Demographic, behavioural and socio-economic influences on the survival of retired people - Evidence from a ten year follow up study of the General Household Survey, 1994

Abstract

This article reports on a longitudinal analysis of a sample of residents who were aged 65 or over when interviewed for the 1994 General Household Survey. It investigates the associations between various personal characteristics as established at the interview and the probability of survival during a ten year follow-up period. Smoking was the most important factor of those measured in determining the likelihood of survival. Smokers at the time of the interview had a mortality risk, on average, 78 per cent greater than non-smokers. Non-drinkers and those who had less than one unit of alcohol per week at the time of the interview appeared to have a higher mortality risk than those who drank between one unit and the recommended government maximum per week. However, further analysis suggested that this result might be a product of the health status of individuals at the time of the interview. Type of housing tenure and region of residence were better predictors of mortality risk than occupation-based social class. Office for National Statistics Box one Variables included in this analysis Variables Definition Demographic variables Marital Status Married/cohabiting or Other (single/divorced/separated/widowed) Behavioural variables Smoking Current smoker or non-smoker Alcohol consumption Average weekly units 'negligible' less than one unit per week 'moderate' 1 to 21 units per week (men) 1 to 14 units (women) 'heavy' greater than 21 units (men) greater than 14 units (women) Socioeconomic variables Tenure Owns/rents Availability of car None/with car Registrar General's Social Class (RGSC) Own RGSC Occupation was originally classified to Socioeconomic Group (SEG): this was subsequently mapped to RGSC General environmental variables Region Region of residence Northern and Western including the North, Yorkshire and Humberside North West and West Midlands Southern and Eastern including the East Midlands, East Anglia, Greater London, Outer Metropolitan area, Outer South East and South West.

Socio-Geographic Variations in Mortality in a Large Retired UK Population

“…it is a kind of weak proxy measure of accumulated human capital), and hence lifetime earnings. Final salary may therefore discriminate better between risk groups in the retired population than measures of current income or pre-retirement occupational status (Breeze et al 1999;Nazroo et al 2008;Johnson and Langford 2009). It is perhaps for this reason that final salary is a better predictor of mortality risk for male pensioners than Carstairs.…”

“…Nor have such studies made much use of area-based measures of deprivation, presumably due to the sample sizes that are required for such analyses. Johnson and Langford (2009), in one of the few published studies which uses both socioeconomic and geographical variables in its analysis of influences specifically on mortality in the retired population, found that region appeared to have an effect on male mortality which was independent of socioeconomic position. Their sample, however, was small (approx.…”