Old Sick and Alone Living arrangements health

Old, Sick and Alone ? Living arrangements, health and wellbeing among older people RGS-IBG Annual International Conference London, 2006 Harriet Young and Emily Grundy London School of Hygiene and Tropical Medicine

Background Ø Demographic changes over the last century have led to older age structures throughout Europe. Ø Declines in the proportions of older people living with children and increases in proportions living alone.

Background What are the implications for health and well-being ? Research to date: l Those living with spouse are the most healthy l l l Contradictory evidence for those not living with spouse: more healthy living with other relatives or living alone ? • Mor et al (1989), Grundy (2001) Selection effects Effects vary according to cultural and socioeconomic context

Research Objectives Ø Describe the magnitude of differences in living arrangements of those aged 60+ between regions of Europe Ø Analyse associations between living arrangements, health and happiness among older people l l For all countries together Examine variations between regions

Dataset: European Social Survey (ESS) Ø Two cross-sectional rounds of data – 2002 and 2004 Ø Using data from 18 countries – l l Excluded 5 countries with response rates below 50% (Italy, France, Czech, Lux, Switz) 3 additional countries (Ireland, Iceland, Israel) Ø Sample size 17, 208 people aged 60+

Country groupings Nordic West Sweden Norway Finland Denmark Germany Portugal Belgium Greece UK Spain Austria Netherlands Poland Slovenia Slovakia Hungary Ukraine Estonia N=3621 N=5867 N=3863 Total population size=17, 208 South N=3857 East

Variables Ø Health and well-being: l l l Self-rated health Happiness scale (0 -10) Satisfaction with life (0 -10) Ø Other variables l l l Ø Living arrangements: MARRIED: l spouse only, l spouse and other, NOT MARRIED: l Children or others only, l Alone l l l Age Gender Region Marital status Socio-economic status Social contacts

Analysis Ø Carry out all analyses separately by gender Ø Descriptive analyses Ø Logistic regression with binary outcome objective): l l Ø Married: living with spouse and others versus spouse only; Non-married: living with non-spouse others versus living alone Logistic regression with ordinal outcome (2 nd objective) l l poor, fair, good health happiness scale (1 st

Results: Objective 1

Table 1: Adjusted odds ratios from logistic regression of region on living arrangements for women, ESS * Married: Spouse + others v spouse only OR Not married: With others v alone OR Region (reference: Nordic) West 2. 4 *** 2. 9 *** South 9. 8 *** 26. 2 *** 13. 9 *** 34. 2 *** East * Model also controls for age, income, feelings about income, educational qualifications, health, and whether currently widowed for the non-married. ** p<0. 01 *** p<0. 001

Results: Objective 2

Table 2: Adjusted odds ratios from ordinal logistic regression of living arrangements on health and happiness for women, ESS* Self rated health OR (lower OR= less healthy) Happiness score OR (lower OR= less happy) Living arrangements (ref: spouse only) Spouse & other 0. 9 1. 0 Alone 0. 9 0. 5 *** Other only 1. 2 0. 9 * Model also controls for age, region, indicators of socio-economic status, indicators of social contacts, and whether currently widowed ** p<0. 01 *** p<0. 001

Table 3: Odds ratios from ordinal logistic regression of living arrangements & other factors on health and happiness for non-married women, ESS* Self rated health OR (lower OR= less healthy) Happiness score OR (lower OR= less happy) Living arrangements (living with non-spouse other vv alone) All countries 1. 25 1. 7 *** * Model also controls for age, region, income, education, feelings about income, indicators of social contacts, whether currently widowed, & region for Europewide model ** p<0. 01 *** p<0. 001

Table 3: Odds ratios from ordinal logistic regression of living arrangements & other factors on health and happiness for non-married women, ESS* Self rated health OR (lower OR= less healthy) Happiness score OR (lower OR= less happy) Living arrangements (living with non-spouse other vv alone) All countries 1. 25 1. 7 *** Nordic 1. 16 1. 2 West 1. 23 2. 1 *** South 1. 16 2. 0 ** East 1. 31 1. 4 By region: * Model also controls for age, region, income, education, feelings about income, indicators of social contacts, whether currently widowed, & region for Europewide model ** p<0. 01 *** p<0. 001

Limitations Ø Varying response rates for different countries may introduce bias. Ø Exclusion of institutional population. Ø In the ESS there are no objective health measures. Only self-rated measures. Ø Cross-sectional analysis so unable to control for selection effects. Ø Small samples sizes in some analyses.

Conclusions Ø Region strongly associated with living arrangements after control for other factors. Ø Regional gradient for self-rated health & happiness: (worst) East–South–West–Nordic (best) Ø HEALTH: No significant associations between living arrangements and health, except in Nordic region: living alone less healthy. Ø HAPPINESS: Women least likely to be happy living alone, significantly so for South and West.

Acknowledgements Ø This research is funded by the ‘Understanding Population Trends and Processes’ (UPTAP) programme of the ESRC. Ø If you would like more information, do get in touch: Harriet. young@lshtm. ac. uk THANK YOU