Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Treatment at the end of life forms a major component of aggregate health care expenditure. Expenditure, however, begins to increase several years before death and varies substantially across individuals. This paper investigates heterogeneity in expenditure profiles across a 36 month period preceding death using group-based trajectory models. A mixture of generalised linear models with four components fits the data best, and identifies decedents in to high cost late rise, medium-high cost late rise, medium-low cost, and low cost late rise expenditure profiles. Approximately 35% of the sample is allocated to the high cost late rise trajectory with average monthly expenditure of £493 36 months prior to death rising linearly for about 28 months before exponential growth to £4000 in the month preceding death. Health conditions at the beginning of the period increase the risk of being in a higher cost trajectory with cancer having the largest impact. The existence of concurrent morbidities substantially raises the probability of membership to the high-cost late rise profile group. A better understanding of the determinants of expenditure profiles in the run up to death contributes to informing policies aimed at mitigating costs while not compromising quality of care.