Evaluation of the impact of a hospice at home service on place of death

Fewer patients are able to die at home than would wish to do so. A literature review showed that palliative home care patients are more likely to die at home than others. However, findings may be due to case mix differences, as variables which are positively associated with home death, are also positively associated with access to palliative home care. The thesis investigated the impact of a hospice at home (HAH) service on place of death. An observational, case control study compared 121 patients referred to HAH with 206 patients not referred. Multivariate logistic regression analysis showed that HAH care was strongly positively associated with home death. However, case mix effects could not be discounted. A randomised controlled trial (RCD compared 186 patients allocated to HAH care with 43 controls. Analysis was intention to treat. Intervention patients were not significantly more likely to die at home than control patients (67% versus 58%). The RCT suffered loss of power and dilution of the treatment effect. Post hoc multivariate logistic regression analysis showed a positive association between actual HAH input and home death. However, this association was no stronger than that between less intensive home care services and home death. Concerns about case mix effects remained. Content analysis of professional and family carers' explanations of endstage inpatient admissions for RCT patients suggested such admissions were mainly perceived to result from factors unrelated to insufficient home care. While some patients may have benefited from added home support, this may not have affected place of death. Additional home care introduced on top of good existing provision, among patients who are already likely to die at home, may have little impact on home deaths. Careful consideration of service aims, target group, other health service context is required before introduction of further hospice at home services