• Michael C Lu, Moshe Fridman, Lisa M Korst, Kimberly D Gregory, Carolina Reyes, Calvin J Hobel, and Gilberto F Chavez.
• Department of Obstetrics and Gynecology, David Geffen School of Medicine, UCLA, Los Angeles, USA. mclu@ucla.edu
• Matern Child Health J. 2005 Sep 1; 9 (3): 297-306.

ObjectiveBecause postpartum hemorrhage may result from factors related to obstetrical practice patterns, we examined the variability of postpartum hemorrhage and related risk factors (obstetrical trauma, chorioamnionitis, and protracted labor) across hospital types and hospitals in California.MethodsLinked birth certificate and hospital discharge data from 507,410 births in California in 1997 were analyzed. Cases were identified using International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) codes. Comparisons were made across hospital types and individual hospitals. Risk adjustments were made using 1) sample restriction to a subset of 324,671 low-risk women, and 2) Bayesian hierarchical logistic regression model to simultaneously quantify the effects of patient-level and hospital-level risk factors.ResultsPostpartum hemorrhage complicated 2.4% of live births. The incidence ranged from 1.6% for corporate hospitals to 4.9% for university hospitals in the full sample, and from 1.4% for corporate hospitals to 3.9% for university hospitals in the low-risk sample. Low-risk women who delivered at government, HMO and university hospitals had two- to threefold increased odds (odds ratios 1.98 to 2.71; 95% confidence sets ranged from 1.52 to 4.62) of having postpartum hemorrhage compared to women who delivered at corporate hospitals, irrespective of patient-level characteristics. They also had significantly higher rates of obstetrical trauma and chorioamnionitis. Greater variations were observed across individual hospitals.ConclusionThe incidence of postpartum hemorrhage and related risk factors varied substantially across hospital types and hospitals in California. Further studies using primary data sources are needed to determine whether these variations are related to the processes of care.

14 keywords...

hide…