Background Hospice deaths are a small but increasing proportion of deaths in England (3.4% in 1993; 6% in 2012). Deprivation is associated with place of death; the gap between levels of hospice deaths in the least and most deprived areas is increasing. Proximity to healthcare facilities is one aspect of accessing services. There is some evidence that hospices in England are located in less deprived areas, but the relationship between distance to hospice and deprivation has not been previously examined. We aimed to investigate whether distance to the nearest hospice of decedents is associated with deprivation and varies according to place of death.

Methods This population-based study used individual death registrations in 2012 in South London and location of hospices, provided by ONS and the London Cancer Alliance respectively. The Integrated Transport Network (Ordnance Survey www.edina.ac.uk) was applied to calculate network distance between the usual residence and nearest hospice using ArcGIS (v10.1 ESRI). Deprivation was measured by the Index of Multiple Deprivation (IMD) score (2010) and linked to the decedent data at the Lower Super Output Area (LSOA) level; higher IMD scores reflecting more deprivation.

Maps visualised borough-level deprivation and distances to hospices in South London. IMD scores and distance to nearest hospices were evaluated using correlation analysis (Spearman’s ρ), for all deaths and for deaths stratified by place of death (hospice, home, hospital, care home).

Results 18,165 deaths of those aged over 25 years were included. Maps demonstrate the variation in median deprivation levels and distances to nearest hospice in South London boroughs. The correlation analysis shows a weak negative relationship between IMD score and distance to hospice (rho = −0.22; p < 0.0001). By place of death, IMD scores of care home deaths and distance to hospice have the strongest negative correlation (−0.27; p < 0.0001). The IMD scores of those who died at home (−0.21; p < 0.0001), hospice (−0.19; p < 0.0001) and hospital (−0.21; p < 0.0001) are similarly correlated to distance.

Conclusion Decedents residing in more deprived areas were closer to hospices than less deprived. These results were similar for people dying at home, in hospices, hospitals or care homes. There are likely other additional factors accounting for differences in place of death between different deprivation levels. Further analysis to assess levels of association in a wider geographical areas with adjustment of potential confounding factors could provide more evidence for service providers, commissioners and policy makers to address inequity of access to hospices.