Inter-hospital variation in use of obstetrical blood transfusion: a population-based cohort study

Abstract

Objective: To identify the extent of hospital-to-hospital variation in use of obstetrical blood transfusion.

Design: Population-based cohort study linking provincial perinatal and blood transfusion registries.

Setting: British Columbia, Canada, 2004-2015.

Population: All pregnant women delivering at or beyond 20 weeks' gestation at any British Columbia hospital.

Methods: Mixed-effects regression models were used to estimate hospital-specific transfusion rates after sequentially accounting for (1) the role of random variation, (2) maternal medical and obstetrical characteristics (i.e. patient case mix) and (3) institutional and delivery factors (such as use of instrumental or caesarean delivery).

Main outcome measures: Hospital-specific use of obstetrical red blood cell transfusion.

Results: Among 44 hospitals, crude institutional transfusion rates across the study period ranged from 3.7 to 23.6 per 1000, with an average of 8.3 per 1000. After adjusting for maternal characteristics, institution and delivery risk factors, a nearly three-fold difference in rates between the 10th and 90th percentile remained (5.4-14.5 per 1000). Twelve sites had rates significantly higher or lower than the provincial average. Women residing in remote areas were 2.5-fold (95% CI 1.8-3.5] more likely to receive a blood transfusion than were women residing in metropolitan areas.

Conclusions: Meaningful variation between hospitals in use of blood transfusion during pregnancy was not explained by differences in patient case-mix or institutional factors, suggesting that over- or under-utilisation of this resource may be occurring in obstetrical care.

Tweetable abstract: Use of blood transfusion in pregnant women varied broadly between hospitals in British Columbia, Canada.

Keywords: Blood transfusion; clinical practice variation; hospital comparisons; pregnancy; severe maternal morbidity.