![[ Visit the Fifth Decennial online ]](images/banner.gif)
Background: Mandatory reporting of hospital-onset MRSA (HO-MRSA) rates requires standardized definitions for inter-facility comparisons. However, there is variability in how hospitals define HO-MRSA events (numerator counts). Despite the CDC definition of >48 hours, hospitals commonly use >2 or >3 calendar days after admission for ease of case finding. These differing definitions can alter hospital rankings since a >3 day definition will find fewer cases. This can be somewhat corrected by using at-risk denominators compared to total patient days.
Objective: To assess the impact of differences in numerator and denominator definitions on HO-MRSA incidence density among hospitals in a large metropolitan county.
Methods: We conducted a countywide survey of Infection Control and Prevention programs in Orange County, California hospitals to obtain definitions and values of HO-MRSA rates for a consecutive 12-month period from 2007-8. We then calculated HO-MRSA incidence density using each facility's cases as numerators and evaluating both total and at-risk patient days as denominators. At-risk denominators for each facility were determined by subtracting the first 2 or 3 hospital days of each admission depending on the facility's definition of HO-MRSA events. Hospital rankings using total and at-risk denominators were compared in quartiles. In addition, t-tests were used to assess differences among actual HO-MRSA rates in the top and bottom quartile.
Results: Of 21 hospital respondents, 5 (24%) hospitals were excluded due to use of a ≥48 hour case definition since at-risk patient days could not be calculated. Among the remaining 16 (76%) hospitals, 50% used >2 calendar days from admission, and 50% used >3 calendar days from admission for HO-MRSA events. Correction for at-risk days resulted in a median increase in HO-MRSA rates of 76% (hospital range 26-176%). HO-MRSA rates were significantly different when comparing rates in the best and worst quartiles regardless of denominator choice (t-test, p≤0.01). In contrast, when rankings were evaluated, 38% of hospitals changed quartiles when using at-risk denominators.
Conclusions: Despite national attempts at standardization, hospitals use non-uniform definitions for HO-MRSA case finding. Selection of an easy-to-apply standardized definition is needed for inter-facility comparison. If calendar day definitions are to be used, then either >2d or >3d definition should be enforced. In addition, use of at-risk denominators increased rates across a range of 26 to 176%, and influenced hospital rankings. Lack of standardized definitions and use of total instead of at-risk patient days are only part of the concern with public reporting. Further work is necessary to discover the impact of other effects on HO-MRSA rates among hospitals, such as case mix adjustment.
