of the three questionnaires.
Two other Tswana speakers, with good English skills then carried out back-translations of the Tswana versions into English.A bilingual physiotherapist compared the back-translations with each other and with the original English questionnaires and highlighted any gross inconsistencies in the content of the translated versions.The bilingual physiotherapist and one of the original translators then jointly reviewed and fine-tuned the pre-final Tswana versions.
Lastly the pre-final translated questionnaires were tested in a pilot study on ten Tswana speaking subjects.They were briefly interviewed to check what they thought were meant by each question and the chosen response.All the findings were evaluated and the Tswana versions of the RMDQ, QDS and WDI were then finalised.

Research participants
A total of one hundred Tswana speaking subjects complaining of LBP agreed to take part in the study.The study was conducted over a period of two months.During the first month data were collected from Wilmed Park and Sunningdale private hospitals as well as Duff Scott and Westvaal mining hospitals in Klerksdorp.In the second month the battery of questionnaires was tested at Dr George Mukari government hospital in Tswane.Written permission was granted by all of the hospitals.Ethical clearance was obtained from the University of the Witwatersrand Committee for Research on Human Subjects.

Procedure
Subjects who consented to participate in the study were asked to complete a questionnaire booklet, which contained an information sheet, consent form, the Tswana versions of the RMDQ, QDS, WDI, DRI and 100mm VAS-Pain.A brief personal data questionnaire was also completed which was kept separately to maintain confidentiality.A Tswana speaking research assistant was present to read (not explain) each question of the Tswana versions of the questionnaires to those subjects who could not read.The researcher was also present during data collection to ensure the coordination of the study logistics.
Construct validity means that an instrument relates to other measures in a way that is to be expected (Trochim, 2001).The use of the DRI as a correlation tool to establish the construct validity of other disability scales has been supported by the literature (Feise & Menke, 2001, Grotle et al., 2003).To test the construct validity, a significant correlation between the RMDQ, QDS, WDI and the DRI was hypothesized.
Concurrent validity is when a measurement correlates highly with the current performance on some other test (both tests are administered at approximately the same time).If a questionnaire is a valid measure of activity limitation, and if the limitation is due to pain, one would expect a significant correlation of questionnaire scores with self rated pain.Concurrent validity was determined by comparing the scores of the RMDQ, QDS and WDI with the results of the VAS (pain).
Internal consistency indicates the strength of the association between all the items within the test instrument, thus assessing the extent to which items within a scale measure a single underlying trait.Cronbach's alpha was used to determine the internal consistency of the questionnaires.
Test-retest reliability is the consistency or repeatability of scores of a measuring instrument when administered on two different occasions.In this study, 31 subjects with LBP were asked to complete the second questionnaire booklet, containing the four disability questionnaires and the VAS pain scale, 24 hours after they completed the first questionnaire booklet.

Statistical analysis
Internal consistency was measured by means of Cronbach's alpha.Test-retest reliability was determined by using the Intraclass Correlation Coefficient.Construct and concurrent validity were determined by the Pearson Correlation Coefficient.

Demographic data
The demographic data of the study population are shown in Table 1.
Of the 100 subjects who participated in the study 95 were female and only 5 were male.The reason for the majority being female subjects in the study was mainly due to the occupations of the subjects (cleaners, kitchen staff, laundry workers and nurses).The majority of subjects had a grade 10 or higher education level.

Translation and cross cultural adaptation
A few noteworthy difficulties arose during the development of the questionnaires as there were also certain cultural differences that made the direct translation from the original English versions of the questionnaires into Tswana difficult.For this reason some modifications were performed during the translation process.

QDS Question 8:
During the pilot study "Walk a few blocks (300-400m)" was found not to be very well understood and was changed to "To walk for a certain distance (300-400m)".Question 10: "Reach up to high shelves" was not that comprehensible during the pilot study and was changed to "To reach up to the top cupboards".Question 12: "Run one block (100m)" was changed to "To run for a certain distance (100m)".

WDI
Question 1: "30-40 pound suitcase" was dropped as pounds are generally not used in South Africa.The example of a 3-4 year old child was a good enough example of a heavy object and was well understood by the pilot study sample.Question 3: "Travelling in a bus or car..." As many of the pilot study sample make use of public transport the word car was changed to taxi.Thus it was changed to "Travelling in a bus or taxi...".Question 9: "Help required with footwear (tights, socks, tying laces etc.)" was translated as "To be unable to put your shoes on or to tie them yourself".It was agreed that this would be better understood.

Validity
The Pearson correlation coefficient of the Tswana versions of RMDQ, QDS and WDI with the DRI and VAS (pain) is illustrated in Table 2.
Construct validity of the Tswana versions of the RMDQ, QDS and WDI was measured by determining the correlation between them and a Tswana adaptation of the DRI.The QDS showed a strong correlation with the DRI while the WDI and RMDQ had a moderate correlation.
Concurrent validity of the Tswana versions of the RMDQ, QDS and WDI was assessed with their correlation to the VAS pain scale.Moderate correlation was found between the RMDQ, QDS and WDI summed scores and the VAS pain scale.

Internal consistency
Reliability of the Tswana version of the RMDQ estimated by the internal consistency reached an overall Cronbach's alpha of 0.9240.Coefficients ranged from 0.9107 (question 14) to 0.9244 (question 15).For all the 20 items, internal consistency for the Tswana translation of the QDS reached a Cronbach's alpha of 0.9538.Coefficients ranged from 0.9494 (question 15) to 0.9538 (question 20).Internal consistency reliability for the Tswana translation of the WDI reached a Cronbach's alpha coefficient of 0.7505.Coefficients ranged from 0.6966 (question 4) to 0.7473 (question 9).In the concurrent validity analysis, moderate correlation was found among the RMDQ, QDS and WDI summed scores and the VAS pain scale.The correlation between the Tswana version of the RMDQ and the VAS pain scale was 0.63 which is higher than the Persian (0.36) (Mousavi et al., 2006), Norwegian (0.32) (Grotle et al., 2003) and Spanish (0.35) (Kovacs et al., 2002) versions of the RMDQ, respectively, but lower than the German (0.81) (Wiesinger et al., 1999) and Brazilian (0.79) (Nusbaum et al., 2001) versions.The Tswana version of the QDS also showed moderate correlation with the VAS pain scale (0.68) which is similar to the original English version (0.7) (Kopec et al., 1995) and higher than the French (0.45) (Kopec et al., 1995) and Persian (0.46) (Mousavi et al., 2006) versions of the QDS.The Tswana version of the WDI correlated somewhat better with the VAS pain scale (0.74) than the Tswana RMDQ and QDS.Mousavi et al. (2006) argue that although the association between LBP disability scales and pain rating scales is expected to be good, it should not be very high, otherwise it would suggest that the two instruments are carrying identical information.

Internal consistency
The Cronbach's alpha for the Tswana version of the RMDQ of 0.92 is good when compared to previously reported values in the Persian (0.83) (Mousavi et al., 2006),German (0.81) (Wiesinger et al., 1999), Spanish (0.84) (Kovacs et al., 2002), Turkish (0.85) (Kucukdeveci et al., 2001), Greek (0.88) (Boscainos et al., 2003), and Japanese (0.86) (Fujiwara et al., 2003), versions of the RMDQ.In this study, the Cronbach's alpha for the Tswana version of the QDS of 0.95 was higher than the Tswana RMDQ and WDI and similar to the alpha coefficient reported by the developers of the scale (0.96) (Kopec et al., 1995) and the Persian version (0.92) (Mousavi et al., 2006) of the QDS.The Cronbach's alpha of the Tswana version of the WDI was 0.75, which is similar to the coefficient previously reported by Davidson & Keating (2002).
Bland & Altman (1997 and Nunnally (1978) agree that for comparing groups, alpha values of 0.7 to 0.8 are considered satisfactory but that for clinical application much higher values of alpha are needed and a minimum of 0.9 is desirable.Thus in this sense the Tswana versions of the RMDQ, QDS and WDI should all be suitable for group analysis.However the Tswana version of the WDI may not be suitable for the interpretation of individual scores in the clinical setting.The small number of items in the WDI probably contributed to the relatively low alpha value compared to the other two scales.

Test-retest reliability
In this study, 31 patients with low back pain were asked to complete the second questionnaire booklet, containing the four disability questionnaires and the VAS pain scale.The RMDQ, QDS and WDI all showed excellent test-retest reliability with ICC values of 0.93, 0.91 and 0.84.
The ICC reported in the present study for the Tswana version of the RMDQ of 0.93 was higher than those reported in previous studies which used the same 24 hour retest interval, Persian (0.86) (31 subjects reteted) (Mousavi et al., 2006), German (0.82) (20 subjects retested) (Wiesinger et al., 1999).The ICC of the Tswana version of the QDS of 0.91 are in accordance with the ICC reported by the developers of the scale (0.93 for English speaking respondents, 0.88 for French speaking respondents) (Kopec et al., 1995) and the Persian version (0.86) (Mousavi et al., 2006).The ICC for the Tswana version of the WDI was 0.84 and although no studies have reported on the ICC of the WDI it is above the generally considered acceptable value of 0.7 (Fayers & Machin, 2000).These results show the high agreement between all the questionnaires' measurements recorded on two occasions over a 24 hour period.

CONCLUSION
The results suggest that the Tswana versions of the RMDQ, QDS and WDI validated in this study are easy to understand, valid and reliable instruments for the measurement of functional disability caused by LBP in a Tswana speaking population.Therefore these translated instruments may be useful clinical methods for collecting standardised data on activity limitations resulting from LBP in a Tswana speaking population.The Quebec Disability Scale (Kopec et al. 1995) This questionnaire is about the way your back pain is affecting your daily life.People with back problems may find it difficult to perform some of their daily activities.We would like to know if you find it difficult to perform any of the activities listed below, because of your back.For each activity there is a scale of 0 to 5. Please choose one response option for each activity (do not skip any activities) and circle the corresponding number.

Table 2 : Pearson correlation coefficient of the Tswana versions of RMDQ, QDS and WDI with the DRI and VAS (pain)
Today, do you find it difficult to perform the following activities because of your back?
Today, do you find it difficult to perform the following activities because of your back?