Is Partner''s Education Important to Understand Health Status at Older Ages? Insights from the Spanish Case
Jeroen Spijker, Autonomous University of Barcelona (UAB)
Jordi Gumà, Universitat Pompeu Fabra
We use data from around 1400 retired couples aged 66-80 from the 2012 Spanish EU-SILC sample. We apply logistic regression models to assess the association between individual health and own and partner’s SES status. We use additive and combined approaches to ascertain whether the latter provides any additional explanation of educational differences in health beyond what is obtained with an additive model.
Results show that adding partner’s educational attainment does not reduce the significant association between own education and health among men. Regarding women, including the partner’s educational level actually nullifies the association of her education on self-reported health. When we consider the combined educational variable for men, this is not significant, but among women we observe that they have worse health when both they and their partner are lowly educated. This suggests that economic and shared behavioural risks factors that lead to health deterioration are also mediated through the educational status of the partner when the partner has the same educational level.
Theprotective health effect of living with a partner is well established from anindividual approach. Less is known about health differences among those wholive with a partner. Combininginformation from both partners permits testing whether the frequently established positiveassociation between socioeconomic status (SES) and individual health is alteredby considering the combined own and partners SES.We built on previous researchthat showed that apartners educational level is only meaningful among low-educated middle-agedadults in Spainby analysing what occurs among elderly couples.
Thehealth advantage of co-residing individuals has been explained by a set ofbenefits related to living with a partner: First, being subjected to control bytheir social environment, particularly partners. Especially men who live as acouple may be encouraged to abandon risky behaviours more often observed amongsingle individuals, or to follow healthier practices. Second, the combination ofboth partners social nets creates a wider source of possible resources toreduce the impact of events that may have a negative impact on health, such asbecoming unemployed or the death of a loved one. And third, the optimization ofcommon resources as a result of economies of scale and specialization of taskswithin the household enhances the SES of both partners.
Inthis context, education facilitates the acquisition of knowledge regardinghealth-damaging behaviour and the ability to optimise the use of healthservices. Positive associations have also been found between partners educationalattainment and individuals health according to different outcomes, including mortality,self-assessed health orrisky behaviour.All studies (full paper will contain references) stated that the inclusion ofpartners education adds meaningful information to better understand healthinequalities, even after controlling for own educational level: the higher the partnerseducational attainment, the lower the probability of dying, the higher thechance of having good subjective health or having stopped with smoking. Acommon explanation was that couples usually pool their resources (material andnon-material). Based on the social causation hypothesis, which assumes that anindividuals education (or SES in general) affects material, behavioural andpsychosocial factors and that these in turn have effects on health, also the partnersSES must have an additional effect on health.
Dataon retired Spaniards aged 66-80 were obtained from the 2012 cross-sectionalsample of the European Union Statistics on Income and Living Conditions survey (EU-SILC)of 2012. EU-SILC has information on SES and health from all household members (privatehousehold is the sample unit).Only native heterosexual Spaniards living with a partner are analysed,irrespective of their marital status. The age criteria is applied to the oldestpartner only to ensure the same sample size for both sexes, leaving us withabout 1400 couples with complete information on all the variables.
Ourdependent variable is individual health status, measured as self-assessedhealth. The possible answers very good, good, fair, bad and verybad are dichotomised into good health (combining the first two) and less thangood health (combining the last three categories).
Ourmain explanatory variable educational attainment of both partners is obtainedby crossing the educational information from both partners. The original sevensurvey categories, based on the ISCED classification, were aggregated into primaryor lower secondary (compulsory education; ISCED 0-2) and upper secondary andtertiary (ISCED 3-6).
Wealso test several covariates known to affect the health status of individualsliving with a partner to ascertain whether the effect of our main variables ofinterest is mediated by other factors: Household ability to make ends meet, partnerhealth status, and age.
Logisticregression models are used to identify which of the individual- andpartner-level combination of educational attainment as well as the other factorsare associated with differences in health status among middle-aged and retiredindividuals living with a partner. We take a sequential approach to assesswhether there is a meaningful advantage in including both partners educationalattainment in combination instead of both separately. The analyses are performedseparately for men and women.
Results show that adding partnerseducational attainment does not reduce the significant association between owneducation and health among men (Table1). Regarding women, including thepartners educational level actually nullifies the association of her educationon self-reported health (Table2). When considering the combined educationalvariable for men, this is not significant, but women have worse health whenboth they and their partner are lowly educated. This suggests that economic andshared behavioural risks factors that lead to health deterioration are alsomediated through the educational status of the partner when the partner has thesame educational level.
Gumà,Spijker (2016). Are couples really a homogeneous cluster in terms of health inEurope? Paper presented at the EPC, Mainz.
Presented in Session 1170: Health, Wellbeing, and Morbidity