We performed a population-based case-control study of risk factors for primary invasive Haemophilus influenzae type b (Hib) disease in metropolitan Atlanta from July 1, 1983, through June 30, 1984. Active surveillance identified 102 cases in children less than 5 years of age, 89 of whom lived in households with telephones. We used random digit dialing to select 530 controls who were postmatched to cases by age. Multivariate analysis showed a significant association between Hib disease and two independent exposure factors, household crowding (odds ratio (OR) 2.7, 95% confidence limits (CL) 1.3 to 5.6) and day-care attendance. For day-care attendance, relative risk was highest in 2- to 5-month-old infants (OR 17.7, 95% CL 5.8 to 54.4) and declined to 9.4 (4.3 to 20.9) at ages 6 to 11 months, 5.0 (2.7 to 9.3) at 12 to 23 months, 2.7 (1.3 to 5.5) at 24 to 35 months, and 1.4 (0.5 to 4.0) in 36- to 59-month-old children. Fifty percent of all invasive Hib disease that occurred during the study period was attributable to exposure to day-care; the attributable risk for household crowding was 18%. Dose-response effects were observed for hours per week of day-care attendance and extent of household crowding. Breast-feeding was protective for infants less than 6 months of age (OR 0.08, 95% CL 0.01 to 0.59). After controlling for socioeconomic and other confounding factors, we could demonstrate no effect of black race on cumulative risk of invasive Hib disease. Our study defines high-risk groups and provides a population-based model of the interrelationship between risk factors associated with invasive Hib disease.