Redefining the National Healthcare Safety Network's definition of catheter-associated urinary tract infections: the hazard of including Candida species.

allowed for cases of HCA UTI to be diagnosed on the basis of clinical symptoms only, meaning that we are not necessarily comparing like with like. However, in this scenario, the assumption is that the incidence would be underestimated in a laboratory system. Using the formula as it is presented by Rhame and Sudderth for converting the prevalence to incidence instead shows the laboratory-based incidence to be an overestimation. Earlier studies have commented on the questionable accuracy of this formula for converting incidence and prevalence data. There are important arguments for the poor performance of the conversion. First, the types of data being compared represent different diagnostic criteria. Second, reflecting the argument put forward by Hoare et al, patients who contract HCA UTI are generally older and have comorbidities. Therefore, using their time in the hospital from diagnosis to discharge to reflect duration of infection is unreasonable. Using antibiotic treatment as a proxy for infection duration greatly improved the comparability of the prevalence and incidence values, and we would recommend amending this formula to the following: and I p P # (LA/DRx) P p I # , in which I is incidence, P is prevalence, LA is (DRx/LA) mean length of stay for all patients, and DRx is the duration of infection treatment.