Este Blog resume (en inglés) el trabajo “Partners in Care! Health effects of providing personal care to spouses or partners”, por Luís Filipe, y que obtuvo el Premio SESPAS a la mejor Comunicación de Salud Pública en las XXXVIII Jornadas de Economía de la Salud celebradas en Las Palmas de Gran Canaria del 20 al 22 de junio de 2018.
Ageing of population brings new challenges to developed countries. Increasing life expectancy comes with more years to be lived at latter stages of life, and consequently, frailty. The lack of capacity to perform everyday tasks, due to the decline in physical and mental capabilities, force the elderly to resort to the support of a third party (family, friends or neighbours), thus affecting other members of society. This paper chooses a subcategory of potential informal caregivers, spouses and partners, and assesses the consequences of providing personal care on the health outcomes of these carers. Partner caregivers are a very special group due to their unique characteristics. Unlike other informal providers, spouses and partners tend to be about the same age as the care recipient and live under the same roof. The type of tasks they perform exclusively as personal care providers, as well as the associated burden, may differ from conventional children caregivers.
The study uses data from the fifth and sixth waves of the Survey of Health, Ageing and Retirement in Europe (SHARE). The dataset includes a partner caregiver identifier and a matching code for partners, thus allowing the researcher to explore the link between carers’ outcomes and patients’ characteristics.
The paper compares partner caregivers to other individuals whose partners are not receiving any type of personal care. The comparison is made using physical health and mental health as outcomes.
The results are presented in two steps. First, the paper evaluates the difference between models with and without the care recipients’ characteristics. In this context, implementing a model conditioning on patients’ health states partially filters effects such as emotional and (household) behavioural effects. Second, the study proceeds with the implementation of a propensity score matching estimator due to the unbalanced structure of the sample.
Given the categorical nature of the dependent variables, the study runs an Ordered Probit regression for each outcome. When patients’ characteristics are not considered, the estimated coefficients on the partner providing personal care are all indicative of worst physical health and depression levels. However, the inclusion of those variables causes relevant changes in the results. In the alternative methodology, providing care has a positive effect on self-assessed physical health and a (negative) non-statistically significant effect on depression. Matching analysis shows positive but non-significant results for both outcomes when conditioning the matched group on the same partner characteristics.
Two conclusions are drawn from this paper. First, the health of the partner caregiver is correlated with the health of his/her counterpart, being through health investment decisions (e.g. consumption of health care and prevention or behavioural risks/habits) or emotional aspects. Therefore, lower health of the patient increases the likelihood of lower health status of the corresponding partner caregiver, which is not a consequence of the isolated act of caring, but rather a consequence of both members of the household sharing health investment decisions and emotional links. Thus, all analysis on caregiver burden in co-residential care must account for this selection problem. Second, personal care provided by spouses/partners is potentially welfare improving in terms of physical health, likely due to higher activity indices.