Socioeconomic variation in the impact of obesity on health-related quality of life
Introduction
Rising obesity levels are a major problem in many countries. In England in 1993 15% of adults aged 16 years and over were obese (13% of men, 16% of women); by 2008 this figure had risen to 25% (24% of men, 25% of women) (National Centre for Social Research & Department of Epidemiology and Public Health (UCL), 2009). This trend is worrying because obesity is an important risk factor for a number of diseases including coronary heart disease, type II diabetes, hypertension and stroke (NHLBI, 1998).
In England 7% of all deaths are attributable to obesity (House of Commons Health Committee, 2004). Obesity also decreases life expectancy. For instance, in the UK a 30-year old non-smoking man with a BMI of 35 kg/m2 is projected to lose five years of life compared to a similar person with a BMI of 24 kg/m2 (Mayhew, Richardson, & Rickayzen, 2009). The analogous result for women is a loss of two years. Results of a similar order of magnitude have been found in other countries (see, e.g., Peeters et al., 2003).
As well as affecting premature mortality and life expectancy there is also increasing evidence that obesity is associated with a loss in health-related quality of life (HRQL) (Barofsky et al., 1997, Brown et al., 1998, Doll et al., 2000, Fine et al., 1999, Ford et al., 2001, Groessl et al., 2004, Han et al., 1998, Heo et al., 2003, Hill and Williams, 1998, Jia and Lubetkin, 2005, Kortt and Clarke, 2005, Laaksonen et al., 2005, Larsson et al., 2002, Lean et al., 1998, Lean et al., 1999, Le Pen et al., 1998, Macran, 2004, Sach et al., 2007, Wee et al., 2008, Yan et al., 2004). This evidence shows that the negative effects of obesity on HRQL persist, but at a lower level, even after controlling for a range of confounding variables.
While the negative effects of obesity on HRQL are established, little attention has been paid to variations in these effects between population groups. Such considerations are important given recent interest in “Pathways to health” (Birch, Jerrett, & Eyles, 2000), in which health is determined by a range of factors which interact with one another and which are not easily separable. According to this approach – which is based in Grossman’s (1972) human capital model where health is an output that is ‘produced’ by individuals from a variety of inputs into the production process – medical care, lifestyle, SES and other social determinants of health are inputs into the production of health that interact with one another in complex ways. SES and other social determinants of health affect health directly (see, e.g., Marmot, 2010). They also modify the association between medical care and lifestyles – other inputs into the production of health – and health (Birch et al., 2000). One reason for this ‘modifying’ role is that the underlying cause of unhealthy lifestyles may affect the impact that lifestyles have on health, and these may vary by SES (Birch et al., 2000). For example, the underlying causes of obesity among those earning low incomes may be due to the consumption of cheaper less nutritional food, whereas among those earning high incomes it may be due to limited non-work leisure time making it difficult to undertake time-intensive physical activity (Butland et al., 2007). Another reason is that the strength of the association between unhealthy lifestyles and health varies by SES (Birch et al., 2000). For example, the production relationship between obesity as an input and health as an output may vary between population groups. Hauck, Shaw, and Smith (2002) suggest that variations in health arise inter alia from systematic variations in health production functions between population (e.g., socioeconomic) groups, implying that individuals with more favourable social determinants of health are likely to be more efficient in producing health.
There is some evidence of a modifying role of social determinants of health in the context of the association between lifestyles and health. For example, Davey Smith and Shipley (1991) show that the association between smoking status and 10-year mortality risk depends on occupational grade and car ownership. Birch et al. (2000) show that the association between smoking status and self-assessed health status depends on household income, employment status and education. In the case of obesity the evidence is limited. Laaksonen, Sarlio-Lahteenkorva, Leino-Arjas, Martikainen, and Lahelma (2005) show that the association between BMI and health status, measured using the physical and mental health component summaries of the SF-36, is modified by occupational class and working conditions. The association between BMI and health status did not significantly change when SES and working conditions were controlled for; there was some evidence that the association between BMI and physical health depended on working conditions.
The aims of this study are to investigate the relationship between HRQL and obesity, and to investigate whether or not this relationship varies by SES. We undertake the analysis using data from a large individual level health survey, which includes interviewer-measured rather than self-reported height and weight, plus a comprehensive set of individual and household characteristics that allows us to control for confounding factors that affect the relationship between obesity and HRQL.
Section snippets
Data and variables
The analysis is based on data from four rounds (2003–2006) of the Health Survey for England (HSE) (National Centre for Social Research & Department of Epidemiology and Public Health (UCL)). The HSE is a cross-sectional representative national survey which draws a different sample every year of individuals living in England. Respondents are interviewed on a range of topics including their health (including obesity and EQ-5D score), and their socioeconomic status. Only participants aged 16 or
Results
The total number of respondents in the HSE in 2003–2006 was 61,603. Forty five thousand nine hundred and eighty five were aged 16 or above and 42,002 had EQ-5D data. Thirty three thousand seven hundred and sixteen observations were included in the income regression and predicted SES values were computed for 42,825 observations. A total of 33,105 observations (15,142 men, 17,963 women) had EQ-5D scores and predicted income data and were used in the EQ-5D models. The numbers of observations in
Discussion
The aims of this study were to investigate the relationship between HRQL and obesity, and whether or not this relationship varies by SES. Our main findings are that obesity and HRQL are negatively correlated, and that the relationship between HRQL and obesity varies significantly by SES.
We provide evidence to show that overweight and obese men and women have significantly lower HRQL than those of normal weight. After controlling for individual and household characteristics, we obtain similar
Acknowledgments
An earlier version of this paper was presented at the Health Economists’ Study Group meeting in Sheffield, England, in July 2009 and at the Nordic Health Economists’ Study Group meeting in Reykjavik, Iceland, August 2009. We thank participants at both meetings for their helpful comments. This work was undertaken at UCL who received a proportion of funding from the Department of Health’s NIHR Biomedical Research Centres funding scheme.
References (43)
- et al.
Heterogeneity in the determinants of health and illness: the example of socioeconomic status and smoking
Social Science & Medicine
(2000) - et al.
Beyond BMI: the value of more accurate measures of fatness and obesity in social science research
Journal of Health Economics
(2008) - et al.
Body mass index and quality of well-being in a community of older adults
American Journal of Preventive Medicine
(2004) - et al.
Impairment of health and quality of life in people with large waist circumference
Lancet
(1998) The impact of obesity on employment
Labour Economics
(2007)- et al.
Pain in the obese: impact on health-related quality-of-life
Annals of Behavioral Medicine
(1997) - et al.
What is a healthy weight for middle aged women?
International Journal of Obesity and Related Metabolic Disorders
(1998) - et al.
Foresight: Tackling obesities: Future choices – Project report
(2007) - et al.
Confounding of occupation and smoking: Its magnitude and consequences
Social Science & Medicine
(1991) Modelling valuation for EuroQol health states
Medical Care
(1997)