History of income benefits and health at older ages

Transcription

1 History of income benefits and health at older ages Amélie Quesnel-Vallée Canada Research Chair on Health and Social Policies, McGill Miles Taylor Florida State University Renée Carter Australian Bureau of Health Information Johan Fritzell Aging Research Center, Karolinska Institutet

2 Unemployment and health Unemployment is associated with both morbidity and later mortality (Roelfs et al. meta analysis). But the extent to which this is a causal relationship, and if so, the mechanisms by which it operates have yet to be fully elucidated (see O Campo et al. for realist review on possible theoretical mechanisms). One of the main hypothesised pathways is material, through income loss. This of course raises the possibility the deleterious effects of unemployment can be mitigated by unemployment insurance. And indeed, studies have shown that unemployment insurance may have a positive effect on health (Cylus and Avendano in Health Affairs; Cylus, Glymour and Avendano AJPH).

3 Confounding of unemployment to health relationship The unemployed are likely a heterogenous group: those experiencing a job loss that is truly exogenous from their personal characteristics, and those for whom these personal characteristics are endogenous. Among those experiencing exogenous job loss, unemployment insurance should mitigate the negative effects on health of the financial loss of job income. But among those who are experiencing endogenous job loss, given that the association is spuriously determined by prior poor health, unemployment insurance will likely not exert a significant mitigating effect as the causal pathway lies outside the income loss to health relationship.

4 Income benefits as a proxy Few studies are able to measure health prior to unemployment Especially when using administrative income data But a history of different income benefits may provide an indirect handle on this problem

5 Study design Association between up to 20 years of historical income tax records from different sources : Family income trajectories Income from employment insurance, welfare, disability credit And well-being among Canadians 40 to 64 years old Self-rated health Disability Stratified by sex

6 Data Longitudinal and International Study of Adults, linked to T1FF Biennial longitudinal survey started in 2012 Linked with income tax data, pensions N=32,133 longitudinal respondents in 2012

7 Linked administrative data Historical income tax records T1 (T1FF) T Canada pensions Longitudinal Immigrant Database Landing File component Business Register Linkage

8 Sample 40 to 64 years old in 2012 Having received some employment income in the past Income measured annually (~ 90% linkage) Well-being measures in 2014

9 Control variables Mother and father s level of education Respondent born in Canada Respondent living alone Respondent s level of education Age in 2014

10 Analyses Logistic regression Poor health Disability status Weighted analyses Likelihood ratio tests support the addition of source of income variables over the income levels

11 % Family income trajectories Latent class analysis with Mplus Family income By sex: both fit best with four classes 0,6 0,5 0,4 0,3 0,2 0,1 0 Income trajectories Femmes Hommes

12 Ever received income benefits/subsidy Invalidité Assurance emploi, assistance sociale, invalidité Assistance sociale; Assistance sociale et invalidité * * Femmes Hommes Assurance emploi et assistance sociale Assurance emploi et invalidité * Assurance emploi * * 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 %

13 Income trajectories (ref. low income), ORs and 95% CIs Mauvaise santé Incapacité 1,2 1, ,8 0,8 0,6 0,6 0,4 0,4 0,2 0,2 0 Moyen-bas Moyen-haut Élevé 0 Moyen-bas Moyen-haut Élevé Note: estimates for models including all controls

14 Income trajectories (ref. low income), after adding ever received benefits Mauvaise santé 1,4 1,2 1 0,8 0,6 0,4 0,2 0 Moyen-bas Moyen-haut Élevé Incapacité 1,4 1,2 1 0,8 0,6 0,4 0,2 0 Moyen-bas Moyen-haut Élevé Note: estimates for models including all controls

15 Poor health by receipt of income benefits history, ORs and 95% CIs Assurance emploi Assurance emploi et assistance sociale Invalidité Assurance emploi et invalidité Assurance emploi, assistance sociale et invalidité Assistance sociale; Assistance sociale et invalidité

16 Disability by receipt of income benefits history, ORs and 95% CIs Assurance emploi Assurance emploi et assistance sociale Invalidité Assurance emploi et invalidité Assurance emploi, assistance sociale et invalidité Assistance sociale; Assistance sociale et invalidité

17 Discussion History of income benefits is associated with health in older adults net of historical levels of income Those with a history of only unemployment insurance are not different from those with only employment income All other combinations of benefits are associated with increased likelihood of poor health or disability Important heterogeneity exists in the association between unemployment and health in older adulthood Likely reflects strong selective processes that should be taken into consideration in future studies