Hospital-acquired infections (HAI) are a major and largely preventable cause of morbidity and morbidity worldwide. Very few reports on the prevalence of HAI in sub-Saharan Africa have been published and most of those that have appeared in the press have focused on surgical-wound infection. In the present, questionnaire-based, point-prevalence study, in which the doctor on the ward round was used as the primary informant, the prevalences of all HAI among all the inpatients at a tertiary referral hospital in northern Tanzania were estimated. On the day of the study, there were 412 inpatients (in 15 ward areas) and 61 cases of HAI were identified, giving an overall HAI prevalence of 14.8%. The prevalences of HAI were particularly high in the medical intensive-care unit (40%), the surgical (orthopaedic and general surgery) wards (36.7%), and one of the general medical wards (22.2%). Factors associated with a patient having a HAI were hospitalization for >30 days [odds ratio (OR) = 4.07; 95% confidence interval (CI) = 2.07-7.99]; being a patient on the orthopaedic and general surgical ward known as 'Surgical 2' (OR = 2.14; CI = 1.02-4.46); and being referred from another health facility (OR = 1.90; CI = 1.02-3.42). The most commonly identified HAI in the hospital were urinary-tract infections (14 cases), followed by surgical-wound infections (10 cases) and then lower respiratory-tract infections (six cases). Twenty HAI were 'unspecified'. The study was rapid and cheap to carry out. The results not only gave a baseline estimate of HAI in the study setting but also identified key areas for interventions to reduce HAI.