Cross-cultural adaptation and validation of a Bengali Health Assessment Questionnaire for use in rheumatoid arthritis patients

Authors


Abstract

Aim

To translate and adapt the Health Assessment Questionnaire Disability Index (HAQ-DI) into Bengali (B-HAQ) for use in Bangladeshi populations and to test its reliability and validity in patients with rheumatoid arthritis (RA).

Method

The HAQ-DI was translated using rigorous forward-backward protocols and the translated version was subsequently cognitively pretested in a sample of 30 outpatients with RA. The pre-final version of the questionnaire was psychometrically tested for internal consistency and construct validity in a new sample of 100 consecutive RA outpatients.

Results

Ten questions were changed to suit the Bengali culture. Pretests showed that the items included in the B-HAQ were well understood by Bengali patients, while some of the original items were difficult to understand for a majority of patients. The resulting B-HAQ showed good internal consistency and construct validity in the psychometric validation study.

Conclusion

This study suggests that the B-HAQ is a reliable and valid instrument for measuring functional disability in a Bengali-speaking population with RA. Future studies should examine the test–retest reliability and responsiveness of the B-HAQ before it can be confidently recommended as an outcome measure in intervention studies.

Introduction

Physical function is a core outcome domain in clinical studies in rheumatoid arthritis (RA).[1, 2] Since its introduction in the 1980s, the Health Assessment Questionnaire Disability Index (HAQ-DI) has become a standard outcome measure of physical function in clinical studies in RA.[3] Overall, psychometric studies suggest that it has credible construct validity and adequate test–retest reliability and responsiveness to change, although some studies noted a considerable percentage of patients scoring the best possible score.[4, 5]

The HAQ-DI has been translated or culturally adapted for use in over 60 countries.[5] Because HAQ-DI items, like other patient-reported outcome measures of physical function, refer to activities of daily living, modifications of culturally idiosyncratic items have sometimes been necessary, especially when the HAQ-DI was translated to non-Western countries.[6, 7] However, validation studies suggest that the construct validity, internal consistency and test–retest reliability of translated versions is often similar to the original HAQ-DI.[5]

To date, the HAQ-DI has not been translated into Bengali. With nearly 300 million speakers, Bengali is the sixth most popularly spoken language around the world, the second in India and is the national language in Bangladesh. The objectives of this study were to translate and adapt the original HAQ into Bengali (B-HAQ) for use in Bangladeshi populations and to examine its reliability and validity in RA patients.

Materials and Methods

The original HAQ-DI is a self-report questionnaire that contains 20 items referring to basic activities of daily living, grouped into eight categories of functioning. Respondents are asked to rate the degree of difficulty they experience in carrying out each activity on a 4-point rating scale ranging from 0 (without any difficulty) to 3 (unable to do). The highest response in each category is divided by 8, yielding a total disability score of 0–3, where 3 is extreme disability.

Translation and cultural adaptation of the HAQ-DI

The translation of the HAQ-DI into Bengali (B-HAQ) followed the guidelines proposed by Beaton et al.[8] Two independent forward translations of the original US English HAQ-DI were performed (T1 and T2) by native Bengali speakers. One translator was a rheumatologist aware of the concept being examined in the B-HAQ. The other translator did not have a medical background and was naïve to the concept examined. The two translators produced a single synthesis version (T12), which was back-translated into English by two translators with a Masters degree in English literature who were totally blinded to the original version (BT1 and BT2).

To produce a preliminary version of the questionnaire, an expert committee was formed which included health professionals, methodologists and the translators involved. The committee reviewed all translations (T1, T2, T12, BT1, BT2) and the source items, made decisions to achieve equivalence between the English and Bengali versions by reaching consensus on any discrepancies, and developed a preliminary questionnaire for field-testing. The expert committee made critical decisions in four types of equivalence (semantic, idiomatic, experiential and conceptual equivalence) between the source and target versions. For source items that were considered potentially problematic, the committee developed culturally appropriate additional items by discussing the activities more commonly practiced by Bengali people. This resulted in a preliminary Bengali version of the HAQ-DI that included the 20 original items and 23 additional or modified items.

The preliminary version of the questionnaire was field-tested in 30 outpatients with a clinical diagnosis of RA as defined by the American Rheumatism Association (ARA) 1987 revised criteria.[9] After obtaining verbal consent, an interviewer administered the questionnaire and used a series of probes to evaluate whether the patients' understanding of the items matched their intended meaning. Based on the feedback obtained during this phase, the expert committee developed the pre-final version that was used in the psychometric evaluation study. The pre-final version of the B-HAQ, along with the rationale for the modifications to the original items in the HAQ-DI, are presented in Table 1.

Table 1. Modification of Health Assessment Questionnaire (HAQ) items
QuestionModificationReason for modification
  1. N/A, not applicable.

Dress yourself, including tying shoelaces and doing buttons?None requiredN/A
Shampoo your hair?Wash your hair with soap?Shampoo is not commonly used in Bangladesh.
Stand up from a straight chair?Stand up from a jalchoki or mora?A jalchoki or mora are used for sitting in rather than chair.
Get in and out of bed?None requiredN/A
Cut your meat?Cut vegetables or fruits or can you tear meat or sort bones from fish?Although it was translatable, it differs in conceptual equivalence in our culture.
Lift a full cup or glass to your mouth?None requiredN/A
Open a new milk carton?Tear off a new packet of salt or chips?Bangladeshi are used to tearing off new packets of salt and chips.
Walk outdoors on flat ground?None requiredN/A
Climb up five steps?Climb up the ghat of a pond or river?For experiential equivalence the committee added this. A ghat is a flight of steps leading down to the water level.
Wash and dry your body?None requiredN/A
Take a tub bath?None requiredN/A
Get on and off the toilet?None requiredN/A
Reach and get down a 5-pound object (such as a bag of sugar) from just above your head?None requiredN/A
Bend down to pick up clothing from the floor?None requiredN/A
Open car doors?Open the drawerMost Bangladeshi do not have a car and the original item was comprehended correctly by only 26% of patients in the pretest.
Open jars which have been previously opened?None requiredN/A
Turn faucets on and off?Open the cap of a bottle of medicine/syrup/oil?Bangladeshi are more used to opening the cap of a bottles rather than a tap.
Run errands and shop?Walk from house to house or door to door in an out-patient department for treatment?Many women in Bengali culture do not run errands or shop.
Get in and out of a car?Get in and out of a rickshaw/auto rickshaw?Rickshaws are more commonly used than cars.
Do chores such as vacuuming or yard work?Can you sweep and clean your floor/Can you pray in the usual wayThis item was understood properly by only 40% of patients in the pretest.

Psychometric evaluation of the B-HAQ

Patients

A second group of 100 consecutive adult RA patients attending the outpatient clinic of the rheumatology wing of Bangabandhu Sheikh Mujib Medical University (BSMMU) was enrolled to assess the psychometric properties of the pre-final version of the B-HAQ. All patients fulfilled the ARA 1987 revised criteria for the classification of RA.[9] The study was approved by the Ethics Committee of the BSMMU and performed in accordance with the Declaration of Helsinki principles. Informed verbal consent was obtained from all participants before enrolment. As most patients were illiterate, we explained the method verbally to the patients and their families and provided several opportunities to ask questions.

Measures

The interviewer-administered survey contained sociodemographic questions, the 20-item version of the B-HAQ (see Table 1) and 0–10 visual analogue scales (VAS) to measure pain, morning stiffness and patient and physician global assessment of disease activity, all with higher scores indicating worse disease severity. In addition to the self-report data, a 68 swollen joint count (SJC) and tender joint count (TJC) was administered and blood samples were taken to determine the erythrocyte sedimentation rate (ESR).

Statistical analysis

Internal consistency of the B-HAQ was examined by first evaluating the unidimensionality of the B-HAQ using principal components analysis (PCA). Subsequently, Cronbach's alpha coefficient was calculated. An alpha > 0.70 was considered sufficient for group comparisons and a value between 0.90 and 0.95 for individual comparisons. To evaluate construct validity, hypotheses concerning expected Spearman's correlations between the B-HAQ and the other clinical and patient-reported outcome measures were specified, according to previous work on the construct validity of physical function scales in RA populations.[4] Specifically, B-HAQ scores were expected to correlate most strongly with pain and least strongly with ESR. The remaining clinical outcomes were expected to occupy intermediary positions. With respect to the absolute magnitude of correlations, we expected the B-HAQ to correlate strongly with pain (> 0.60) and moderately with the remaining measures (0.30 < < 0.60). All statistical analyses were performed using SPSS for Windows version 12 (SPSS Inc., Chicago, II, USA).

Results

Pretest of the preliminary version

The results of the pretest showed that some of the source items were not understood well by the patients. The item ‘Reach and get down a 5 lb object (e.g., a bag of potatoes) from just above your head’ was comprehended as intended by 66.7% of patients, while the items ‘Open car doors’ and ‘Do chores such as vacuuming, household work or light gardening’ were comprehended correctly by only 26.6% and 40% of patients, respectively. In contrast, all the modified items were comprehend by > 95% of patients. Among the 20 items of the original HAQ-DI, 10 questions required modifications to suit the Bengali culture. Although half of the questions had to be changed, changes were often minor. For eight items, minor changes in wording were required to ensure that objects or activities that items refer to will be more familiar to Bangladeshi people responding to the questions. For instance, most Bangladeshi are unfamiliar with handling car doors and therefore in item 15 of B-HAQ respondents are asked to rate the amount of difficulty they experience in opening drawers. Item 18: ‘Are you able to run errands and shop’ required a bigger change to the content of the item to be made, because in Bangladesh women often do not run errands and shop. Item 7: ‘Are you able to cut your meat’ was changed, although it was properly understood by most respondents, for reasons of cultural equivalence.

Psychometric evaluation of the B-HAQ

Twenty men (20%) and 80 women (80%) were included with a mean age of 40.4 years and a standard deviation (SD) of 13.6 years (range 18–80 years). The duration of their illness varied from 1 year to 30 years (Table 2).

Table 2. Sociodemographic and clinical characteristics of patients
 Mean ± SD or %Range
Age, years40.4 ± 13.618–80
Sex,% female80% 
Duration of morning stiffness, h2.6 ± 0.91–6
Disease duration, years5.7 ± 5.51–30
Number of tender joints47.3 ± 13.05–68
Number of swollen joints16.8 ± 12.34–47
Pain4.4 ± 1.72–10
Patient global assessment4.5 ± 1.72–9
Physician global assessment4.4 ± 1.61
Erythrocyte sedimentation rate69.3 ± 23.510–140
Rheumatoid factor,% positive60% 

Principal components analysis indicated a strong first component that explained 57.2% of the observed variance. Factor loadings ranged from 0.495 to 0.925, indicating a high degree of homogeneity of the B-HAQ. With a value of 0.96, Cronbach's alpha was satisfactory for individual-level analyses.

Table 3 summarizes correlations between the B-HAQ and different disease severity parameters. In contrast to what was expected, the correlation of the B-HAQ with pain was only moderate. However, correlations with TJC, SJC, morning stiffness and ESR were of the expected magnitude. Moreover, although the B-HAQ did not correlate most strongly with pain, as initially anticipated, its correlation with ESR was weakest, while the correlations with the TJC and morning stiffness occupied intermediary positions relative to pain and ESR. Overall these results were well in accordance with the pre-specified hypotheses, indicating adequate construct validity for the B-HAQ.

Table 3. Spearman's correlations between the Health Assessment Questionnaire into Bengali (B-HAQ) and other measures of disease severity
 r
Pain0.451
Morning stiffness0.437
Tender joint count0.429
Swollen joint count0.515
Erythrocyte sedimentation rate0.258

Discussion

This study describes cross-cultural translation and evaluation of the HAQ-DI into Bengali, according to rigorous methodological standards. Psychometric analyses of the translated version demonstrated adequate construct validity and internal consistency of the B-HAQ in RA patients in Bangladesh.

Ten items of the original HAQ-DI were modified for use in Bengali populations. These modifications were made while trying to maintain optimal semantic, idiomatic, experiential and conceptual equivalence between the source and target versions. We believe that the modifications made are important to ensure that the questions will be relevant and understandable to all Bengali patients. However, considering the amount of items that were modified, future research could be directed at evaluating the cross-cultural measurement equivalence of the original and Bengali versions of the HAQ to ensure that research results can be compared or pooled across versions.

The pattern of correlations of the B-HAQ with other disease activity measures largely corresponded to our pre-specified hypotheses. Considering that the hypotheses were formulated in accordance with previous international work on physical function scales in RA,[4] these results provide convincing evidence for the construct validity of the B-HAQ.

A limitation of the current study is that its cross-sectional design did not allow us to directly analyze the responsiveness and test–retest reliability of the B-HAQ. Although previous research has shown that the original and a variety of translated versions of the HAQ-DI show excellent longitudinal measurement properties,[4, 5, 10] study of the longitudinal performance of the B-HAQ is an important area of future research.

In summary, the current study describes the process of cross-culturally translating the HAQ-DI to Bengali using established guidelines for cross-cultural translations. Results of the psychometric analysis provide preliminary support for the measurement properties of the B-HAQ. Future research should be directed at studying the longitudinal performance of the B-HAQ.

Conflict of Interest

All the authors responsible for this article have no conflicts of interest of any kind.

Ancillary