Measuring device days versus patient days for patients with urinary catheters

BACKGROUND: Device-related infection rates are useful measures and have been calculated using number of discharges (admissions) as denominator. Today, the preferred measure calculates rates with a denominator based on device days. This measure is thought to provide a more accurate portrayal of the burden of infection but the data require considerable effort to collect. The objective is to determine whether patient “hospital days” can serve as a surrogate for “device days.” We undertook a study to compare urinary catheter–related infection rates using each of these measures. METHODS: This large, nationwide healthcare system introduced systematic, computer-based data collection several years ago. Data over a 2.5-year period were reviewed to determine infection rates as a percentage of the population, and infection rates per 1000 urinary catheter days. Rates for hospitals of similar size and location were compared. RESULTS: The study included observations on 99,834 patients during 2002–2004. The infection rates per 1000 days over the entire period were 4.72 per 1000 patient days and 6.48 per 1000 device days. The correlation of infection rates per month using device days compared to patient days was 0.978 (p < 0.001) using Pearsonʹs and 0.987 (p < 0.001) using Spearmanʹs. This shows that the values are not independent. CONCLUSIONS: Rates calculated on the basis of patient discharges correlated highly with rates calculated using device days. The high correlation between these rates suggests that for some purposes rates based on discharges can serve as a surrogate for rates based on device days. Such a change would reduce the data-gathering burden on infection control practitioners.