In 2012, the Suicide Prevention Resource Center challenged researchers and policy makers with this question: "How do we shift our thinking from a focus solely on the individual in crisis and move more intently to efforts to examine the communities where people live and work and the systems they visit to receive care?"
This proposed research responds to this challenge by examining how major life transitions in work (e.g., retirement, going on disability, changing jobs, unemployment) and housing (e.g., moving in with adult children, foreclosure) relate to suicide risk among middle-aged and older adults. This study will investigate how these transitions relate to suicidal behavior (i.e., passive suicidal ideation and completed suicide) using two complementary data sources: (1) Restricted Access Data from the National Violent Death Reporting System, the most comprehensive registry of suicides in the US (NVDRS: n>50,000 suicide decedents aged ≥ 50 years from 27 states, 2003 – 2016); and (2) the Health and Retirement Study (HRS), the largest nationally-representative longitudinal study of health and aging in the US (n~20,000 adults aged >50 years interviewed each biennial wave, 2006 – 2016).
We will use multiple approaches to investigate how employment and housing transitions relate to suicide risk in these two datasets. We will use Natural Language Processing machine learning algorithms to analyze how anticipated and actual transitions in work and housing relate to completed suicide in the NVDRS. We will use Marginal Structural Modeling, an approach for identifying causal relationships with time-varying confounders, to examine how transitions in work and housing contribute to passive suicidal ideation and related constructs (e.g., depression, hopelessness) in the HRS. Finally, we will leverage the fact that the HRS survey data are linked to mortality records and conduct a matched case-control study of completed suicide nested within the HRS cohort to explore about how major life transitions contribute to the linkages between suicidal ideation and completed suicide over time. By using two distinct analytic approaches and two independent data sources to triangulate on a common set of scientific questions, we aim to develop a more comprehensive understanding of the intersection between work and housing transitions and suicidal behavior than could be achieved by using either approach or dataset in isolation.
The findings from this project will help close persistent scientific gaps in our understanding of how suicide risk emerges, persists, and remits during major life transitions in mid- and later life. Finally, by working with policy stakeholders (e.g., UM Institute for Healthcare Policy Innovation, the National Governors Association, and others) we will undertake this research with the explicit intention of translating out findings into policies and practices aimed at preventing suicide during these life transitions, while also efficiently disseminating our findings to a wide range of agents and decision-makers beyond the healthcare sector.
