People with a lower education level used GP services slightly less often as those with a higher level of education in most countries (except for Belgium and Germany). At the same time, higher educated people used specialist care services significantly more often in all countries (except for the Netherlands). Educational inequalities in utilisation of specialist care among women were slightly larger than among men in some countries, although the general pattern of use was similar for both men and women. Inequalities in utilisation of specialist care were equally large in Eastern European and in Western European countries, except for Latvia where the level of inequalities was somewhat larger. Similarly large was the level of inequalities in utilisation of specialist care among patients with chronic diseases, diabetes, and hypertension.
The high percentage of non-response in some countries could have biased our results if both the educational level and the reported utilisation of services had been unequally distributed among respondents and non-respondents. Although some studies reported that non-response is related to socioeconomic status[13–15], previous evaluations showed that the association between utilisation of services and socioeconomic status would not greatly change if non-respondents were included with respondents[16, 17]. Nevertheless, in the present study we cannot exclude the possibility that an over-representation of sicker lower educated people in the non-response group may have led to some underestimation of the pro-rich inequalities in prevalence rates of utilisation reported here.
We used education as an indicator of socioeconomic position. Education allows the classification of individuals who do not work, prevents reverse causation, and facilitates international comparisons due to its relative ease of measurement. In addition, recent studies suggest that in some countries education has an independent effect and is more strongly related to the likelihood of health services utilisation than income and employment status[18, 19]. On the other hand, educational level might not accurately indicate an older person's current socioeconomic position since it is acquired early in life and may inadequately reflect changes in socioeconomic position during adult life.
There were large differences between countries in the educational distribution. These differences reflect, in part, the real situation of educational attainment in different countries of Europe. However, there is a possibility that the ISCED classification is not flexible enough to accommodate different national schemes. To cope with the differences in educational classification we used the RII, a measure that takes educational distribution into account[11, 12]. Additionally, RII has the advantage that it can be applied in a comparable way to all countries provided that the educational classifications are strictly hierarchical.
The recall period for use of GP and specialist services was shorter in the Netherlands and Belgium than in the other countries. A longer recall period would have influenced the overall utilisation rates for the total population. It is, however, unlikely that it would have a differential effect on utilisation of services by different educational groups.
Self-assessed health was used in order to control for the health care needs of the population. Although the measure of self-assessed health is often used in health care research due to its wide availability and good comparability, it does not completely encompass the full spectrum of need. A better control for need would likely result in greater inequalities in specialist visits, while inequalities in GP visits might have also emerged in some countries.
Most European countries have achieved universal access to health care. However, the results of the present study show that universal access does not mean equal use. One might argue that differences in utilisation do not directly reflect inequalities in access to care. The decision to use health care services and the type of provider is, after all, a personal choice. Nevertheless, this personal choice is affected to a large extent by various enabling and predisposing factors. People from lower socioeconomic strata are likely to have fewer enabling factors and more barriers to use specialist care.
European countries have very different health care systems. For example, some countries operate with GP gate keeping (e.g. the UK, the Netherlands), others have more direct access to specialists and hospital care (France); some countries use only public insurance (Germany, the Netherlands), others only private or a combination of the two (Spain, Portugal); some countries use co-payments, others do not; etc. Regardless of the way the system is organised, we find a generalised pattern of differential access to primary and secondary care for people with different socioeconomic positions. Such a universal pattern indicates that patients with a lower socioeconomic position encounter barriers that are common in all countries, and thus lie beyond the national structure and organisation of the health care system.
Proper communication between the patient and health provider where the patient not only receives information about his disease, diagnostic procedures, and treatment, but also feels understood and helped is essential. Successful communication contributes to both patient outcomes[22, 23] and general satisfaction with services[24, 25]. People with a lower socioeconomic position may better appreciate communication with the GP than with a specialist, as the former may be clearer in discussing the disease, be better at understanding and addressing the needs of the patient and, thus, be perceived as more trustworthy. On the other hand, patients with a higher socioeconomic position may trust a "higher specialised" provider and request contact with the specialist, or seek this contact directly thus avoiding the primary care provider. It is suggested that patients with lower education, lower income and ethnic background express more preference to see a GP for their initial care than better educated, higher income white patients, although research in this area is very limited and sometimes contradictory. Higher educated patients that chose a GP for their initial contact (either as personal choice or due to organisational enforcement, as in countries with a gate keeping system) are usually better able to articulate their needs for the specialist and have greater assertiveness regarding being referred to one[28, 29], leading to a higher number of referrals.
One may suggest that a simple substitution of care occurs i.e. equal quality care for the same problem, which is performed by one type of provider instead of another without any consequences for the health outcomes of the patient. Our data, however, indicate that lower-educated people use GP services slightly less often compared to higher-educated people in most European countries, while inequalities in the use of specialists are large. A better control for need of care may even reveal pro-rich inequalities in the use of GP services. Thus, we do not find evidence for the substitution of care. Others also showed that the likelihood to consult a specialist increases given a consultation with the general practitioner.
Another common feature of the health care system is its enormous complexity: whichever type of organisation exists in a country it is never easily understood, particularly by those with a lower socioeconomic position. This complexity is often coupled with constant changes in the way the system operates that may disorient even well-educated patients. Since primary care (GP practices) is the easiest, most accessible and least changeable type of care, people with a lower socioeconomic position may not feel inclined to go further up the hierarchy of the health care organization, in order to avoid this confusing complexity.
Within the generalised pattern of differential utilisation of different types of services, there remain some variations that indicate that national health care systems may play an additional role in (dis-)motivating patients to use particular types of care. For example, compared to other countries, we observed larger inequalities in the use of specialist care in Latvia and smaller inequalities in the Netherlands. Similar differences were also observed in studies on income inequalities in utilisation of care. It is plausible that these variations in the magnitude of inequalities are driven by differences in health system characteristics, such as sources of finance and service delivery practices. For example, in the Netherlands there is a stronger GP gate keeping system compared to other countries included in this study. A strong GP gate keeping system may allow a better control of the patient flow to specialists that is in accordance with clinical guidelines (and needs of the patients), thus leaving less room for inequalities in the utilisation of specialist care to occur compared to a more free-way system.
We hypothesized that inequalities in access to care in East European countries would be larger than in the West European countries due to disruption of the social protection and health care systems that occurred during the 1990s in many former Soviet countries. Our data do indicate larger inequalities in use of specialist care in Latvia. Compared to the neighbouring countries, Latvia has implemented a system with larger co-payment mechanisms for public health services. Thus, the financial barriers met by the population for the use of health services might have resulted in much lower utilisation rates and the highest level of inequalities observed in the present study. Also in Hungary, in addition to large inequalities in utilisation of specialist care, there were significant pro-rich inequalities in the use of GP services, indicating gross general inequalities in utilisation of health services. Our findings are supported by studies reporting larger inequalities in mortality amenable to medical care found in East European countries compared to West European countries [32–34]. However, in Estonia the magnitude of inequalities in utilisation of care was similar to that of West European countries, which indicates that the problem is limited to particular countries and can not be generalised to all East European countries.
The present study paid particular attention to people with chronic diseases. The results show large inequalities in utilisation of specialist services in this vulnerable group. Hampered access to specialist care might have a more severe impact on the health status of patients with high need, such as the chronically diseased, compared to the general population. Thus, there is an urgent need to investigate and remove barriers to the use of specialist care among patients with chronic diseases.