Latent Class Analysis for Asthma Patient Outcome Measures

• Activity Number: 547
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Wen-Ling Kuo*+ and Davis Gates and Ruji Yao
• Companies: Merck and Merck and Co. and Merck
• Keywords: Latent Class Analysis ; Asthma Disease Activity Scores ; Asthma Attack ; Multivariate Analysis ; Patient Outcome Measures

Abstract

Asthma Disease Activity Score, a newly derived composite score, has been developed to capture multiple aspects of asthma outcome measures. A disease severity indicated by this composite score is used to predict future asthma attack. However, composite measures are not without the usual assumptions and associated limitations.

Latent class analysis (LCA) is a multivariate approach to classify individuals based on their responses to a set of observable categorical variables into a set of mutually exclusive latent classes. Unlike other statistical methods, LCA makes no assumptions about distributions of the indicators. Thus, it serves an alternative way to assess asthma disease severity based on multiple asthma outcome measures.

This paper will apply the LCA method to patient outcome measures and compare the composite score method with the latent class model through an examination of data from clinical trials.