New rapid methods for the diagnosis of CDI are constantly being developed. One of these, the Oxoid toxin A test (Unipath, Basingstoke, Hampshire, UK), was evaluated by comparison with a cytopathic effect (CPE) method, using 100 strains of C. difficile. The performance of the Oxoid toxin A test was only as good as the CPE method read at 6 hours, with a sensitivity of 725 and a specificity of 94%. This is not likely to be sufficiently accurate to be relied on as a single test for CDI. Furthermore, with the advent of disease caused by toxin A negative / toxin B positive strains, toxin A tests are no longer to be recommended for CDI diagnosis. The epidemiology of CDI was studied in two ways. 1) The prevalence of C. difficile toxin A negative / toxin B positive strains was measured in 269 isolates from symptomatic hospital patients in Leeds and Bradford, using toxin gene detection PCR. 2) A comparison of C. difficile strain type was made between isolates from symptomatic community patients in Leeds and Truro, using ribo-spacer (RS) PCR. There were 15 isolates from Leeds and 39 from Truro. The epidemiological investigation indicated that there were no toxin A negative / toxin B positive strains apparent in Leeds or Bradford, although these strains have been reported from other areas of the country. Amongst Leeds community patients, 60% of strains isolated were identical to the 'endemic' hospital strain (which accounts for over 80% of sporadic CDI cases in Leeds hospital patients). By contrast, this strain was not present in the Truro population (p= <0.0001). In order to examine the likelihood of CDI following treatment with either cefotaxime (CTX) or piperacillin-tazobactam (PT), a prospective, ward based, crossover study was carried out. This was performed on two well matched care of the elderly wards, with patients who required empirical broad-spectrum antibiotic treatment. Although only 48 patients were enrolled in this study, a highly significant difference was noted in the incidence of CDI and C. difficile colonisation between the two groups. The odds ratio (OR) for CDI in CTX treated patients compared with those who received PT was 14.6 (CI = 1.7-124.7). The number needed to treat (NNT) to save one case of CDI was 2.18.