Bypassing nearby childbirth care facilities: Geospatial modelling of travel to nearest, utilised, and referral facilities in rural Uganda using context-specific travel speeds

Abstract Background Previous Ugandan studies estimated travel time to childbirth services using generic speeds, incomplete facility databases, and nearest-facility assumptions, without accounting for referrals. We estimated travel times to facilities women actually used, incorporating referral pathways and context-specific speeds, and assessed urban-wealth inequities in access and bypassing in rural midwestern Uganda. Methods We assembled spatial data on facilities, socioeconomic status, urbanisation, roads, land cover, and geotraced speeds from car journeys in Kasese and Bundibugyo. From records, we extracted residential addresses and referral pathways for 357 women delivering in all 42 public childbirth facilities during November–December 2024. Using a least-cost path algorithm, we estimated travel times to nearest and utilised facilities under slowest, average, and fastest scenarios; incorporated referrals; examined wealth and urban–rural inequalities; and estimated proportions of women of childbearing age (WoCBA) living within 15, 30, 60, and 120 minutes of the nearest childbirth facility. Results Travel speeds ranged from 8.1 to 49.4 km/h. Mean travel time to the nearest facility was 24 minutes, rising to 56 minutes under the slowest scenario. Under the slowest scenario, 99.3% of WoCBA lived within 2 hours of the nearest facility, but only 52.1% within 30 minutes. Travel times were longer for rural and poorer women. Overall, 65.1% used their nearest facility, and referrals added a mean of 21 minutes. Conclusions Travel times were longest for poorer rural women, with bypassing and referrals increasing journey time. Investigating bypassing and reducing unnecessary referrals is needed. Utilised-facility travel with referrals better reflects access than nearest-facility models.