Background: Temporal aspects of suicide risk have been relatively neglected in the investigation of significant life events (LEs). Treating social risk factors for suicide as crude binary states ignores temporal transitions in suicide risk in the anticipation and aftermath of a significant LE. It misses opportunities to identify the points of greatest risk, at which suicide prevention interventions might be targeted pre-emptively. Such work has potentially greatest impact for commonly occurring risk factors such as divorce or unemployment.
Aims: We will address this evidence gap by applying new epidemiological methods to understand the temporal nature of suicide risk after divorce and unemployment. We will conduct a set of population-based studies over 44 years of follow-up to identify high-risk periods for mental illness, substance use and suicidal behaviour in the process of these LEs. Our work will build an understanding of the mechanisms underlying the links between mental illness, substance use, divorce, unemployment, and suicide, gaining insights into preventive opportunities.
Hypotheses: That specific time points in the process of two LEs (divorce; unemployment) represent periods of greatest risk for i) suicide/suicide attempt, ii) depression, and iii) substance use disorder (SUD). We predict that i) the risk of SUD increases three years before divorce and one year before unemployment; ii) the risk of depression and suicide attempt emerge one year before the LE; iii) the risk of suicide/suicide attempt, depression, and acute SUD increase during a) all birthday periods and b) all Christmas/New Year periods between the LE and re-employment/re-marriage. We will also test whether there is any heterogeneity by sex or social connectedness.
Methods:
Sample: We will analyse population-based registry data covering the entire Danish population from 1980-2023 (250,143,030 person-years at risk).
Measures: We will define divorce, end of cohabitation and end of employment using Danish registry variables that identify the date of status change with a high degree of granularity. Our primary outcomes are defined as i) suicide, ii) medically serious self-harm (termed suicide attempt), iii) both combined and our secondary outcomes as iv) depression and v) SUD.
Procedures: We will use advanced statistical methods such as the self-controlled case series approach to discover when individuals undergoing significant LEs are at the highest risk of these outcomes. Our analysis will take into account time-varying confounders (e.g. age, sex, household/ individual income, educational level, household size, seasonality of suicide and period effects), fixed confounders (e.g. country of origin as a proxy for ethnicity, and prior psychiatric disorder/SUD/suicide attempt), and modifiers (sex and social connectedness).
Potential impact: Our findings will clarify mechanisms and periods during which clinicians and policymakers should target pre-emptive support. We will set the highest research standards, using study pre-registration and sharing publicly accessible code. This will provide a model and benchmark for future studies investigating temporal risk transitions for other LEs.
Next steps: Our studies will facilitate the development of evidence-based strategies for timely interventions and support around divorce and unemployment, thereby contributing to more targeted public health initiatives for individuals navigating common but challenging life transitions.
