Cluster Analysis for Clostridium Difficile Colitis National Burden Estimates

• Activity Number: 75
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Yi Mu*+
• Companies: CDC
• Keywords: Cluster analysis ; Survey ; Sampling ; Clostridium Difficile Colitis ; Infection ; National burden

Abstract

Clostridium Difficile Colitis (CDI) is an important public health problem. To generate accurate estimates of CDI nationally we rely largely on data from the Emerging Infection Program (EIP), EIP conducts active laboratory- and population- based surveillance for CDI in selected counties across 10 U.S. states, representing 3.6% of U.S. population. Previous study showed rates of reported CDI differ by population characteristics such as demographic traits, social economic traits, access to care and case identification techniques. Current national burden estimates are model based and account for some of the site variations, but the adjusted CDI incidences are weighted equally across sites. In this study, we use clustering analysis to identify groups of states with populations represented by specific EIP sites. Separate weights are calculated for each individual cluster. The estimated CDI national rates are 51.8 (43.48, 59.98)/100,000 and 54.45 (50.36, 58.50)/100,000, based on uniform and clustering weights, respectively. As a conclusion, uniform weight approach might underestimate the national burden and clustering approach can reduce variances and produce narrower confidence.