Evaluation of Risk Factors in Population-Based Incidence of Clostridium Difficile Infection Across Multiple U.S. Geographic Locations

• Activity Number: 350
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Yi Mu*+
• Companies: CDC
• Keywords: risk factors ; infection disease ; mixed model ; variance component ; surveillance

Abstract

The marked increase in Clostridium difficile infection (CDI) incidence and mortality across the United States based on hospital administrative data has made CDI a national public health challenge. To understand the population-level burden of CDI in the United States, reports of CDI infections from 7 states participating in CDC's population-based surveillance for CDI were studied. A mixed model was used to evaluate fixed effect risk factors across sites. Traditionally, a variable was kept at P = 0.05 significance level. We proposed, based on the traditional strategies, to further test and keep the variable if it reduced the variance components (random effects), regardless of its significant level. We calculated and compared the site specific direct standardization rates (SDRs) through the traditional and new modeling strategies using combined data as the standard population. Results showed that extra risk factors reduced the variations of SDRs between sites, indicating that variations between sites were partially explained by them. We concluded that using variance component reductions to further identify isk factors for heterogeneous data is appropriate.