Using DHIS2 routine data for health system preparedness in resource-limited settings: A bayesian predictive approach in Bangladesh

Health systems in low- and middle-income countries (LMICs) like Bangladesh face persistent challenges in delivering timely and equitable care, often exacerbated by poor planning and inefficient resource allocation. Forecasting service utilisation using routine health data can support more responsive and data-driven health system planning, yet such approaches remain underutilised in Bangladesh. By analysing service utilization trends and projecting future service volume at national and regional level, we aim to improve region-specific health planning and promote more efficient and equitable health service provision. We analyzed monthly routine health service data reported into the District Health Information Software 2 (DHIS2) platform between January 2021 and March 2025 in Bangladesh. We examined key indicators across maternal, newborn, child, and hospital-based services. Bayesian log-linear Poisson regression models, adjusted for seasonality and autocorrelation, were applied to forecast service utilisation for the final nine months of 2025 and all of 2026. Relative changes in 2025 and 2026 were calculated using 2024 as the reference year.The analysis revealed rising trends across most service areas relative to 2024 levels. Compared to levels in 2024, normal deliveries are projected to increase by 6% in 2025 and 10% in 2026, while caesarean sections are expected to rise by 4% and 12%, respectively. Low birth weight (LBW) deliveries are forecasted to increase by 10% in 2025 and 25% in 2026, with the largest relative growth in Rajshahi and Dhaka. Coverage of Kangaroo Mother Care (KMC) is projected to rise substantially—by 20% in 2025 and 74% in 2026. Pneumonia treatment is expected to increase by 31% by 2026, particularly in Dhaka. Outpatient visits are forecasted to grow by 32% by 2026. Notable regional disparities persist, with Dhaka and Chattogram showing the highest service utilisation, while Barisal and Sylhet consistently report the lowest levels. Bangladesh’s health system must prepare for increasing service utilisation across all service categories. Forecasting using DHIS2 data supports for proactive planning and equitable resource allocation. Strategic investments in infrastructure, workforce, and data-driven planning are essential for building a resilient health system.