Incidence Trajectories of Psychiatric Disorders After Assault, Injury, and Bereavement
HRs for any psychiatric disorder among individuals exposed to potentially traumatic events compared to matched unexposed individuals or unexposed full siblings from 1987 to 2013. Time-varying HRs and 95% CIs were calculated using flexible parametric survival models, allowing the effect of potentially traumatic events to vary over time. A spline with 4 df was used for the baseline hazard, while 2 df was used for the time-varying effect. Models were stratified by matching identifier (birth year, sex, and birthplace) for population comparison or family identifier for sibling comparison, controlling for age at index date, sex, educational level, family income, marital status, history of somatic diseases, and family history of psychiatric disorders (in population comparison). The longest follow-up time was 17 years for assaults or transportation incident, as data were available from 1997.
HRs with 95% CIs for any psychiatric disorders among individuals exposed to potentially traumatic events compared to matched unexposed individuals or unexposed full siblings, overall and by time during follow-up (<1 and ≥1 year). Estimates were calculated from Cox models with time since the index date as the underlying time scale. Models were stratified by matching identifier (birth year, sex, and birthplace) in population comparison or family identifier in sibling comparison, controlling for age at index date, sex, educational level, family income, marital status, history of somatic diseases, and family history of psychiatric disorders (in population comparison).
