Background

Research onsocio-economic differentials in health and well-being in later life has foundthat, although inequalities persist, they tend to diminish with increasing ageand are less marked for women than men (Foverskov et al 2016). This attenuationhas been attributed to various factors, including selective survival,inadequacy of measures such as past occupational status to fully capturevariations in older women’s socio-economic resources, and the availability ofage related state supports, such as pensions, and, in the US, health care whichmay mitigate the impact of inequalities. Other domains of life, includingpartnership, parenthood and social participation, also influence health andwell-being and it is increasingly recognised that these domains may haveinteractive or offsetting effects (Grundy and Holt 2000; Elder et al 2003;McDonough et al 2015). This is especially true for women, who traditionallyhave been more likely than men to have to prioritise between more workorientated and more family orientated roles. There has been a long standingdebate on the health implications of holding multiple roles and increasinginterest in the longer term implications of work-family life courses (McDonoughet al 2015; Sabbath et al 2015; Lacey et al 2016; Benson et al 2017). Researchindicates that women’s work and family histories are associated with well-beingin later life, with some apparent variations between the USA and the UK. Several of these studies consider women in later mid-life, rather than at olderages, and examine outcomes measured at one point in time. However, the salienceof socio-economic and socio-demographic indicators  may change as people age.For example, the status and stimulation accruing from a professional career maybecome less relevant to well-being many years post retirement whereas supportand social interaction with family and community may become more important.Here we examine how work/family orientation is associated with changes in thehealth and well-being of US women over a five year follow-up.

Analyses andresults

We use six wavesof data from the National Health and Aging Trends Study (NHATS), collected2011-2016 from a nationally representative sample of Medicare beneficiariesaged 65 and older. Our sample consists of 5,572 person-years nested in 1,342American-born women who were aged 65-74 at baseline and had no missing valueson variables of interest.

We conductedlatent class analyses on Wave 1 data to identify distinct life course typesthat manifest themselves in response patterns on a range of survey itemscapturing aspects of respondents’ educational, occupational, family and socialparticipation careers. Fit statistics indicated that a solution with fourclasses was optimal (Table 1).

Virtually allwomen in the first class (“career and family”) had married and a large majoritybecame mothers at a regular age. They are highly educated and are likely towork in the social services, education and healthcare fields. They are unlikelyto have primarily been homemakers. The second class (“family oriented”) alsohas a high probability of marriage and regular age motherhood, but educationalattainment is lower. Women in this class have a moderate probability of havingbeen a homemaker and high probability of having had a sales or administrativejob. Women in the third class (“career oriented”) are highly likely to haveremained childless and unmarried. They tend to be highly educated and areunlikely to have been homemakers or blue collar workers. The fourth class (“precarious”)has moderate probabilities of early motherhood and never having married, lowsocial participation, low educational attainment and a high likelihood of bluecollar jobs.

Multinomiallogistic regression analyses (results not shown in table) showed that Non-Hispanicblack women were less likely than non-Hispanic whites to have “career andfamily” or “family oriented” life course types, and more likely to have a“precarious” life course type. Poor financial circumstances and poor health whengrowing up were associated with a lower likelihood of having a “career andfamily” life course and a higher likelihood of a “precarious life course”.

We conductedlinear growth curve models to estimate whether the levels and trajectories ofwellbeing and self-rated health differed by life course type.  Race and earlylife characteristics were adjusted for.  Women in the “career and family” typeconsistently had the highest scores on both indicators, and the “precarious”group had the lowest scores. The self-rated health model shows some convergenceover time between the “career and family” and the “precarious” groups (Figure1).

These resultsare consistent with other studies indicating advantages for those combiningwork and family roles, but also suggest that these advantages may change overtime. For the ‘precarious’ group the fact that reported well-being appearedstable over the follow-up-in contrast to other groups-might suggest a positiveimpact of access to age related benefits and status, but this needs furtherinvestigation.

Figure 1.Predicted wellbeing and self-rated health by life course type