Intended for healthcare professionals

Education And Debate

As the health divide widens in Sweden and Britain, what's happening to access to care?

BMJ 1997; 315 doi: (Published 18 October 1997) Cite this as: BMJ 1997;315:1006
  1. Margaret Whitehead, visiting fellowa (margaret{at},
  2. Maria Evandrou, visiting fellowb,
  3. Bengt Haglund, associate professorc,
  4. Finn Diderichsen, professora
  1. a Department of Public Health Sciences, Karolinska Institute, S-172 83 Sundbyberg, Sweden
  2. b King's Fund Policy Institute, London W1M OAN
  3. c Centre for Epidemiology, National Board of Health and Welfare, Stockholm, Sweden
  1. Correspondence to: Professor M Whitehead The Old School, Ash Magna, Whitchurch, Shropshire SY13 4DR
  • Accepted 5 August 1997


In both Sweden and Britain, social inequalities in health have been widening over the past 20 years. In Sweden, improvement in life expectancy weighted for health has been greater for men and women in non-manual than in manual occupations.1 Swedish industrial workers experienced increasing mortality in the 1970s and early 1980s, at a time when death rates among professional and managerial workers were decreasing sharply.2

Likewise in Britain, large and widening differentials in mortality have been reported between the most and the least deprived areas in Scotland and northern England.3 4 At a national level, the gap between the death rates of different social classes has also widened over the 1980s and up to 1993.5 6 Mortality is now three times higher in unskilled manual workers than among professionals.6 Large differentials in morbidity have also been found.7 8

The role of equitable health services

Lack of access to essential health care is likely to make only a minor contribution to the overall difference in mortality.9 Nevertheless, health services have an important role, not least in coping with and ameliorating the damage to health caused by inequalities in society. At times of growing inequalities in health, it is more important than ever to ensure equitable access to health care for those hardest hit.10 11

In this respect, both Sweden and Britain have prided themselves on their national health services, based on equal access for equal need. Swedish evidence from the 1970s and 1980s supported the claim of an equitable service. After health status was controlled for, no socioeconomic differences were found in the proportion who had visited a doctor12 13; whereas studies from the 1960s (before user fees were reduced in a major health policy reform) had found higher use of the health service among high income groups.14

In Britain, the argument has shifted over time, with early evidence indicating a “pro-rich” bias15 16 whereas later studies concluded that, in primary care at least, there was a “pro-poor” bias.17 18 19 However, most of these studies were based on data from the 1970s and 1980s, and it is not clear if the situation has changed since then.

Analysing the current situation

To assess what has been happening to access and uptake of health care in the two countries over this period of widening inequalities in health, we used the British general household survey and Swedish survey of living conditions to analyse prevalence of ill health and use of the health service among different socioeconomic groups over three periods between 1984 and 1994. Two calendar years of data were aggregated for each period.

Summary points

In 1993-4, inequalities in access to care appeared in Sweden for the first time since the 1960s

Britain had little socioeconomic inequality in consultations with general practitioners in the mid-1980s, but a “pro-poor” gradient developed in the 1990s; this was not apparent for outpatient visits

In both countries the health divide widened between socioeconomic groups between 1984–5 and 1990–1 and narrowed slightly by 1993-4

This narrowing of health differentials in Britain occurred because of a marked deterioration in morbidity for the professional group

The major shifts in labour market and health policies over the decade provide possible explanations for some of the trends in utilisation

Five fairly comparable socioeconomic groups were constructed for each country. They were based on Sweden's socioeconomic classification and Britain's socioeconomic grouping. In both countries, the socioeconomic groups included all people who were currently employed, as well as unemployed people and the “economically inactive,” provided they had previously been employed. This is important because of the increasing numbers moving out of the labour force.20 The main difference in the classifications is that the British scheme incorporates self employed people into the socioeconomic group with employees of the same occupation, whereas the Swedish classification does not.

Utilisation of the health service was used as a proxy for access to care. Multivariate logistic regression was used to calculate, for each socioeconomic group, the odds ratios for consulting a doctor, adjusted for demographic factors (age, sex, and marital status), a measure of supply (metropolitan and non-metropolitan areas), and the major confounder of health status (model 1). The health status adjustment combined survey responses on self reported general health, longstanding illness, and limiting longstanding illness.

In a second multivariate logistic regression (model 2), employment status (divided into employed, unemployed, and economically inactive) was added to the first model to test whether employment situation was an important mechanism through which socioeconomic status influenced use of health care.

To study the trends in morbidity over the same periods, prevalence rates standardised for age and sex were calculated for the various measures of health status for each socioeconomic group.

Swedish developments

Table 1 shows that in Sweden the proportion of each socioeconomic group visiting a doctor increased over the decade 1984-94. The increase was particularly strong for professionals in the 1990s—their consultations were approaching the level of manual groups by 1993-4.

Table 1

Trends in use of health service, Sweden, 1984-94

View this table:

Lower use in relation to health status among manual workers had developed by 1993-4. Adding employment status to the model did not change the estimates for socioeconomic groups, not even when long term unemployment became prevalent in Sweden in the 1990s.

Are these emerging inequalities in use accompanied by growing inequalities in morbidity? table 2 shows clear inequalities in the prevalence of fair or poor health, ranging from 16% in professionals to 29% in unskilled manual workers in 1984-5. These inequalities widened in 1990-1, then narrowed slightly in 1993–4 with an increased prevalence of fair or poor health among professional and intermediate groups.

Table 2

Respondents reporting fair or poor health, Sweden, 1984-94. Values are percentages (95% confidence intervals), standardised for age and sex

View this table:

British developments

A corresponding analysis for Britain, combining general practitioner and outpatient visits, showed little inequality in access to care (table 3).

Table 3

Trends in use of health service, GreatBritain, 1984-94

View this table:

In Britain, general practitioner consultations can be separated from outpatient visits. This resulted in distinct gradients in use (table 4). Little inequality in NHS consultations with general practitioners was apparent in the mid-1980s, but a gradient favouring manual groups had developed by the beginning of the 1990s and was maintained in 1993-4. In contrast, for outpatient visits there was a (non-significant) gradient favouring professionals in the 1980s (table 4).

Table 4

Odds ratios* (95% confidence intervals) for consultations with general practitioners and outpatient visits, Great Britain, 1984-94

View this table:

Trends in morbidity in Britain show a clear socioeconomic gradient in health within each period (table 5). By 1993–4 the proportion reporting fair or poor health ranged from 27% of professionals to 47% of the unskilled manual group. Over the decade, the British differentials in morbidity widened, then narrowed slightly in the latest period as a result of a strong increase in morbidity in the professional group during the 1990s. This was also found for longstanding illness and limiting longstanding illness.

Table 5

Respondents reporting fair or poor health, Great Britain, 1984-94. Values are percentages (95% confidence intervals), standardised for age and sex

View this table:

Comparisons between countries

The increase in reported morbidity in non-manual groups in 1993–4 was apparent in both countries, though more marked among professionals in Britain, and was consistent for all the health indicators examined. In both countries the health gap between socioeconomic groups widened from 1984–5 to 1990-1, and then narrowed slightly up to 1993-4.

Where the countries differ strikingly is in the trends in access to care. For Sweden, inequalities in use of the health service appeared in 1993-4, favouring the professional group. This was due partly to a substantial increase in utilisation among professionals and partly to increasing inequalities in reported health. In Britain, differentials in utilisation between socioeconomic groups in all three periods were negligible when the measure of access was reasonably comparable with that available in Sweden (general practitioner and outpatient visits combined).

When we disaggregated utilisation, further contrasts became apparent. In Britain, trends in general practitioner consultations showed, as in Sweden, the development of inequalities in use during the 1990s, but unlike in Sweden, these favoured manual groups. There was a slight, non-significant, bias in favour of professionals in outpatient visits in the mid-1980s, but this bias had disappeared by the 1990s: the mirror image of developments in Sweden.

Possible explanations

Market reforms

Traditionally, the Swedish healthcare system has relied heavily on hospital services. In the past decade, market related reforms have included attempts to expand the number of private and public physicians in primary care.21 22 Many of these reforms were introduced in response to discontent among middle class sections of the population, and the way the changes were designed may have facilitated access of professional groups to these services, over and above other groups.

Spending cuts

Tight cost control measures have been in operation in the Swedish health service since the early 1980s, and these became much more stringent in response to the worsening economic situation from 1991 onwards. Spending per capita is now only about 10% higher than in Britain.23 This has the dual effect of reducing the level or quality of services available to the general population and producing high unemployment among health service workers, both of which may hit lower paid workers the hardest. British spending on health care has been low by international standards and has risen only modestly,24 but cuts have not been as rapid and severe as in Sweden.

Increased user charges

Lower paid workers would also be expected to be hardest hit by the sharp increases in Swedish user charges. The flat rate charge for every general practitioner or outpatient visit, as well as charges per day in hospital and for drugs, have increased gradually since the early 1970s and rose sharply in 1996-7. The combination of increased charges and higher unemployment may deter lower income groups from using the service. The British NHS does not currently have user charges for general practitioner or outpatient services.

Unemployment and privatisation of occupational health services

In the 1990s occupational health services have experienced cuts and privatisation. In Sweden occupational health services provide primary health care, particularly for manual workers, the groups showing a relative decline in use of the health service. From 1992, unemployment increased greatly in Sweden, especially among manual workers; the reduced use of the health service by people not in employment may indicate growing barriers to access to occupationally based services. However, we found no evidence that employment status accounted for the changing pattern of health service use in manual groups.

“Pro-poor” bias in Britain

The opposite trend, a bias favouring lower socioeconomic groups in access to general practitioner services developing in Britain over the decade, could be seen as part of a longer term transition. Studies in the 1970s found a bias in favour of richer groups in the NHS and were influential in generating debate about inequalities in service provision.15 25 Since then the NHS, at least at the general practice level, may have been putting more effort into providing services for less advantaged socioeconomic groups. Other studies using the general household survey reported a bias in favour of poorer groups in the NHS in the late 1980s and early 1990s.17 18 26 Our analysis is the first to span several points in time and to pick up a transitional phase.

The conclusions from this study only apply to access to specific primary care and outpatient services; the measures used would not pick up possible differentials in the quality of services. British women and elderly people, and people experiencing social disadvantage, have been found to have poorer access to specialist diagnostic and inpatient services.27 28

In times of growing hardship, accompanied by increasing inequalities in health, it is important to question how well the health sector is responding. This analysis provides an early indication of possible problems with access to care in Sweden developing in the 1990s. It also raises wider concerns about British trends in the health of different socioeconomic groups. Both are crucial areas for further investigation.


We are indebted to the Economic and Social Research Coundil's data archive at Essex University and alto to the Office for National Statistics for making the data files of the general household survey available for analysis.

Funding: FD was funded by the National Institute of Public Health in Sweden.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.