Health, employment, and economic change, 1973-2009: repeated cross sectional study
Health, employment, and economic change, 1973-2009: repeated cross sectional study by Jonathan William Minton and colleagues (BMJ 2012;344:e2316, doi:10.1136/bmj.e2316)
Objective—To see whether adverse relations between social class, health, and economic activity, observed between 1973 and 1993 and previously identified in a 1996 BMJ paper, were still apparent between 1994 and 2009 despite improvements in the general economic climate and overall population health.
Design—Replication of repeated cross sectional analysis from the original paper, using the same source (the General Household Survey) and occupation coding scheme, but extended from the period 1973-93 to 1973-2009, and including women as well as men.
Participants—Men and women aged 20-59 years in each annual survey between 1973 and 2009.
Main outcome measures—Change over time in class specific rates of employment, unemployment, and economic inactivity within subgroups of respondents.
Results—Overall employment rates have decreased for men of working age while increasing for working age women. For men in particular, the gradient of these changes seems to depend on occupational group. Over 37 years, the difference in occupational group specific economic inactivity and employment rates between people reporting and those not reporting a limiting long term illness has increased substantially.
Conclusion—Between 1973 and 2009, the relation between good health and securing and sustaining employment has strengthened for men and women. For men, this has been caused by employment rates decreasing and economic inactivity rates increasing among men with poor health. For women, this has largely been because of a general trend of increased employment and reduced economic inactivity occurring among healthier women, but not in women of poorer health. Some evidence suggests that, since 2005, the relation between health, employment, and economic inactivity for women in the top two occupational groups has become more like that for men, with poor health becoming associated with reduced employment rates.
Why do the study?
During the late 1980s increasing concern was reported among labour market and social policy analysts about the accuracy of the way in which unemployment statistics were calculated and presented. Some suspected that many people (men especially) who would previously have been classified as “unemployed and looking for a job” were being reclassified as “economically inactive” (not looking for a job).
Research in the south Wales and Sheffield areas, where mining and steel making had been decimated by the recession,  had shown that redundant workers did not all sign on as unemployed. Rather, many of them joined the ranks of the “economically inactive” because of long term sickness. Anecdotal reports came from people working in labour exchanges saying that they were being encouraged to advise job applicants to apply for disability benefits rather than continue to seek work. It was not hard to do this because chronic ill health was common among people who had worked in mining and steel making owing to the hazards of their jobs. Effectively, there had always been a lot of workers who continued to work despite experiencing musculoskeletal and respiratory conditions. When the employment for which their skills were relevant disappeared, they could qualify on purely medical grounds for disability benefit.
When unemployment began to fall in the late 1980s, it was accompanied by continuing increases in the proportion of the working age male population receiving long term disability benefits (fig 1) 1 , and as the next economic downturn receded in 1992-3, this figure rose. Suspicions increased that the real rate of unemployment was not reflected in the headline figures. But questions remained, such as were differences between occupational groups also being hidden? And had the two recessions of the 1980s and 1990s affected different occupational social classes in an unequal manner?
What did the authors do?
The cross sectional study by Minton and colleagues takes unemployment trends and asks: are the trends the same in different occupational groups (socioeconomic classes)? Has labour market change affected those in blue collar or manual occupations differently from those in white collar occupations? Has it affected those with long term health problems differently from those in better health? It has long been suspected that people with poorer health are more likely to experience unemployment.
The authors put together a series of data from the General Household Surveys (GHSs). Every year, around 12 000 to 15 000 people, a representative selection of the British population, have been asked about (among other things) their health, their social class based on occupation, and their employment status (this survey was discontinued in January 2012). The GHS is a cross sectional study—that is, a different sample of people are surveyed each year, with no follow-up of any individuals.
Men and women are allocated into social classes on the basis of their occupations, with the most advantaged and highest status group being professionals and managers in large companies (SC 1) and the least advantaged and lowest status being non-skilled manual workers (SC 4). In general, the jobs in SC 1 (such as medicine) require higher levels of education and offer more job security and autonomy. Exposure to industrial hazards and heavy work is greatest in SC 4. The intermediate group is made up largely of lower level managers, clerical workers, and sales workers, whose work conditions are less hazardous but who have lower status and less autonomy at work. Skilled manual workers often have more dangerous and heavy activities included in their jobs, but tend to have more autonomy.
Occupations such as printing, carpentry, and plumbing required a five year apprenticeship until the 1990s. In traditional industrial areas, semi-skilled occupations in coalface mining, steel making, and ship building also carried a high local status because of the hazards involved, and they were often well paid.
In the past, many studies of health inequality, such as the Black report, have used cross sectional datasets that recorded both social class and health, such as the GHS. Today, more use is made of the newer longitudinal studies that follow individuals over the course of their lives. But for the purposes of the authors, the GHS was sufficient.
The data series made it possible to see, in each year, what proportion of those in each social class were in good or poor health, and what proportion in each social class with good or poor health were employed, unemployed, or economically inactive.
What did the study find?
The authors found that the decrease in employment among men with less good health was far greater among those in manual occupations. As the general level of employment rose again in the mid-1990s, employment rates in this group of men did not respond at all, but continued to fall. And surprisingly, as the proportions of those in manual occupations with less good health who were employed decreased, the proportions that were unemployed did not increase. Rather, this group became steadily more likely to be excluded from the labour force as economically inactive.
What are the strengths and limitations of the study?
The study is an excellent example of what can be done with some of the long running series of British social and health data. The results show a neglected aspect of health inequality—that is, inequality in the impact of ill health on the ability to work. Because the GHSs (and their successor the General Lifestyle Surveys) continued to collect the same data in the same way from similarly representative samples of the population, we can be confident that the trends are real.
Could the rising trend in permanent sickness be caused by increasing severity of the long term conditions? This would be plausible only if one could argue that there was some kind of social gradient in the increase in severity. It is more likely that the work environment rather than the nature of the illness creates the gradient that we see. But as there are no data on the diagnoses, we do not know if these have changed over time.
Another limitation is that there is no way of telling whether people’s responses to “the same” illnesses have changed over time. Do people now react more negatively to less than perfect health, making them more likely to give up participating in work? And, if such a change has occurred, has the trend differed according to social class, as it would have had to, to result in the widening social gradient in the impact of ill health on employment that we see?
What does the study mean?
By 2009, there was an astonishing 60% economic inactivity rate in men whose usual job was semiskilled or unskilled manual and who had a long term illness (fig 2 2 ). In this group, the rate of economic inactivity hardly improved during the upturns in the economy of the 1990s and early 2000s. This contrasts with an inactivity rate of only 20% in men who answered the question on long term ill health in exactly the same way, but whose last occupation was professional or managerial.
The data for women are different and harder to interpret. Women’s participation in employment has risen sharply, and economic inactivity is more likely to take the form of home and family care than long term disability.
Why are so many more people permanently sick when, for the population as a whole, mortality is plummeting? At a time when employers have more applicants for the available jobs, they may become choosier. The rising trend of employment for women creates more competition for men in the job market. It is probably easier for women to fit into the service sector jobs that now make up a higher proportion of all jobs. Comorbidity of physical disorders (such as musculoskeletal problems) with psychological issues has probably become more important, particularly in deprived areas.
The demands of work have changed in such a way that some non-life threatening conditions such as depression, anxiety, and autism spectrum disorders, which have increased in prevalence, now make it far harder to find and keep work. Many modern jobs—for example, working in a call centre—put a lot of emphasis on communication skills. Another likely influence is that there are more people chasing each job (5.7 unemployed people per vacancy in January 2012), so that relatively minor health problems render the individual less desirable to employers. A musculoskeletal illness (the most common cause of long term work disability, followed by mental health problems) may not impair communication skills but may result in more sickness absence than average, and eventually result in dismissal.
The findings show the increasing importance of “health maintenance” strategies in the current labour market. Both physical and particularly mental fitness seem to have become far more important factors in the ability to remain in employment. Fortunately, policies are now being developed that focus on adapting the workplace to the abilities of individuals. Improved working environments and more supporting managers can do a lot to maintain fitness for work. Primary care also has a vital role in forestalling health problems before they limit people’s ability to work. It will be interesting to see what happens if recent government policies do succeed in removing entitlement for long term sickness benefits from more people. In the present economic climate, with few jobs available for those no longer entitled to sickness benefits, the most that can be hoped for is a sharp increase in the numbers of people needing unemployment benefits. But nobody seems to have thought of that.Mel Bartley, professor of medical sociology
1Department of Epidemiology and Public Health, University College London, London, UK
Correspondence to: firstname.lastname@example.org
Competing interests: MB’s work on this paper was supported by ESRC grant RES-596-28-0001 International Centre for Life Course Studies.
Provenance and peer review: Commissioned; not externally peer reviewed.
- Working Party on the Measurement of Unemployment in the UK. The measurement of unemployment in the UK (with discussion). J R Stat Soc [Ser A] 1995;158:363-418.
- Harris CC. Redundancy and recession in south Wales. Blackwells, 1987.
- Westergaard J, Walker A, Noble I. After redundancy: the experience of economic insecurity. Polity Press, 2002.
- Beatty C, Fothergill S. Labour market adjustment in areas of chronic industrial decline: the case of the UK coalfields. Reg Stud 1996;30:627-40.
- Black C. Employee support: good business sense [blog]. The Work Foundation, 2012. www.theworkfoundation.com/blog/760/Employee-Support-Good-Business-Sense
- Fenton WS, Stover ES. Mood disorders: cardiovascular and diabetes comorbidity. Curr Opin Psychiatry 2006;19:421-7.
- Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 2012;May 10. doi:10.1016/S0140-6736(12)60240-2 [published online.]
- Office for National Statistics. Labour market statistics. May 2012. www.ons.gov.uk/ons/publications/re-reference-tables.html?newquery=*&newoffset=50&pageSize=25&edition=tcm%3A77-230793.
Cite this as: Student BMJ 2012;20:e4095