Population-based cohort study of hospital delivery volume, geographic accessibility, and obstetric outcomes

Objective To determine associations between geographic accessibility, delivery volume, and obstetric outcomes. Methods Population-based cohort study of linked hospital administrative, census, and geospatial data (2006-2009) from all Canadian jurisdictions except Quebec. Perinatal mortality and major maternal morbidity/mortality were compared across categories of road distance and hospital delivery volume. Results Among 820 761 mothers delivering 827 504 neonates, travel distance had minimal effect on perinatal mortality. Compared with mothers travelling 0-9 km, the odds of adverse maternal outcomes was decreased for women travelling modest distances (20-49 km, odds ratio, 0.80 [95% confidence interval, 0.75-0.86]), and increased thereafter (50-99 km, 0.99 [0.89-1.10]; 200-299 km, 1.44 [1.10-1.87]; >400 km, 2.22 [1.06-4.63]). Relative to high-volume hospitals (>2500 deliveries/year), adverse maternal outcomes were less likely for hospitals with 1000-2499 (0.90 [0.86-0.95]), and roughly equivalent for hospitals with 200-499 (1.34 [1.22-1.48]) and 500-999 (1.27 [1.17-1.39]) deliveries/year. Odds of perinatal mortality ranged from 1.04 (0.73-1.49; 100-199 deliveries/year) to 1.50 (1.04-2.16; 50-99 deliveries/year); the pattern did not suggest causality. Conclusion Maternal outcomes worsen when travel distance is greater than 200 km, and improve when delivery volume exceeds 1000 deliveries per year.