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

Introduction

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.

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.

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.

British developments

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

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).

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.

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