Gender inequalities in health in later life: Does informal caregiving matter?

Damiano Uccheddu, Department of Sociology, University of Groningen
Anne H. Gauthier, NIDI
Nardi Steverink, Department of Health Psychology, University Medical Center Groningen (UMCG), University of Groningen
Tom Emery, NIDI

Numerous studies have shown that women generally report worse health conditions than men, even though they live longer. These gender inequalities in health may partially stem from caregiving responsibilities, precisely because they are often performed by women. This study aims at examining the mediating role of caregiving in the association between gender and health in later life. We used panel data from 28,109 individuals (76,845 observations) older than the age of 50 who participated in the Survey of Health, Ageing and Retirement in Europe (SHARE) between the years 2004-2015. We employed random-effects logistic regression models and the Karlson, Holm, and Breen (KHB) decomposition method – adjusting for important covariates – to decompose the total effect of gender on health in its direct and indirect components. Caregiving activities included spousal and parental care. Health was measured by self-perceived health, depressive symptoms (EURO-D scale), activities of daily living (ADL), and instrumental activities of daily living (IADL). Preliminary results suggest that women, for each of the four health outcomes, report worse health conditions than men. Contrary to expectations, the health gaps between women and men appear not to be associated with differences in caregiving responsibilities. In our further analyses, we aim at investigating the role of the broader institutional context in shaping the relationships between gender, care, and health. Moreover, we aim at examining in a more detailed way the role exerted by different types of care activities.

IntroductionResearch has so far highlighted several explanations forgender inequalities in health, typically referring to a combination ofbiological, psychological, behavioural, and social factors that can impact thehealth of women and men in different ways (Readand Gorman 2010). Among them, socioeconomic status (SES)is widely recognized as a key mechanism (Dentonand Walters 1999; Mackenbach et al. 1999; Östlin 2002).However, with very few exceptions (Bambraet al. 2009), there is little research into how SES may bedifferentially associated to the health conditions of women and men living indifferent institutional contexts.This comparative study contributes to the literature inaddressing the question of whether the magnitude of socioeconomic inequality inhealth – among the older part of the population – differs between gendersacross different welfare state regimes.Our theoretical expectation is that if the welfarestate decommodifies health as well as labour (Bambra2007),then there should be a less strong association between SES and health in highlydecommodifying welfare states (Sweden and Denmark). Since these lattercountries are characterized by a high commitment to gender equality, we alsoexpect higher levels of gender inequalities in health in countries thatreinforce traditional gender roles, and in those with higher reliance on family(Italy and Spain).  Data and methodsWe use data from 5 waves (2004-2015) of the Survey ofHealth, Ageing and Retirement in Europe (SHARE). The sample consisted of 59,660respondents (135,779 observations) of age 50 and older living in nine Europeancountries (Austria, Belgium, Denmark, France, Germany, Italy, Spain, Sweden,and Switzerland). Sample characteristics are reported in Table 1.The outcome variable, a 40-item frailty index (FI), isconstructed in accordance with standard procedures (Romero-Ortunoand Kenny 2012) to capture the multidimensional nature ofhealth at older ages. The FI is defined as a count of deficits divided by thetotal number of deficits evaluated (i.e. 40). The FI ranges from 0.0 (nodeficits present) to 1.0 (all deficits present). Cronbach’s alpha is 0.896.Gender and SES are the key explanatory variables. SESis operationalised using three indicators, namely education, income, andwealth. Education is based on the international classification ISCED-97: low(ISCED 0, 1 and 2), medium (ISCED 3 and 4), and high (ISCED 5 and 6).Country-specific and wave-specific quartiles of income and wealth (referring tothe year preceding the interview) were estimated at the household level and adjustedfor family size, consumer price index (CPI), and purchasing power parities(PPPs). Control variables include age, current job situation, marital status,number of children, drinking and smoking behaviours, SHARE waves, and countryof residence.Countries are classified into three generic welfareclusters (Bambra2007):Southern Europe (Italy and Spain); Western Europe (Austria,Belgium, France, Germany, and Switzerland); Northern Europe (Denmark andSweden).Linear hybrid models (Allison2009)were used to examine the associations between SES and frailty, controlling for time-invariantunobserved individual heterogeneity.  ResultsTable 2 shows the estimates of different sets ofhybrid regression models. It appears that once relying solely onwithin-individual variance, the longitudinal association between income andwealth with frailty is no longer statistically significant for both genders in Northern European countries (Denmarkand Sweden), and for men living in Southern European countries (Italy andSpain). Considering the between-effects, educational gradient in frailty appearto be more consistent for women, in all the welfare clusters, and lesspronounced – for both genders – in the Northern European countries.  ConclusionsThere are some insights, from these preliminaryanalyses, that may support the hypothesis that welfare state may play afundamental role in shaping socioeconomic inequalities in health betweengenders. References 

Allison,Paul D. 2009. Fixed Effects Regression Models. Thousand Oaks,California: SAGE Publications.

Bambra, Clare. 2007.“Going beyond The Three Worlds of Welfare Capitalism: Regime Theory and PublicHealth Research.” Journal of epidemiology and community health 61(12):1098–1102.

———. 2009. “Gender,Health Inequalities and Welfare State Regimes: A Cross-National Study of 13European Countries.” Journal of epidemiology and community health 63(1):38–44.

Denton, Margaret, andVivienne Walters. 1999. “Gender Differences in Structural and BehavioralDeterminants of Health: An Analysis of the Social Production of Health.” Socialscience & medicine 48: 1221–35.

Mackenbach, Johan P. etal. 1999. “Socioeconomic Inequalities in Mortality among Women and among Men:An International Study.” American journal of public health 89(12):1800–1806.

Östlin, Piroska. 2002. “GenderPerspective on Socioeconomic Inequalities in Health.” In ReducingInequalities in Health: A European Perspective, eds. Johan P. Mackenbachand Martijntje Bakker. London: Routledge, 315–24.

Read, Jen’nan Ghazal, andBridget K. Gorman. 2010. “Gender and Health Inequality.” Annual Review ofSociology 36(1): 371–86.

Romero-Ortuno, Roman, andRose Anne Kenny. 2012. “The Frailty Index in Europeans: Association with Ageand Mortality.” Age and Ageing 41(5): 684–89.
Appendix  

Presented in Session 1171: Health, Wellbeing, and Morbidity