Context. Patients and the public could benefit from identification of factors that prevent drug users'heavy reliance on inpatient care; however, optimal health care delivery models for illicit drug users remain ill-defined. Objective. To evaluate associations of outpatient medical and drug abuse care with drug users' subsequent hospitalization rates. Design mull Setting. Retrospective cohort study of data from longitudinally linked claims for all ambulatory physician/clinic services and drug abuse services covered by the New York State Medicaid program. Subjects. A total of 11 556 human immunodeficiency virus(HIV)-positive and 46 687 HIV-negative drug users. Main Outcome Measures. Hospitalization in federal fiscal year (FFY) 1997 compared by 4 patterns of care in FFY 1996: regular drug abuse care (>6 months in 1 program), regular medical care (>35% of care from 1 clinic, group practice, or individual physician), both, or neither. Results. Hospitalization occurred in 55.6% of HIV-positive and 37.5% of HIV-negative drug users, with a mean of 27.5 and 24.5 inpatient days, respectively. In HIV-positive drug users, the adjusted odds ratio (AOR) for hospitalization was lowest among those with both regular medical and drug abuse care (AOR, 0.76; 95% confidence interval [CI] 0.67-0.85) followed by those with regular medical care alone (AOR, 0.82; 95% CI, 0.74-0.91) and regular drug abuse care alone (AOR, 0,85; 95% CI, 0.76-0.96) vs those with neither. In HIV-negative drug users, the AOR of hospitalization was lower for those with regular medical and drug abuse care (AOR, 0.73; 95% CI, 0.68-0.79), regular drug abuse care alone (AOR, 0.71; 95% Cl, 0.66-0.76), and regular medical care (AOR, 0.91; 95% Cl, 0.86-0.95) vs those with neither. Both types of care showed favorable effects for all but drug abuse-related hospitalizations. Conclusion. Our data indicate that regular drug abuse care with regular medical care for drug users is associated with less subsequent hospitalization. (Author' s abstract)