Hospital admission at the time of a postpartum psychiatric emergency department visit: the influence of the social determinants of health

Aims — Social determinants of health have the potential to influence mental health and addictions-related emergency department (ED) visits and the likelihood of admission to hospital. We aimed to determine how social determinants of health, individually and in combination, relate to the likelihood of hospital admission at the time of postpartum psychiatric ED visits.

Methods — Among 10 702 postpartum individuals (female based on health card) presenting to the ED for a psychiatric reason in Ontario, Canada (2008-2017), we evaluated the relation between six social determinants of health (age, neighbourhood quintile [Q, Q1 = lowest, Q5 = highest], rurality, immigrant category, Chinese or South Asian ethnicity and neighbourhood ethnic diversity) and the likelihood of hospital admission from the ED. Poisson regression models generated relative risks (RR, 95% CI) of admission for each social determinant, crude and adjusted for clinical severity (diagnosis and acuity) and other potential confounders. Generalised estimating equations were used to explore additive interaction to understand whether the likelihood of admission depended on intersections of social determinants of health.

Results — In total, 16.0% (n = 1715) were admitted to hospital from the ED. Being young (age 19 or less v. 40 or more: RR 0.60, 95% CI 0.45-0.82), rural-dwelling (v. urban-dwelling: RR 0.75, 95% CI 0.62-0.91) and low-income (Q1 v. Q5: RR 0.81, 95% CI 0.66-0.98) were each associated with a lower likelihood of admission. Being an immigrant (non-refugee immigrant v. Canadian-born/long-term resident: RR 1.29, 95% CI 1.06-1.56), of Chinese ethnicity (v. non-Chinese/South Asian ethnicity: RR 1.88, 95% CI 1.42-2.49); and living in the most v. least ethnically diverse neighbourhoods (RR 1.24, 95% CI 1.01-1.53) were associated with a higher likelihood of admission. Only Chinese ethnicity remained significant in the fully-adjusted model (aRR 1.49, 95% CI 1.24-1.80). Additive interactions were non-significant.

Conclusions — For the most part, whether a postpartum ED visit resulted in admission from the ED depended primarily on the clinical severity of presentation, not on individual or intersecting social determinants of health. Being of Chinese ethnicity did increase the likelihood of admission independent of clinical severity and other measured factors; the reasons for this warrant further exploration.