Abstract
Abstract
Background
Strengthening health research and information systems in low- and middle-income countries requires evidence on how data quality failures in routine health systems translate into poor health outcomes. In Malawi, stock-outs of essential medicines remain persistent despite the national rollout of the Open Logistics Management Information System (Open-LMIS). However, the statistical relationship between LMIS-derived performance indicators and stock-out burden has not previously been quantified at district level using real national LMIS data. This study examined the association between LMIS availability metrics, order fill rates, and essential medicine stock-out burden across government-managed health facilities at Blantyre District Health Office (DHO).
Methods
A cross-sectional study was conducted at six government-managed health facilities under Blantyre DHO. Nine essential medicines were extracted from the national Open-LMIS dataset for five months (January to June 2023; April was absent from the national LMIS data pull), generating 196 facility-product-month observation units. Spearman rank correlation analysis tested associations between mean facility-level availability percentage, order fill rate, accumulated stock-out days, and stock-out frequency. Bonferroni correction was applied for four simultaneous comparisons (adjusted α = 0.0125). Descriptive Key Performance Indicator (KPI) analysis was structured around the Supply Chain Operations Reference (SCOR) model domains. The study is reported in accordance with the STROBE statement for cross-sectional studies.
Results
Madziabango Health Centre had the highest accumulated stock-out burden (269 days; 10 events; mean availability 69.1%), while Chilomoni Health Centre had zero stock-out days (mean availability 100.0%). Phenobarbitone 30mg had the highest accumulated stock-out days (251 days; 37.0% observation-level stock-out rate). Morphine sulphate had the highest observation-level stock-out rate (60.0%). Mean order fill rate across all facilities was 28.24%. Spearman rank correlation demonstrated strong, statistically significant negative associations between mean availability and both accumulated stock-out days (rₛ = −0.829, p = 0.042) and stock-out frequency (rₛ = −0.829, p = 0.042). Order fill rate was not significantly correlated with either stock-out outcome at facility level (rₛ = −0.257 to − 0.600, p = 0.208–0.623).
Conclusions
Mean medicine availability recorded in Open-LMIS is a strong, statistically significant predictor of stock-out burden across government health facilities in Blantyre DHO. These findings validate LMIS data quality metrics as proxy indicators suitable for health system monitoring, early-warning system design, and policy prioritisation in LMIC public health supply chains. Investment in LMIS data quality and adherence is a high-priority, low-cost health system strengthening intervention with direct implications for medicine access and patient safety.