Diversity of End-of-Life Care in the Boom of Deaths at Old Age

The trend of prolonging life expectancy in the world population, though interrupted by the COVID-19 pandemic, is expected to resume soon, but whether the reduced mortality comes with less morbidity in later life is still debatable.¹ Despite such uncertainty, many aged societies in our world have started to see an unprecedented boom of deaths in their old-age populations. As highlighted by Ebeling et al. in the current issue of AJPH (p. 786), for example, Sweden will witness an annual increase of deaths by 25% in the next 30 years. By 2050, more than 30% of the world’s countries are predicted to reach or surpass the current level of population aging in Sweden.² Thus, there has never been a better time to promote studies on death and dying among older people.

Ebeling et al. investigated changes in end-of-life care use among Swedish older people (≥ 70 years). It is comparable to an early study by Aaltonen et al.,³ which examined the end-of-life care transfer of older people in Finland (≥ 70 years). Both studies utilized data on care use from the official registers of the respective countries and classified how end-of-life care is arranged at the population level. Based on data at two time points—two years and six months prior to death—the Finland study focused on care transfer between community, hospital, and residential care institution, reporting that about 60% of home dwellers moved to institutions or hospitals within the last six months of their lives. The current study on Sweden instead investigated a wide range of care use one year before death and at death, including home-based and institutional care as well as four medical services of clinical, acute, outpatient, and inpatient care. Six types of care use were identified: about 40% of people were categorized according to two “dependent” types (staying in a residential care institution and receiving either more or less medical care), whereas the other four types were all home dwellers, most receiving medical care. The current study, together with the Finland study and other related literature, revealed a great diversity of end-of-life care patterns, suggesting that the last stage of later life is uneven regarding where, when, and how various kinds of care are needed and delivered. This sends a strong message to stakeholders of the elder care sector, particularly given the escalating old-age deaths in many societies of the world.

Unlike in the Finland study, end-of-life transitions between home and a residential care institution were not common in Sweden as revealed by Ebeling et al. The use of medical services, however, was intensive in Sweden, although inpatient care was less utilized among older individuals living at home. One limitation of the current study, however, is the information shortage for the year prior to death. Taking into account the care situation at sixth months prior to death, which the literature shows is an important time-of-care changes in dying,^3,4 may help to capture more care transitions and yield a more valid typology. Moreover, the care classification in the current study was based on the Latent Class Analysis, a data-driven approach, so the conceptualization of these six types was based on summarized features of data clusters rather than direct empirical evidence. In this regard, there were efforts in the article to further examine the profiles of age, gender, and causes of death across the six groups, which enhanced the legitimacy of the classification. However, black boxes remain. For example, what kind of cancer was behind the group “terminal ill”? What type of cardiovascular disease was responsible for “sudden death”? And why were men more likely to be in the “terminal ill” group but women in the “dependent” group? Further work is called for to clarify mechanisms underlying the observed typology.

One essential point raised by Ebeling et al. is the contrast between the observed patterns of end-of-life care and the standard of a “good” death, which is expected to entail less care and more self-control. Such inconsistency is in line with the literature, which shows that a significant proportion of older people did not die at home even if they preferred to.⁵ As implied by findings of the current study, the trend of population aging is leading to more old-age deaths in residential care facilities. This may challenge the mainstream model of aging now in place and the ongoing movement toward deinstitutionalization.⁶ With the mounting number of deaths with institutional long-term care, promoting aging-friendly residential care facilities with good end-of-life services becomes urgent. In this regard, as highlighted by Ebeling et al., how to help seniors afford such services is a valid concern, especially considering the present social inequality in dying.⁷

The multiple types of care use identified by the current study are considered by the authors to represent “fast and slowly progressing trajectories,” the latter of which featured deaths in oldest old ages with a greater variation in causes of death. Consequently, the authors raised a meaningful hypothesis: the longer life span may partially come from a prolonged dying process with substantial care needs. This echoes well the classical hypothesis of morbidity expansion with population aging,⁸ which predicts more morbidity with population aging among older people (failure) when life expectancy extends (success). In the ongoing debate between the failure of success versus success of success (or benefits of success vs costs of success),⁹ the current study reminds us to give more weight to the end-of-life stage.

Lastly, the current study’s applicability beyond Sweden is worth discussing, as the impact of increasing old-age deaths would differ by context. One recent article, for example, discussed the rise of old-age deaths in Japan, another world-leading nation in population aging.¹⁰ Taking Japan as a “death-laden society,” the author argued that the lack of end-of-life services, greater risk of lonely deaths, and debate on physician-assisted suicide could be eminent issues for Japan in the boom of old-age deaths.¹⁰ In particular, given the specific socioeconomic and institutional context of the study by Ebeling et al., follow-up studies may pay attention to informal care by family members, which was not accounted for in the current study but is prevalent in non-Western societies.¹¹