A spreadsheet simulation model of hepatitis A disease was developed to evaluate the cost effectiveness of an inactivated [corrected] hepatitis A vaccine ('Havrix', SmithKline Beecham) in high risk groups in France. Gammaglobulin prophylaxis, systematic vaccination without screening and vaccination of nonimmune persons after systematic screening were compared with the reference situation of no prevention over a 10-year period. It was found that both vaccination strategies would prevent 98% of new cases of hepatitis A, and would generate savings of FF4.2 to FF4.7 million ($US1 = FF5, 1995) in alternative service volunteers [initial seroprevalence (IS) 26%] stationed in countries with high hepatitis A endemicity. The cost per symptomatic case avoided [i.e. the cost-effectiveness ratio (CER)] was found to vary from FF177,612 with screening to FF281,463 without screening in adult tourists (IS 77%). In hospital workers, screening before vaccination (CER = FF65,108) would be about half as costly as systematic vaccination (IS 55 to 79%). Recommendations for vaccination should take into account the specific collective or individual risk, age, seroprevalence and probability of compliance with the prevention protocol.