Previous studies in Russia have shown a wide gap in mortality between socioeconomic groups, and the differences are greatest among young and middle age people with the largest relative difference in the age group 20-39 years [2]. However, mortality rate does not reflect the burden of disease attributable to chronic disabling disorders that are not life-threatening. Thus, more refined measures are needed in order to compare health differences between different social groups.

Self-rated health has been shown to be a valid and reliable measure of overall health status in other countries [3,4]. Previous studies of self-rated health in post-Soviet Russia have revealed that better self-rated health consistently correlates with higher household income and, in recent studies, with higher education as well. Some of the previous studies have been conducted in urban parts of Russia (Moscow [5], Taganrog [6,7]), while others cover the whole country [8,9].

In general, the variation in self-rated health by socioeconomic status can be explained by differing health behaviour, environment, genetic predisposition, and health service utilisation [10]. Previous studies have not, however, analysed differences in self-rated health together with utilisation of health services in Russia.

The current Russian health service system is based on mandatory health insurance that should guarantee universal access to public sector health services without user charges at the point of delivery [11-14]. In practice, patients have to contribute both semi-formal user charges and informal payments. The public sector primary health care in urban areas, such as St. Petersburg, is based on a network of neighbourhood clinics (polyclinics) that are responsible for the population in their catchment area. In addition, a network of specialist polyclinics, such as women's clinics, provides specialist services without the need for prior referral from the neighbourhood clinic. More specialised services are provided by hospital in- and out-patient units. Furthermore, a number of private clinics also provide health services in St. Petersburg today, and gynaecology is one of the specialties where private services have developed most rapidly [15].

Some clinics (public and private) have financial support from voluntary health insurance companies. A small number of employers of the richer companies tend to purchase voluntary health insurance for their staff to cover extra services provided by the clinics that have an agreement with private health insurance companies [16]. Not many people in Russia have voluntary health insurance (VHI) policies. Only 3% of those surveyed are holders of VHI policies [17].

There is very little data on how well the mandatory health insurance system has been able to maintain equal access to health services. A study comparing eight countries that were previously part of the Soviet Union revealed that in Russia, people with lower education refrain from seeking health services more often than those with higher education. Use of health services was markedly lower among those with fewer household assets, suggesting that equal access to health services has been compromised [18].

The survey was conducted in 2003-2004 in the catchment area of three women's clinics in two districts of the city, namely Krasnogvardeysky and Primorsky. The Krasnogvardejsky district consists of different areas with diversity in socioeconomic status. District residents are mainly employed in industry; the place is not very attractive because of higher environmental pollution and less convenient public transportation.

The industrial development of the Primorsky district was less intense during the Soviet period and this district is more attractive for those with higher income because of newer buildings.

The potential respondents were randomly selected. For the survey we did not have any formal sample size calculation. Based on feasibility 2.8% from a total of 90 532 reproductive age women were invited to participate. The original target sample size was 2501 women. The sample was drawn from a database maintained by the District Authority Police department.

The survey was approved by the Ethical Committee of the St Petersburg Medical Academy for Postgraduate studies. For a detailed description see earlier reports [19,20]. The survey questionnaire was anonymous and confidentiality was guaranteed [19,20].

A total of 782 women (31%) were excluded from the sample for the following reason: women were not reached due to a difference in their official address and their real living address or due to not reaching the person because of living abroad or elsewhere in Russia. The final sample consists of 1719 women, and the response rate was 66.7% (n = 1147). The recruitment procedure started with an invitation letter that informed about the purpose of the study and invited women to come to clinic. The invitation letter was followed by a phone call. During the phone call the location of the participant was verified, the willingness of the potential respondent to participate was confirmed and the participation method was agreed (clinic visit or completing the questionnaire at home).

The questionnaires were self-administered, consisting of multiple-choice questions. They contained several sections including background information, pregnancies and children, use of health services related to pregnancies and deliveries, values of childbearing, health and health behaviour, and use of health services [19,20].

The analysis revealed a significant association between socioeconomic status and health indicators (Table S2, see additional file 2). University education has a negative association with poor self-rated health, 35.7% of women with the highest education reported poor self-rated health in comparison to 45.5% of those with the lowest education. Women with a university degree were also less likely to assess their quality of life as poor: 10.6% vs. 5.6% among women with the lowest and the highest educational level, respectively. After adjusting for age the correlation between education and self-rated health was significant: women with low education were more likely to report that they had poor health status in comparison to those with a university degree (OR = 1.48, p = 0.011), in addition they were more likely to report that they have poor quality of life (OR = 2.02, p = 0.013). However, no association between level of education and chronic disease was found.

A borderline positive association between low personal income and poor self-rated health as well as chronic disease prevalence was revealed with logistic regression analysis (Table S2, see additional file 2). However, there was a strong positive relationship between personal income and self-assessed quality of life. Women with a low personal income were almost four times more likely to report a poor quality of life than those with high income (age-adjusted OR = 3.92, p = 0.005).

Family income was the strongest determinant of self-rated health. Half of women with low family income had rated their health as poor, in comparison with only a quarter of those with high family income (age-adjusted OR = 2.31, p = 0.003). Almost as strong an association was found in the prevalence of chronic conditions: women with low family income were more likely to report chronic conditions (age-adjusted OR = 2.00, p = 0.013). A strong relationship was also found between self-perceived quality of life and family income. As many as 14.1% of women with a low family income reported a poor quality of life, in comparison with only 1.2% of those with a high family income, (age-adjusted OR = 13.23, p = 0.012).

As for utilisation of health services, 58.8% of women had visited a physician during the past year independently of SES. The type of service provider varied significantly by socioeconomic status (Table S3, see additional file 3). Differences between educational groups were not discovered in the utilisation of policlinic, but university education was correlated with the use of additional outpatient services. 16.6% of respondents with a university degree had used private and 6.6% occupational health services, while, for comparison, only 5.6% and 2.7% of women with the lowest education had used private and occupational health services, respectively. After adjusting for age the correlation between education and outpatient service use was significant: women with a university degree were more likely to report private health care use (OR = 3.37, p < 0.001), and more likely to report occupational health care use (OR = 2.60, p = 0.016) in comparison to the respondents with low education. On the other hand, women with a university degree were less likely to have been hospitalised, with borderline significance (OR = 0.69, p = 0.078).

High personal income was strongly associated with the use of private services (age-adjusted OR = 3.56, p < 0.001), and middle income with occupational services (age-adjusted OR = 1.83, although with a borderline significance, p = 0.077); 23.4% of women with high income had visited a physician at a private clinic and 6.2% of women with middle income had used occupational health services. The association between personal income and hospitalisation was not found to be significant.

Family income had a specific influence on the utilisation of health services. Women with middle and high family income were significantly more likely to use private services (age-adjusted OR = 1.81, p = 0.048 and 2.81, p = 0.004; respectively) in comparison to the poorest respondents. Those who did not know their family income were also more likely to use private services, but this association was not statistically significant. However, those who did not know their family income were significantly less active in seeking public health services.

Table S4 (see additional file 4) describes health service utilisation during pregnancy. As in the case of overall utilisation, women with a university degree were more likely to visit private clinics (age-adjusted OR = 5.90, p = 0.007) and other centres (age-adjusted OR = 2.41, p = 0.051). In St. Petersburg, "other centres" most likely refers to highly specialised centres that are connected to teaching hospitals, medical academies and scientific research institutes.

Income was not significantly associated with health service utilisation during pregnancy.

Participation in preventive health examinations also varied significantly across socioeconomic groups (Table S5, see additional file 5). University education had a significant positive correlation with Pap smear and gynaecological examination (age-adjusted OR = 2.06, p < 0.001 and 1.62, p = 0.002; respectively) and high personal income had a significant positive correlation with mammography (age-adjusted OR = 3.32, p = 0.003), Pap smear (age-adjusted OR = 1.72, p = 0.004), and ultrasound examination (age-adjusted OR = 2.09, p = 0.005), while high family income correlated only with mammography for women over 40 years of age (age-adjusted OR = 4.34, p = 0.012). Women who did not know their family income were less likely to have undergone gynaecological examination.

Some study limitations should be taken into consideration. First of all the cross-sectional design does not allow for any conclusions about causality in regard to health. It is possible that health-related social mobility may contribute to a positive association between health and higher income and education. However, that kind of reverse effect is not likely for socioeconomic differences in health care utilisation; for example, primary prevention measures such as cancer screening are unlikely to affect a person's socioeconomic status.

Reporting bias is another potential problem with a cross-sectional study. Women with low income may be more likely to report poor health. In this case the association between good self-rated health and high income may be overestimated. However, according to studies from other countries, self-rated health is a reliable measure of health status [3,4].

It is also possible that women with higher education are more likely to recall and report the exact names of tests they had undergone, e.g. Pap smear. However, given that women with lower education had also reported a lower frequency of gynaecology examination one can assume that the correlation between high education and Pap smear is not biased.

Selection bias may contribute to some data misinterpretation. If women with higher income did not participate in the study, the influence of society stratification may be underestimated. This study limitation is difficult to overcome given the lack of registers on education and income in Russia and the widespread underreporting of income.

The generalisability of the results requires a note, as well. St. Petersburg is a rather privileged and wealthy part of Russia and does not provide a typical example of the whole country. The population has a higher than average education, for example. In our study, the prevalence of a university degree is 34% and corresponds with official statistics [22]. Furthermore, the income differences are likely to be wider within St. Petersburg than within rural and semi-urban areas of Russia. However, it is likely that similar tendencies of inequality are to be found in other regions of Russia as well.

Our findings on the correlation between good self-rated health and higher education and income are consistent with the results of previous studies conducted in Russia [5-9]. However, in our study women reported poor health less often. In part this may be explained by the fact that our study population is younger.

A lack of association between self-reported chronic illness and education may have resulted from information bias; those with lower education may have been less aware of their illnesses. This view is supported by earlier findings that women with higher education are more likely to seek medical help at an earlier stage of the disease and use a variety of opportunities to control their health status more actively [23]. Therefore, they are more likely to discover the disease and get medical care at an earlier stage to preserve their health. On the whole, our results support the earlier view that those with higher education and income feel healthier and are more satisfied with their quality of life.

The results on utilisation of public health care services reveal hardly any differences between different socioeconomic groups. However, those with a university education visit a variety of service providers in addition to public sector services and those with a higher income visit the private sector more often. During the first two decades of post-Soviet Russia, both upward and downward social mobility were rapid and common. High income and high education do not necessarily go hand in hand in Russia, as they commonly do in Western Europe and the United States. Many highly educated occupational groups have been impoverished, while new business opportunities have paved the way to prosperity for those whose education may be lower.

Earlier studies from St. Petersburg have revealed how patients, as a result of being dissatisfied with public sector services, use a variety of strategies in order to gain access to better quality treatment and to less expensive or free-of-charge treatments [23-27]. These strategies tend to be patterned according to a person's socioeconomic status in such a way that those with higher education use their personal networks in order to access services, while those with lower education and relatively good financial resources prefer paid services. The socioeconomic patterning is likely to reflect the more extensive networks of those with higher education [25]. Against this background, it seems likely that the larger variety of health services used by women with a university education in our study is related to their personal networks, which enable them to access additional services regardless of their financial position.

High personal, and middle and high family income were related to more common use of private health services, but not with other additional services. Private services sometimes involve considerably higher user charges, but nevertheless they can be accessed easily if the user has the financial assets. Private services tend to be provided in more comfortable settings and they usually guarantee a higher level of privacy [15]. At the same time, distrust towards and dissatisfaction with public sector services is widespread [23-26]. Thus, it is not surprising that those who can afford it seek additional care from the private sector.

As for the utilisation of preventive health services, we found that, independently of personal and family income, women with a university degree were more likely to obtain a Pap smear and gynaecological examination, although in all education groups the prevalence of Pap smear and gynaecological examination was lower than recommended. Both examinations are essential and widely accepted health promotion measures for reproductive-age women. For comparison, a US study reported that 91.4% of reproductive-age women (18-40 years old) had been for a Pap smear, while in our sample the proportion was only 36.3% and 47% for those with a university education [28]. Our results suggest that the higher the woman's education, the more aware she is of the importance of preventive examinations. Financial barriers are unlikely to explain the whole story behind the underutilisation of the Pap smear and gynaecological examination, as the fee for the PAP smear is very small (less than 3 euro in 2004) and the latter can be obtained free-of-charge at public sector women's clinics. However, screening for oncology conditions can lead to the discovery of the disease, resulting in further treatment, which could act as a psychological as well as a strong financial barrier to the Pap smear among those worse-off, even though the Pap smear itself is cheap.

Women with higher personal income had a mammography more often, irrespective of their educational status and family income. This suggests that high personal income enabled women to have the diagnostic procedure done.