The performance of marginal structural models for estimating risk differences and relative risks using weighted univariate generalized linear models
Austin PC. Stat Methods Med Res. 2024; Apr 24 [Epub ahead of print].
Aim — To determine factors associated with home death in patients with cancer in Ontario, particularly to assess the association between death at home and (1) patients' rural/urban residence and (2) neighborhood income in urban areas.
Materials and Methods — We conducted a retrospective cross-sectional study in Ontario (2003-2010) using linked administrative databases. In order to account for clustering phenomenon, multivariable generalized estimating equation model was used to evaluate factors associated with home death. Analysis was performed in both rural and urban areas. For urban areas, neighborhood income was tested as a determinant of the place of death.
Results — A total of 193 783 deaths were analyzed, 9.1% of which occurred at home. In urban areas, home death was more likely for patients living in richer neighborhoods (odds ratio 1.69 for the highest compared to lowest neighborhood income quintile, 95% confidence interval: 1.54-1.86). The odds of dying at home when living in a rural area were no different from those living in the poorest urban neighborhood. Other variables associated with lower odds of home death were comorbidity index, certain cancers, and year of death.
Conclusion — The likelihood of dying at home significantly increases with living in higher-income urban neighborhoods and decreases with rural residence. Urban neighborhoods with lowest income have odds of home death similar to rural areas. These findings underline the importance of targeting proper populations for public support at the end of life.
Raziee H, Saskin R, Barbera L. J Palliat Care. 2017; 32(1):11-8. Epub 2017 Jun 29.
The ICES website uses cookies. If that’s okay with you, keep on browsing, or learn more about our Privacy Policy.