In Rajahmundry town in India, 234 community cases of jaundice were interviewed for risk factors of viral hepatitis B and tested for markers of hepatitis A–E. About 41% and 1·7% of them were positive for anti-HBc and anti-HCV respectively. Of 83 cases who were tested within 3 months of onset of jaundice, 5 (6%), 11 (13·3%), 1 (1·2%), 5 (6%) and 16 (19·3%) were found to have acute viral hepatitis A–E, respectively. The aetiology of the remaining 60% (50/83) of cases of jaundice could not be established. Thirty-one percent (26/83) were already positive for anti-HBc before they developed jaundice. History of therapeutic injections before the onset of jaundice was significantly higher in cases of hepatitis B (P = 0·01) or B–D (P = 0·04) than in cases of hepatitis A and E together. Other potential risk factors of hepatitis B transmission were equally prevalent in two groups. Subsequent studies showed that the majority of injections given were unnecessary (74%, 95% CI 66–82%) and were administered by both qualified and unqualified doctors.