VALIDATING THE ACCURACY OF A NOVEL METHOD FOR IDENTIFYING HEALTHCARE-ASSOCIATED INFECTIONS
Author(s)
Lee J, Imanaka Y, Sekimoto M, Nishikawa H, Ikai H, Motohashi TKyoto University, Kyoto, Japan
OBJECTIVES: The identification of healthcare-associated infections (HAIs) is a necessary first step for elucidating infection patterns and evaluating control measures. While the potential advantages to using administrative data for this purpose are recognized, previous attempts have proved unsuccessful. We have developed a novel method of HAI identification based on a series of criteria involving antibiotic utilization patterns available in administrative data. The objective of this study was to validate the accuracy of our HAI identification method using comparative analyses with chart review-based identification. METHODS: We retrospectively and contemporaneously identified HAIs in 584 gastrectomy patients admitted to 4 Japanese hospitals (A~D) using both our method and chart reviews. Chart review analysis was based on CDC criteria for nosocomial infections. The accuracy of our method was tested using Cohen's Kappa coefficient to quantify the non-random agreement between the two methods, as well as sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). RESULTS: Overall HAI incidence was 21.6% based on chart review identification. Cohen's Kappa coefficient was 0.78, and sensitivity and specificity were 0.93 and 0.91, respectively. PPV was 0.75 and NPV was 0.98. CONCLUSIONS: These results indicate that our method of identification has high accuracy, and may be used to estimate HAIs in large populations of patients at the hospital level and above. While the NPV was observed to be higher than the PPV, the relatively low prevalence of HAI incidence may have influenced this result. It is important to note that this method was not designed to pinpoint individual infections, but rather to identify identifications in a large number of patients using administrative data with a far higher accuracy than previously available. Therefore, our method has applications for HAI identification in large groups of patients, elucidating HAI patterns and trends, downstream economic evaluations and multi-institutional comparisons.
Conference/Value in Health Info
2010-09, ISPOR Asia Pacific 2010, Phuket, Thailand
Value in Health, Vol. 13, No. 7 (November 2010)
Code
ME2
Topic
Real World Data & Information Systems
Topic Subcategory
Health & Insurance Records Systems
Disease
Infectious Disease (non-vaccine), Respiratory-Related Disorders