Questionnaires were sent to 360 UK medical microbiologists to determine the incidence of Clostridium difficile infection in the UK between 1993-1996, and to establish the current laboratory testing protocols. Replies were received from 104 laboratories (29% response rate), 86, 7, 4 and 3% of which are in England, Scotland, Wales and Northern Ireland, respectively. The laboratories serve a total of approximately 90,000 hospital beds (median 750). C. difficile testing was performed by 83% of the laboratories, and 52, 45 and 31% used toxin A +/- B kits, cell cytotoxicity, and culture +/- isolate toxigenicity testing methods, respectively. Forty-seven percent of laboratories only performed testing when specifically requested, 19% if antibiotic use was stated, 15% tested all diarrhoeal specimens, and 14% examined all specimens except those from infants and community patients. The annual totals of positive C. difficile reports and cases increased from 3132 to 12,775, and from 1576 to 8211, respectively, between 1993-1996. In 1993 C. difficile infection caused ward closures in 5% of hospitals but 16% in 1996. Antibiotic policy changes, due to C. difficile infection, occurred in 21% of hospitals in 1996 compared with only 4% in 1993. C. difficile infection appears to be increasing markedly in the UK with major implications on hospital inpatient activity. Wide variations in laboratory selection and testing methods are likely to be masking the true epidemiology of C. difficile infection, and consensus is required on optimal protocols.