Statistical Analyses of Longitudinal Observational Data from Adults Living in Marginalized Housing, with Particular Attention to Cognitive Function Characterization

An ongoing longitudinal study named by Hotel Study focuses on investigating the physical and mental health condition of people who lived in Vancouver's Downtown Eastside (DTES). The study's primary objective is to provide better healthcare services for the target population. Study participants could not follow closely the study's predetermined visit schedule and thus have caused their study data with nonignorable missing. That has motivated this project. We aim to understand the data missing mechanism and then account for it in analyses of the participants' cognitive test results. We begin with a descriptive analysis of all the study's available data. The analysis indicates that the study participants who were recruited from different sources represent different populations. That leads us to focus on the participants from single-room occupancy (SRO) hotels of the DTES in this project. We explore the frequency and rate of participants' study visits and their attendances to the cognitive tests over time by conducting regression analyses under various models. Participants who joined the study early appear to experience less missing visits. We conduct regression analyses of the study's available cognitive test scores with adjustment for the data missing. The test scores appear to be strongly associated with when the participants' joined the study. We also find that the participants' decision-making abilities are associated with their age and sex.