Background. Hospitalization rates for childhood asthma are three times as high in Boston, Massachusetts, as in Rochester, New York; New Haven, Connecticut, rates are intermediate. We undertook this study to determine how care for children admitted for asthma varies across these communities. Methods. We performed a community-wide retrospective chart review. We reviewed a random sample of all asthma hospitalizations, from 1988 to 1990, of children 2 to 12 years old living in these communities (n = 614). Abstracted data included demographics, illness severity, and treatment before admission. Results. Compared with Rochester children, Boston children were less likely to have received maintenance preventive therapy (inhaled corticosteroids or cromolyn [odds ratio (OR), 0.4 (0.2, 0.9)]), acute "rescue" therapy (oral corticosteroids [OR, 0.2 (0.1, 0.4)]), or inhaled betaagonist therapy [OR, 0.5 (0.3, 1.0)]. A larger proportion of admitted asthmatic patients in Boston (34%) were in the least severely ill group—oxygen saturation 95% or above—compared with patients in Rochester (20%). Conclusions. The quality of ambulatory care, including choice of preventive therapies and thresholds for admission, likely plays a key role in determining community hospitalization rates for chronic conditions such as childhood asthma.