Geue Claudia, Briggs Andrew, Lewsey James, Lorgelly Paula
Health Economics and Health Technology Assessment, Institute of Health and Wellbeing, University of Glasgow, 1 Lilybank Gardens, Glasgow, G12 8RZ, UK,
Eur J Health Econ. 2014 Nov;15(8):885-96. doi: 10.1007/s10198-013-0543-7. Epub 2013 Nov 29.
Health care expenditure (HCE) is not distributed evenly over a person's life course. How much is spent on the elderly is important as they are a population group that is increasing in size. However other factors, such as death-related costs that are known to be high, need be considered as well in any expenditure projections and budget planning decisions.
This article analyses, for the first time in Scotland, how expenditure projections for acute inpatient care are influenced when applying two different analytical approaches: (1) accounting for healthcare (HC) spending at the end of life and (2) accounting for demographic changes only. The association between socioeconomic status and HC utilisation and costs at the end of life is also estimated.
A representative, longitudinal data set is used. Survival analysis is employed to allow inclusion of surviving sample members. Cost estimates are derived from a two-part regression model. Future population estimates were obtained for both methods and multiplied separately by cost estimates.
Time to death (TTD), age at death and the interaction between these two have a significant effect on HC costs. As individuals approach death, those living in more deprived areas are less likely to be hospitalised than those individuals living in the more affluent areas, although this does not translate into incurring statistically significant higher costs. Projected HCE for acute inpatient care for the year 2028 was approximately 7% higher under the demographic approach as compared to a TTD approach.
The analysis showed that if death is postponed into older ages, HCE (and HC budgets) would not increase to the same extent if these factors were ignored. Such factors would be ignored if the population that is in their last year(s) of life were not taken into consideration when obtaining cost estimates.
医疗保健支出（HCE）在人的一生中分布并不均匀。老年人的医疗支出数额很重要，因为这一人群规模正在不断扩大。然而，在任何支出预测和预算规划决策中，还需要考虑其他因素，比如已知较高的与死亡相关的成本。
本文首次在苏格兰分析了应用两种不同分析方法时，急性住院护理支出预测是如何受到影响的：（1）考虑生命末期的医疗保健（HC）支出；（2）仅考虑人口结构变化。同时还估计了社会经济地位与生命末期HC利用及成本之间的关联。
使用了一个具有代表性的纵向数据集。采用生存分析以纳入存活的样本成员。成本估计来自一个两部分回归模型。两种方法都获得了未来人口估计数，并分别与成本估计数相乘。
死亡时间（TTD）、死亡年龄以及这两者之间的相互作用对HC成本有显著影响。随着个体临近死亡，生活在较贫困地区的人住院的可能性低于生活在较富裕地区的人，尽管这并未转化为在统计学上显著更高的成本。与TTD方法相比，人口结构方法下2028年急性住院护理的预计HCE高出约7%。
分析表明，如果死亡推迟到老年，若忽略这些因素，HCE（和HC预算）不会以相同幅度增加。如果在获取成本估计时不考虑处于生命最后一年的人群，这些因素就会被忽略。
