Validation of the IDF-DAR risk tool for fasting in Ramadan for adults with diabetes mellitus in primary care: A nationwide multicentre study in Malaysia.

Introduction: Fasting during Ramadan is a religious obligation for Muslims but poses health risks for individuals with diabetes mellitus. The International Diabetes Federation-Diabetes and Ramadan Alliance (IDF-DAR) introduced a risk stratification tool in 2021 to guide clinicians, though its utility in primary care settings remains limited.

Methods: A retrospective observational study was conducted on adults with diabetes attending government health clinics in Malaysia from April 15 to June 15, 2024. Medical records of those who attempted fasting during Ramadan were reviewed. The primary outcome was a composite of hypoglycaemia, hyperglycaemia, diabetes-related hospitalization, or dehydration leading to breaking the fast. Discriminative performance of the IDF-DAR tool was evaluated using area under the receiver operating characteristic curve (AUC). Calibration was assessed via the Hosmer-Lemeshow test.

Results: A total of 310 patients were included (99 % with type 2 diabetes). The mean age was 61 years and the median diabetes duration was 7 years. Adverse fasting outcomes were observed in 18.4 % of patients, with hypoglycaemia being the most common (13.5 %). The IDF-DAR risk stratification tool demonstrated good discriminative ability, achieving an area under the ROC curve (AUC) of 0.78 (95 % CI: 0.72-0.84). At the recommended cut-off for distinguishing low-moderate from high-risk categories, the tool achieved a sensitivity of 92.9 % and a specificity of 40.9 %. The Hosmer-Lemeshow goodness-of-fit test indicated poor agreement between observed and predicted adverse outcomes, with a statistically significant result (P < 0.05).

Conclusion: The IDF-DAR risk stratification tool identifies high-risk patients fasting during Ramadan in primary care. However, its poor calibration and specificity as highlights the need for refinement of the model to improve its predictive accuracy. Enhancing the tool's calibration could allow for better individual risk estimation and more precise clinical decision-making in diverse primary care settings.