Practice Variability in the Inpatient Treatment of Pelvic Inflammatory Disease

Abstract

Objective. To examine variability in the practice of hospitalization and the factors predictive of hospitalization for women with diagnosed Pelvic Inflammatory Disease (PID). Design. Population-based chart review. Setting. Twenty-three hospitals and 40 clinics in King County, Washington. Patients. 2,803 women with a clinical diagnosis of PID who were evaluated at participating hospitals and clinics. Results. Thirteen percent of women diagnosed with PID were hospitalized. Women presenting for care to hospitals were more likely to be referred for inpatient treatment than were those presenting to non-hospital-affiliated clinics (adjusted odds ratio 8.7, 95% confidence interval 5.4 to 13.9). Of clinical features recommended as criteria for hospitalization by the Centers for Disease Control (CDC) that we could evaluate, only patients′ complaints of fever, chills or sweats, indicative of severity of illness, significantly predicted hospitalization. Other factors independently associated with more hospitalizations were lack of birth control use and positive cervical culture for Neisseria gonorrhoea, particularly for women presenting to clinics. Conclusions. The strongest predictor for hospitalization among women diagnosed with PID was site of presentation. Several criteria for hospitalization proposed by the CDC were not predictive.