Submission ID 130792
| Issue/Objective | Globally, Postpartum hemorrhage (PPH) is a leading cause of maternal mortality, and delays in receiving blood transfusions can cause preventable deaths. Over the past three decades, Rwanda has achieved notable success in reducing maternal mortality, with the maternal mortality ratio declining from 1,071 to 203 per 100,000 live births. Despite this progress, maternal mortality remains high, and reductions have slowed between 2015 and 2020, suggesting that further gains may require new and innovative approaches. In October 2016, Rwanda launched a drone-based delivery system for blood products to the first health facilities. This initiative, implemented in partnership with Zipline, has since expanded to serve over 450 health facilities across the country, excluding Kigali. Operated from two strategically located distribution hubs, these drones fulfill both routine resupply requests and urgent, life-saving deliveries based on real-time facility orders. At present, drones account for approximately 75% of blood deliveries to facilities outside of Kigali. Despite this substantial scale, the extent to which drone delivery has translated into improved health outcomes, particularly reductions in postpartum hemorrhage (PPH) mortality, remains unclear. |
|---|---|
| Methodology/Approach | Using data from the Rwanda Health Management Information System from 2014 to 2024, we assessed the impact of drone blood delivery on the maternal mortality ratio due to postpartum hemorrhage using controlled Interrupted time series analysis. We calculated the Maternal Mortality Ratio (MMR) by dividing the total number of PPH-related maternal deaths per 100,000 live births. For each intervention facility, we defined time = 0 as the quarter in which drone-delivered blood products began. We then constructed a balanced panel by including data from 12 quarters before and 20 quarters after the intervention start date. For control facilities, a placebo intervention date was determined. We applied mixed-effects negative binomial regression to account for clustering and overdispersion in the count data. To determine whether the intervention's impact varied across different types of facilities, we performed subgroup analyses based on four key characteristics: facility level (primary, secondary, or tertiary), health facility size (small, medium, or large), and distance to drone distribution center ( <96.6 km or >96.6km). This distance threshold was derived from the drone's cruise speed ( 96.6 km/h) under the assumption that a massive transfusion must be initiated within an hour to be effective. To evaluate the robustness of our findings, we restricted the study cohort exclusively to hospitals (n=30 intervention; n=7 control). |
| Results | We found that the intervention was associated with a large immediate reduction in postpartum hemorrhage mortality in intervention health facilities relative to control health facilities (RR: 0.34; 95% CI: 0.15 to 0.76; p=0.008). This corresponds to a 66% reduction in the risk of maternal mortality in the intervention group relative to the control group. There was no evidence of a differential change in post-intervention trends between the two groups (RR: 0.98, 95% CI: 0.89 to 1.07). In subgroup analyses by health facility level, the use of drones to supply blood products was associated with differential immediate effects, while no differences were observed in post-intervention trends. Tertiary facilities in the intervention group had a 90% reduction in PPH-related mortality relative to the control group (RR: 0.10; 95% CI: 0.02 to 0.68; p=0.019), while secondary facilities experienced an 87% reduction (RR: 0.13; 95% CI: 0.02 to 0.96; p=0.046). Beyond the facility level, proximity to the drone distribution center also played a key role in mortality outcomes. Facilities located within a 96.6km radius experienced an 81% immediate reduction in PPH-related mortality relative to the control group (RR: 0.19; 95% CI: 0.07 to 0.49; p<0.001). In contrast, we observed no statistically significant immediate change among facilities located more than 96.6km. When examining results by facility size, the intervention appeared most effective in large institutions, with a differential change. Specifically, large facilities experienced a 93% immediate reduction in PPH mortality relative to the control group (RR: 0.07; 95% CI: 0.02 to 0.29; p < 0.001). The sensitivity analyses found similar results. First, by excluding health centers, the cohort was restricted to 30 intervention and 7 control facilities. In this subset, the immediate impact of drone implementation was more pronounced, with a 78.9% immediate reduction in the maternal mortality ratio (RR: 0.21; 95% CI: 0.08 to 0.54) in the intervention group relative to the control group. The post-intervention trend remained stable (RR: 0.95; 95% CI: 0.85 to 1.06). |
| Discussion/Conclusion | This study provides empirical evidence that the introduction of drone-based blood delivery in Rwanda was associated with a clinically important immediate reduction in maternal mortality due to PPH, particularly in larger health facilities that were within one hour of drone centers. These findings highlight the potential of drone technology to address critical delays in emergency obstetric care. As Rwanda and other low- and middle-income countries continue to invest in digital and logistical health innovations, integrating drone delivery into broader maternal health strategies could play a role in reducing preventable maternal deaths. |
| Presenters and Affiliations | Corneille Killy Ntihabose University of British Columbia Nisingizwe Marie Paul Ministry of Health, British Columbia Sriram Veena University of British Columbia Stucchi Andrea University of British Columbia Cheng Lucy University of British Columbia Remera Eric Rwanda Biomedical Centre hutcheon Jennifer Anne University of British Columbia Law Michael Robert University of Calgary |
