Abstract
BACKGROUND: Adherence to long-term therapy for diabetes remains low. Accurately measuring adherence is the primary step in improving adherence. We translated and validated the Sinhalese version of the Brief Medication Questionnaire (BMQ) in patients with diabetes. METHODS: The study was conducted at the National Hospital of Sri Lanka between April and December 2017, including 165 patients with diabetes. BMQ was translated into Sinhalese using the translation-back translation method. The translated questionnaire validation included evaluation of internal consistency, temporal stability, and performance in regard to a gold standard (HbA1c). RESULTS: Mean age (±SD) was 60.6 ± 11.1 years, and 46.1% were males. Mean duration of diabetes in the participants was 13.4 ± 7.8 years. Mean HbA1c was 8.3 ± 1.7%, with poor glycaemic control (HbA1c ≥ 8.5%) identified in 41.8%. Medication adherence measured by the BMQ regimen, belief, and recall screens were 39.4%, 75.8%, and 18.8%, respectively. In the analysis of temporal stability, the overall BMQ and the regimen, belief, and recall screens demonstrated good concordance between test and retest with significant gamma correlation coefficients of r = 0.85 (p < 0.001), r = 0.81 (p < 0.001), r = 0.84 (p < 0.001), and r = 0.91 (p < 0.001), respectively. The overall BMQ had a Cronbach α coefficient of 0.65 (95% CI: 0.61-0.70). The questionnaire performance with regards to the gold standards for the overall BMQ AUC was 0.73 (95% CI 0.65-0.80), while the BMQ regimen screen AUC was 0.61 (95% CI 0.53-0.70). The overall BMQ score with a cutoff value of 2 presented better equilibrium between sensitivity and specificity for the gold standard. Those with low adherence had a significantly higher percentage of poor glycaemic control (HbA1c ≥ 8.5%). CONCLUSION: The translated questionnaire demonstrated good reliability (internal consistency), temporal stability (test-retest reliability), and validity when assessed using a gold standard for disease control. Using culturally validated tools to evaluate adherence may help clinicians to identify low adherence and institute corrective measures.