In-Home Care at the End of Life—How Much Is Needed?

We spend a large proportion of our lives in our home environment, so on face value it is a natural assumption to consider dying at home an important goal for quality end-of-life care.¹ As the need for care at the end of life grows in the face of an aging population, understanding how best to configure and fund services for people approaching the end of life is increasingly important. It requires consideration of what outcomes we are aiming to achieve and for whom (the patient, their carer, and/or the health system), and whether those who are at the end of their life prioritize these factors similarly.² In terms of models of care, the challenge is to determine the role, configuration, and optimal timing for in-home services, as well as the proportion of overall services that should be performed in the community. When home is the preferred place of death, in-home services are critical to ensure we can meet this preference while maintaining adequate symptom control and support for both the person themselves and their caregivers.³ Meta-analyses support increased odds of dying at home when adults with advanced illness receive home-based palliative care, with reduced symptom burden.³

The retrospective cohort study from Abe et al⁴ using claims data from 572 059 decedents in Japan’s long-term care insurance system found the proportion of in-home deaths was higher in those who had used in-home care services at least once prior to their death (14.3%), and that each day increase in the use of in-home care services at the end of life was associated with a 3.6 percentage point increase in the probability of in-home death.⁴ Older people who died at home were more likely to have a spouse and have lower care needs.⁴ The use of instrumental variable estimation aimed to account for the expectation that if the older person lived in a municipality with more care workers, they could use in-home services more frequently. It is important to note this study assessed the in-home services provided by care workers to assist with instrumental and activities of daily living, which did not include in-home medical services provided by physicians and nurses (such as palliative care), and focused on services during the narrow time frame of the month prior to death. The authors highlight limitations of population-based approaches. These include the difficulty in classifying the number of unintended home deaths (where a home death was not the preference) and that those who have not received in-home services could potentially have either not needed services or been affected by other factors driving inequity of access, which would lead to an underestimation of the population denominator of older people who may have benefited.

This study highlights the complexity of defining the “exposure” in this context. Most health systems include several elements of in-home care, as identified in this example of Japan’s long-term care system, which includes direct physical and personal care as well as medical and nursing expertise for specific clinical services such as palliative care. It is likely that a combination of these services, delivered when and where they are needed by clinicians and care workers with the required competencies, will be needed to best optimize outcomes, whether this is meeting preferred location of death or other aspects rated important by people at the end of their lives.²

This study also has made an assumption that it is care close to the end of life that influences capacity to meet preferences, but increasingly we are understanding palliative care needs occur much earlier, and that early intervention (including care delivered during hospitalization or in ambulatory care) improves outcomes. As the authors rightly conclude, future work to fully understand how best to meet end-of-life care at a population level will require analysis to assess the impacts of exposure to all elements of in-home care, including clinical services such as palliative care and out-of-hours care, in the local context of availability of other care options at the end of life (eg, hospitalization and institutional care). Similarly, we need to also ensure our models of care are cost-effective,⁵ and the models are inclusive of those who may be at higher risk of not having their needs met (eg, those without an informal carer or with higher physical needs).

Overall, although this study has demonstrated that in-home care worker services do increase the rate of in-home death, the proportion of deaths at home are at the lower end in comparison with other countries.⁶ This raises 2 issues. Policymakers need to have a good understanding of epidemiology about preferences for location of care and death,¹ and how these may change over the course of an illness, for the population for whom they are developing service models. Similarly, population-based methods to evaluate health services need to use the thresholds based on known preference for location of death within their community to evaluate success. Researchers also need to account for a composite of outcomes, in which case location of care and death is only 1 element. Preference for location of care and/or death is not a simple dichotomous outcome; a number of trade-offs inform these decisions, and clinical experience tells us that the priorities of the person with a life-limiting illness and their carers often present a moving goalpost.¹ The preference of location of death at the top of the hierarchy cannot be assumed, with a 2021 study⁷ using discrete choice methodology to consider preferences for an older person with advanced cancer in the last 3 weeks of life finding that people prioritized patient and carer well-being and symptom control, with less importance placed on location of care and death. Similarly, studies have shown that the home of a relative or family member is not considered an equal alternative to the person’s own home.¹ Other studies of home-based palliative care have included a range of outcomes, including physical symptom control, psychological, social, and spiritual well-being, and caregiver outcomes such as burden, mastery, coping, and effects on bereavement.³ Other important elements that patients identify include trust in the treating clinical team, effective communication, and goals around life completion (resolving conflicts and saying goodbye).²

Population studies will continue to play an important role in policy and service planning to ensure our health care systems can meet the growing need for end-of-life care, and there is work that can be done to optimize our administrative and clinical data sets to make these more fit for purpose. However, they will need to be supported by prospective comparative effectiveness studies of models of care evaluating impacts on end-of-life needs and preferences, and optimal cost-effectiveness.

Published: November 8, 2021. doi:10.1001/jamanetworkopen.2021.33019

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Agar MR. JAMA Network Open.

Corresponding Author: Meera R. Agar, MBBS, PhD, Professor of Palliative Medicine, IMPACCT, Faculty of Health, University of Technology Sydney, 235 Jones St, Ultimo 2007 Australia (meera.agar@uts.edu.au).

Conflict of Interest Disclosures: None reported.