Cross-cultural translation, adaptation and validation of a Japanese version of the functional index for hand osteoarthritis (J-FIHOA)

Abstract

Purpose: Hand osteoarthritis (hand OA) is a highly prevalent and heterogeneous disorder. Although well-designed clinical trials and guidelines have been published, there are comparatively few hand OA studies concerning Asian countries. To both pursue appropriate clinical courses for patients and engage in future research, physicians need to discriminate between symptomatic and non-symptomatic hand OA. The Functional Index for Hand OsteoArthritis (FIHOA) has been well-studied and is accepted as the gold standard instrument with excellent reliability and responsiveness. Approximately twenty linguistic versions are currently available. Our objective was to make a Japanese version of the FIHOA (J-FIHOA) and to validate it among Japanese hand OA patients. Methods: Translation and cultural adaptation process: The FIHOA was translated following a process distilled from the established guidelines. An expert committee consisting of 6 members (two language professionals, three health professionals and an American native English speaker) was convened. Forward translation was performed independently by one professional translator who had no prior knowledge of this study and two language professionals serving on the expert committee. The committee assessed all results and synthesized them into a preliminary version of the J-FIHOA. It was translated back into English independently by two English native professional translators, one with a medical background and one without. Most of the FIHOA items were deemed equivalent, but discrepancies between the original and retranslation revealed several problems. Additional forward/backward translations were undertaken for specific linguistic points in need of revision. To resolve a few major issues, the committee consulted with the FIHOA developers. After all this process the J-FIHOA was finalized. Validation process: Our university and 16 other hospitals recruited hand OA patients at the outpatient departments from September 2017 to September 2018. New or follow-up patients who met the ACR classification criteria were eligible for inclusion. Patients with other rheumatic diseases or post-traumatic OA were excluded. A medical record review, X-rays of both hands and J-FIHOA in addition to the following patient-rated questionnaires were collected at an enrollment visit: the Japanese version of the Stanford Health Assessment Questionnaire (J-HAQ), numerical rating scale for pain (NRS pain) and the Japanese version of the Short Form 36 Health Survey (J-SF-36). The J-FIHOA scores at the enrollment visit were used to assess Cronbach’s alpha coefficient/item-total correlation and compared with the J-HAQ, NRS pain and J-SF-36 yielding respective Spearman’s rank correlation coefficients. Correlation was categorized as weak (r = 0.30-0.49), moderate (r = 0.50-0.69) or strong (r = 0.70-1.00). As item 7 is a gender-role question, an item total correlation needed to be calculated independently for women and men. We used test-retest methods for the patients whose treatment and symptoms were unchanged within the previous 3 months. The interval was 1-2 weeks. The reliability of the J-FIHOA was assessed with the intraclass correlation coefficient (ICC). Those who started new pharmacological treatments (oral acetaminophen, NSAIDs and/or tramadol) required answering the patient-rated questionnaires and a 1-month follow-up visit. Differences between pre- and post- pharmacological treatments were assessed to evaluate effect size. Results: Twenty-nine male and 145 female hand OA patients participated. Mean age at the enrollment visit was 65 years (42 to 88). One hundred fifty-seven longitudinal data sets were pooled and used to evaluate reliability (137 for test-retest) and responsiveness (20 for effect size). Cronbach’s alpha was 0.91, which showed high internal consistency. Item-total correlation ranged from 0.46 (item 5) to 0.89 (item 7) among men, and from 0.57 (item 10) to 0.79 (item 6) among women. Construct validity between the J-FIHOA and other scales was: 0.73 (J-HAQ), 0.56 (NRS pain), -0.31 (J-SF-36 physical component) and -0.21 (J-SF-36 mental component). There were stronger correlations between the J-FIHOA and J-HAQ than we expected. Although the FIHOA was designed as an instrument for hand functional impairment and has no pain-related questions, previous reports have shown moderate correlations between the FIHOA and pain. Our result was concordant with these reports. It was reasonable that the J-SF-36 was moderately correlated on the physical but not the mental component. Test-retest data showed that the intraclass correlation coefficient was 0.83 (95% confidence interval, 0.77 to 0.88) and in line with other linguistic versions. The effect size of the J-FIHOA was 0.56 while the J-HAQ was 0.44. We assumed that the higher effect size was because the J-FIHOA was a more disease-specific scale for hand OA. Conclusions: The J-FIHOA has good reliability and responsiveness. Current limitations are a small number of male participants (n=29) and the data sets for responsiveness (n=20). Patient recruitment and data collection will continue through December 2018.