Grant: Project Grant

Depression is a chronic and disabling condition that is the leading global cause of life-years lived with disability. However, most Canadians with depression do not receive adequate care. Thus, the development and implementation of effective depression management programs is urgent. Depression screening has been recommended as a solution. However, when trials have been conducted, they have failed to show benefit from depression screening. One reason is that screening tools have not been accurately calibrated for use with individual patients. The CTFPHC has emphasized the need for high-quality research that generates unbiased estimates of depression screening tool accuracy and considers individual patient risk factors. Individual participant data meta-analysis (IPDMA), which uses actual patient data from primary studies, can address existing problems and incorporate important depression risk factors. Our team has built four large IPD databases including individual and study level characteristics. However, best practices for developing prediction models in the context of IPDMA are unknown because of a number of statistical challenges. Particularly, accounting for study specific intercepts and slopes, dealing with study level variables, and accounting for missing data have not been studied. The proposed research will advance and validate statistical prediction models in the context of IPDMA and apply the advanced models to generate personalized risk probabilities for MDD based on individual screening scores and depression risk factors in actual individual participant datasets, which has not been done previously. Results will be used to improve individualized depression screening and management programs and to make better decision for individual patients in primary, perinatal, and hospital care.