Abstract In a national prevalence survey setting, we studied whether the day of week selected for data collection, and the number of days needed to complete the survey, were associated with the prevalence of hospital-acquired infection (HAI). The EPINE (Estudio de Prevalencia de las Infecciones Nosocomiales en España) database (1990–2002) was analysed for the purposes of the study. Adjusting for the admission day in the week, the number of intrinsic risk factors, the number of extrinsic risk factors and the prevalence length of stay, a ‘weekend effect’ was confirmed in this study. The day of the week selected for data collection was related to the presence of infection in the surveyed patients, showing for the period of Saturday–Monday a higher prevalence of patients with HAI (adjusted OR 1.08, 95%CI 1.05–1.10). There was a crude positive trend between number of weeks and prevalence, but the number of days involved in data collection was finally not associated with the prevalence of HAI, once adjustment for hospital size was made. The percentage of repeated records increased linearly with hospital size, and the frequency of infections was higher within this group (OR 2.8, 95%CI 2.6–3.0). The results of this study highlight the need for encouraging hospitals to shorten the time spent in obtaining a prevalence survey. If it is impossible to carry out the survey within the limits of one day, data collection should then be limited to that period of the week, Tuesday to Friday.