A Model for Combining Case-Control and Cohort Studies in Systematic Reviews of Diagnostic Tests

• Activity Number: 359
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Yulun Liu*+ and Yong Chen and Jing Ning and Haitao Chu and Janice Cormier
• Companies: The University of Texas School of Public Health and The University of Texas School of Public Health and The University of Texas MD Anderson Cancer Center and University of Minnesota School of Public Health and The University of Texas MD Anderson Cancer Center
• Keywords: Composite likelihood ; Diagnostic accuracy study ; Meta-analysis ; Pseudolikelihood ; Systematic review

Abstract

Systematic reviews of diagnostic tests often involve a mixture of case-control and cohort studies. The standard methods for evaluating diagnostic accuracy provide estimates of only sensitivity and specificity and ignore the nature of cohort studies. This approach may be less efficient because it does not consider disease prevalence. More importantly, the standard methods do not provide estimates of the more clinically relevant measures, such as positive and negative predictive values. In this talk, we propose a hybrid model that jointly models the disease prevalence along with the diagnostic test sensitivity and specificity in cohort studies, and models the test sensitivity and specificity in case-control studies. In order to overcome the well-known computational difficulties in the standard full likelihood inference, we propose an alternative inference procedure based on the composite likelihood. We conduct extensive simulation studies to evaluate the finite sample performance of the two inference procedures and illustrate our method in a review of the performance of contemporary diagnostic imaging methods for detecting metastases in patients with melanoma.