Vaccines are traditionally tested under the optimal conditions of clinical trials (efficacy). However, their public health impact is better assessed under the real conditions of a clinical practice (effectiveness). The authors aimed to estimate the effectiveness of a rotavirus vaccine (rhesus rotavirus vaccine-tetravalent (RRV-TV)) to prevent rotavirus-related hospitalization among children <or=3 years of age. They reviewed computer records from an urban pediatric practice in New Orleans, Louisiana, comprising 1,413 children born between April 1, 1998, and June 1, 1999, and who would have been eligible to receive RRV-TV. They also reviewed hospital records to determine rotavirus hospitalizations for gastroenteritis during October 1998-June 2001. A total of 1,099 children were enrolled, 513 unvaccinated and 586 vaccinated. The attack rate of rotavirus hospitalization was 0.34 per 100 child-years-0.52 for unvaccinated (no doses) children, 0.20 for partially vaccinated (one or two doses) children, and 0 for fully vaccinated (three doses) children. Protective vaccine effectiveness was 70% (95% confidence interval (CI): -43, 94) among vaccinated children, 61% (95% CI: -86, 92) among partially vaccinated children, and 100% (95% CI: -120, 100) among fully vaccinated children. One episode of rotavirus-associated hospitalization was prevented per 104 infants partially vaccinated and per 64 infants fully vaccinated.