Comparative forecasting of adverse drug reaction trends among leprosy patients: Insights from the ARIMA and ETS models in a postpandemic context in a tertiary care center in Sri Lanka

Background Adverse drug reactions (ADRs) in leprosy patients receiving multidrug therapy (MDT) present a persistent challenge, exacerbated by healthcare disruptions during the coronavirus disease 2019 (COVID-19) pandemic. Effective forecasting of ADR trends is essential for resource allocation and healthcare planning. This study analyzes ADR trends at Sri Lanka's Central Leprosy Clinic (CLC) from 2016–2023, uses the ARIMA and ETS models to predict future ADR counts and compares model performance. Methods A retrospective analysis of ADR data was conducted. Trends were assessed pre- and post-COVID-19, and forecasts for 2024–2030 were generated via the ARIMA and ETS models. Model performance was evaluated via the MAE, RMSE, AIC, and BIC. Statistical comparisons between periods were conducted via the Mann-Whitney U test. Results ADR counts increased significantly post-COVID-19, peaking at 46 in 2023. The ETS model outperformed the ARIMA model, with lower MAEs (0.84 vs. 8.00) and RMSEs (1.05 vs. 12.29). ETS forecasts indicated a steady rise in ADRs, reaching 32.33 by 2030, whereas ARIMA predictions showed greater variability. The Mann-Whitney U test confirmed significant differences between the pre- and post-COVID-19 periods (p = 0.03577). Conclusion The ETS model offers superior accuracy and stability for ADR trend forecasting, providing valuable insights for healthcare planning. These findings emphasize the need for robust ADR monitoring systems to address ongoing challenges in leprosy management, particularly in the postpandemic era.