Effect of Hospital Delivery Volume and Travel Time on Perinatal Mortality and Transport Delivery in Norway

Background: Access to advanced obstetric and perinatal care is consequential to maternal and neonatal health and survival. In many health care systems, regionalisation of delivery services must be balanced against distance to hospital. Methods: This is a cohort study using nationwide patient level data from the Medical Birth Registry of Norway. We compared outcomes within women moving between hospital catchment areas, using the woman as her own control. In an alternative analysis, we compared women living in neighbouring municipalities, while in different hospital catchment areas. We estimated effects of expected hospital volume adjusted for expected travel time, and expected travel time adjusted for expected hospital volume. The main outcomes were perinatal mortality and transport delivery. Non-linearity was analysed using restricted cubic spline regression. Findings: Among all births in Norway from 1999 to 2016 we included 203,464 births by women moving between hospital catchment areas and 460,776 births by women residing in neighbouring municipalities. Perinatal mortality gradually declined with increasing catchment area hospital volume, levelling off at around 2,000 births/year. Among women moving between catchment areas, perinatal mortality was 55% (95% confidence interval (CI) 3 to 134) higher when residing in a catchment area of a hospital with 500 births/year compared to 2,000 births/year. The estimate for 500 vs. 1,000 births/year was 29% (95% CI 3 to 62). There was a near linear increase in the risk of transport delivery with travel time, but little evidence of any association with perinatal mortality. Results from the alternative analyses were largely similar. Interpretation: A hospital delivery volume of 2,000 births/year may represent a threshold where smaller hospitals have gradual increased perinatal mortality with lower numbers of births. While there was little indication that longer travel time influenced mortality, there was an increase in the risk of transport delivery.