Determinants of Hospital Death for Taiwanese Pediatric Cancer Decedents, 2001-2010

Context Factors influencing pediatric cancer patients' place of death may have evolved with advances in medical and hospice care since earlier studies were done. Objectives To comprehensively analyze factors associated with hospital death in an unbiased population of pediatric cancer patients in Taiwan. Methods This was a retrospective cohort study using administrative data for 1603 Taiwanese pediatric cancer patients who died in 2001-2010. Place of death was hypothesized to be associated with 1) patient sociodemographics and disease characteristics, 2) primary physician's specialty, 3) characteristics and health care resources at both the hospital and regional levels, and 4) historical trends. Results Most Taiwanese pediatric cancer patients (87.4%) died in an acute care hospital. The probability of dying in hospital increased slightly over time, reaching significance only in 2009 (adjusted odds ratio [AOR], 95% CI: 2.84 [1.32-6.11]). Children were more likely to die in an acute care hospital if they resided in the most urbanized area, were diagnosed with leukemia or lymphoma (2.32 [1.39-3.87]), and received care from a pediatrician (1.58 [1.01-2.47]) in a nonprofit proprietary hospital (1.50 [1.01-2.24]) or large hospital, reaching significance for the third quartile (2.57 [1.28-5.18]) of acute care hospital beds. Conclusion Taiwanese pediatric cancer patients predominantly died in an acute care hospital with a slightly increasing trend of shifting place of death from home to hospital. Propensity for hospital death was determined by residential urbanization level, diagnosis, primary physician's specialty, and the primary hospital's characteristics and health care resources. Clinical interventions and health policies should ensure that resources are allocated to allow pediatric cancer patients to die in the place they and their parents prefer to achieve a good death and promote their parents' bereavement adjustment.