Surveillance methods for hospital infection.

Eight selective surveillance methods were assessed for their effectiveness in detecting hospital infection and the time required for data collection. The methods were compared with a reference method which was designed to identify all patients and infections in the study population (patients occupying 122 beds of a district general hospital). The selective methods were: - laboratory based ward surveillance - laboratory based telephone surveillance - ward liaison surveillance - laboratory based ward liaison surveillance - risk factor surveillance - temperature chart surveillance - treatment chart surveillance - temperature and treatment chart surveillance. The proportions of community acquired infection (CAI) and hospital acquired infection (HAI) detected by the selective surveillance methods varied; the highest proportion of CAI (70%) was identified by treatment chart surveillance, and temperature and treatment chart surveillance, and of HAI (71%), detected by laboratory based ward liaison surveillance. The time for data collection ranged from 1.5 hours/122 beds/week for laboratory based telephone surveillance to almost 8 hours for temperature and treatment chart surveillance. The time for the reference method was 22.1 hours/122 beds/week. Using the proportion of patients with HAI detected and time required for data collection to assess the methods, laboratory based ward liaison surveillance was the most effective and an efficient method. This method was revised minimally and introduced into six district general hospitals by infection control nurses. The time for data collection ranged from 4 to almost 8 hours/120 beds/week. The revision did not affect the proportion of HAI detected, however, the proportion of CAI identified was significantly reduced. The reproducibility of laboratory based ward liaison surveillance was good. The results will enable infection control teams to make an objective and rational choice of methods for the surveillance of hospital infection