Article Text

Download PDFPDF
P43 The Healthy Worker Effect? Socio-economic and health status determinants of changing self-rated health in older people in the Lifeways Cross-Generation Study of a Thousand Families
  1. R Somerville,
  2. H Khalil,
  3. R Segurado,
  4. J Mehegan,
  5. M Heinen,
  6. C Murrin,
  7. CC Kelleher
  1. School of Public Health Physiotherapy and Sports Science, University College Dublin, Dublin, Ireland


Background There has been renewed interest in the self-rated health (SRH) of older people. Many cohorts, including the Lifeways Cross Generation Cohort Study, have confirmed the role of SRH in predicting mortality. Many have also examined cross-sectional determinants, but few have described the determinants of change in SRH using the same measures 10 years apart, with a particular interest in assessing the relative contributions of change in physical and psychological morbidity and socio-demographic changes.

Methods Using generalised estimating equations (GEE), we examine predictors of SRH in Lifeways grandparents who participated in the baseline (n = 710) and year 10 follow-up (n = 843) questionnaire surveys. Morbidity score was determined from summing 6 physician-diagnosed conditions that were self-reported at both time-points. For the cross sectional measures, a binomial distribution with logit link was specified, examining predictors of excellent/very good/good SRH versus fair or poor. For the predictors of change in SRH (range −2 to +2) a linear distribution with identity link was specified and two multivariable GEE models were constructed, firstly with only baseline variables, and secondly with variables reflecting change in morbidity or socioeconomic variables measured at both time-points. Each term in the model was examined conditional on the baseline value of SRH.

Results Cross sectional results at both baseline and year 10 confirm that measures of both physical and psychological morbidity, as well as a measure of socio-economic deprivation (eligibility for means-tested General Medical Services – GMS), are predictive of SRH in this cohort. Of the 288 respondents with SRH measured at both time-points the following were significant predictors in the baseline GEE model (B, p): age (−0.018, 0.017), Eastern (urban) region (−0.223, 0.011) and baseline SRH (0.411, <0.001). In the final GEE model, the following were significant (B, p): age (−0.018, 0.011), Eastern region (−0.196, 0.023), baseline SRH (0.337, <0.001), change in morbidity score (0.203, <0.001), developing an activity limiting health condition (−0.701, <0.001), becoming GMS eligible (−0.239, 0.012) and change in employment status from working to other (0.234, 0.011).

Conclusion The results of this on-going analysis show some interesting and novel findings. Change in SRH is related to initial level of SRH, but not initial level of morbidity. Even after accounting for change in objective and subjective morbidity, certain socio-economic factors (region and GMS eligibility) are influencing change in SRH. Of particular interest and meriting further investigation is the finding that leaving employment is associated with a positive change in SRH.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.