This paper estimates the impact of spousal bereavement on hospital inpatient use for the surviving bereaved by following the experience of 94,272 married Scottish individuals from 1991 until 2009 using a difference-in-difference model. We also consider the sample selection issues related to differences in survival between the bereaved and non-bereaved using a simple Cox Proportional-Hazard model. Before conducting these estimations, propensity score approaches are used to re-weight the non-bereaved to generate a more random-like comparison sample for the bereaved. We find that those bereaved who survive are both more likely to be admitted and to stay longer in hospital than a comparable non-bereaved cohort. Bereavement is estimated to induce on average an extra 0.24 (95% CI [0.15, 0.33]) hospital inpatient days per year. Similar to previous studies, we estimate the bereaved have a 19.2% (95% CI [12.5%, 26.3%]) higher mortality rate than the comparable non-bereaved cohort.
Keywords: difference-in-differences; hospitalisation; mortality; spousal bereavement.
Copyright © 2017 John Wiley & Sons, Ltd.