Health services utilization after induced abortions in Ontario: A comparison between community clinics and hospitals

The purpose of this study was to compare postabortion health services utili zation of hospital abortion patients with community clinic abortion patient s using administrative databases. The study was a retrospective cohort stud y. The study group consisted of patients with induced abortions (n = 41,039 ) performed in hospitals or community clinics recorded in the 1995 Ontario Health Insurance Plan claims (OHIP) database. An age matched cohort of 39,2 20 women who did not undergo induced abortions was selected from the same d ata source to serve as controls. The main outcome measures were health serv ices utilization indicators constructed from OHIP data within 3 months post abortion from office consultations, emergency room consultations, and hospi tal admissions. Hospitalization indicators were constructed from Canadian I nstitute for Health Information hospital discharge data with in 3 months po stabortion and included data on hospitalizations for infection, certain sur gical events, or psychiatric problems. Postabortion health services utiliza tion and hospitalization were higher in the patient population, regardless of service location, than in the age-matched co hort. Within the abortion p atient population, hospital day surgery patients had higher rates of postab ortion utilization and hospitalization than did community clinic patients. Multivariate analysis revealed that hospital day surgery patients had a hig her risk of subsequent post abortion hospitalizations for infections (odds ratio [OR] 1.67, 95% confidence interval [CI] 1.23-2.28), surgical events ( OR 1.70, 95% CI 1.30-3.24) and psychiatric problems (OR 2.65, 95% CI 1.77-3 .98) than community clinic patients. The rates of postabortion health servi ces utilization and risk of hospitalization were lower in community clinic abortion patients than in hospital day-surgery patients. However, it is not possible to fully control for important confounding variables when using t hese administrative data.