Objective: To examine relations between characteristics of a child's usual source of primary care and involvement of that source before and during hospitalization.
Design: Medical record review of pediatric hospitalizations.
Setting: All hospitals in Boston, Mass; New Haven, Conn; and Rochester, NY admitting children during the calendar years 1988 through 1990.
Patients: The study included 1875 randomly selected pediatric hospitalizations for five diagnostic groups (i.e., asthma and other lower respiratory tract disease, abdominal pain [including appendicitis], meningitis [bacterial and viral], toxic ingestions, and head injury). Hospital records selected were limited to children aged between 1 month and 12 years and residing in the three study communities.
Outcome measures: Whether the primary care source examined the child before admission to the hospital, referred the child to the emergency department, or served as the in-hospital attending physician.
Results: Of the medical charts reviewed, 85.7% identified primary care sources. Children in Rochester had higher rates of medical visits before admission (P < .04), referrals (P < .001), and in-hospital care provided by the primary care physician (P < .001, chi 2) than children in Boston and New Haven. Patterns of primary care involvement also varied by source of care within cities, after controlling for income and severity of illness. Compared with children from Rochester community-based private practices, children in Boston receiving care from health centers, hospitals, or community-based private practices generally had 25% to 50% lower likelihood of positive findings on all primary care involvement measures. Children in New Haven receiving care from community-based private or hospital-based practices also had lower rates, but involvement rates were not higher when they received care from health centers. Other children in Rochester and children receiving care from health maintenance organizations in all cities demonstrated almost no significant differences compared with data from Rochester community practices.
Conclusion: The source of primary care is associated with patterns of prehospital and hospital care among hospitalized children, although specific associations vary by city.