Sociodemographic and socioeconomic determinants of health services utilization in Greece: the Hellas Health I study

Abstract

The purpose of the study was to estimate the demographic and socioeconomic determinants of utilization of the Greek primary and hospital health care services.

Data were obtained from the cross-sectional nationwide household survey Hellas Health I (2006). The sample (N = 1005) was representative of the Greek adult population in terms of age and residency, and was selected by means of a three-stage, proportional-to-size sampling design.

The presence of a family doctor was reported in a higher degree by participants of higher social classes and private insurance. After adjusting for self-perceived general health and chronic illness, contacts with health care professionals during the past four weeks were found less for residents of rural areas, while contacts with health care professionals during the past 12 months were found less for men than women, for individuals without private insurance and for individuals of lower education. More out-of-pocket payments were reported by the 34–44 age group, rural area residents and individuals with private insurance. Higher use of private health care services was reported by participants of higher social classes and residents of rural areas and private insurance. Only hospital admissions were not directly influenced by demographic and socioeconomic factors.

The findings imply the existence of inequities in access and use of primary health services with clear implications to related policies.

Introduction

National health systems of universal coverage face common challenges today in their attempt of achieving their major goal of equal use for equal need.¹ The aging of the population, the new costly medicines and medical technologies together with major political shifts following the principles of global economy, do threaten equity in health and require targeted governmental health policies. So, a shared challenge in the 21st century is how to redesign primary health care to make it more accessible, continuous, coordinated and patient-oriented, a challenge that comprises an emphasis on promoting health.² The common strive to tackle the pro-reach inequity, exacerbated by private insurance and privatization of health services, remains equally important.^1,3 Consequently, monitoring the access and use of health services at the national level, in a systematic and thorough manner, is vital in order to measure equity in health and assess whether it is threatened by the widespread health sector reforms.

There is general agreement that Greece has not yet fully established a comprehensive and universal health-care service following the introduction of the national health system (ESY) in 1983.⁴ ESY is a mixed-type system, encompassing elements from the two basic health care system types (i.e. Beveridge and Bismarck). Public health services are provided via the national health system (ESY) (i.e. public hospitals, rural health centres and provincial surgeries) and the social insurance funds (i.e. primary health care centres, contracted private doctors and diagnostic centres and a few hospitals), with the main provider known as IKA covering private employees. There is also a well-developed private sector with more than 20,000 doctors running private offices, 400 diagnostic centres and 218 hospitals.⁵ Most of the doctors who offer primary care services are specialists, since general practitioners (GPs) constitute less than 2% of the total Greek doctors and there is no gate-keeping system.

The Greek health system provides free access to health care services to all citizens. Free access, however, does not compensate for the low quality and inefficiency of the health services. The basic aspects of the problem involve a heavily centralized and unorganized administrative framework, low levels of public expenditure, a significant private sector under loose governmental control, a fragmented primary health care network, inequitable coverage of health services, perverse incentives for providers, escalating costs and inadequate national public health policy.^6,4 Greece spends approximately 10% of its GDP on health care – a percentage above the median of the EU-15, yet its per capita GDP is one of the lowest and its citizens are the least satisfied with their health services.⁷ The realization of these problems and the resulting need to improve and modernize the health system has led to various attempts for reform, which have been undertaken over the last 30 years.⁸ However, the low degree of the public's satisfaction with health care services simply confirms the failure of these reforms and stresses the need to provide guidance through empirical research.⁹

Inequalities in the Greek health system have been reported mainly in theoretical papers and to a lesser degree through empirical research. They mainly concern: (1) unequal coverage that results from a fragmented financing system, varying contribution rates and differing benefits packages across different insurance funds; (2) regional inequalities in the quantity and quality of services provided, caused by distortions in the allocation of resources; (3) inequalities in the access and utilization of health services based on various sociodemographic and socioeconomic characteristics. In this paper, we are mainly going to examine the latter category.

Because there is no official system of monitoring the use of health services at the national level, research is needed to fill this gap. So, the aim of the present study was to identify the major factors that determine the utilization of health services by Greek adults and to examine the existence of any socioeconomic inequalities. Both primary and hospital health services utilization were explored in terms of demographic and socioeconomic factors, controlling for the level of health care needs and self-reported health level. It should be noted that this is the first Greek study on health services utilization that uses data from interviews from a country-wide representative population sample. The limited studies conducted so far, were either of regional scope¹⁰ or have been conducted through questionnaires sent by post,¹¹ thus compromising their generalizability and reliability.

Recording and comprehending the factors that influence health services utilization is important in order to assess the demand for health care services, track and reduce potential sources of social, economical or geographical inequalities, monitor and restrict high private expenditure and use of the private sector, and address structural and organizational problems within the Greek health system. In general, the results and conclusions of the study can assist in the development of a comprehensive national health policy and support health system structural reforms.

Methodology

Sample and procedure

Data were obtained from the national household survey Hellas Health I, conducted during 2006 by the Centre of Health Services Research of the Department of Hygiene and Epidemiology, University of Athens. According to the last population census, the survey population consists of approximately 8,880,924 individuals. The survey covered urban areas (2000 or more inhabitants) and rural areas (less than 2000 inhabitants) of all geographical regions of the country.

Candidate respondents were selected by means of a three-stage, proportional to size sampling design. In the first stage, a random sample of building blocks was selected proportionally to size, based on the 2001 Population Census of the National Statistical Service of Greece. In the second stage, in each selected area of blocks, the households to be interviewed were randomly selected by means of systematic sampling. Any person, or group of persons, living in a separate housing unit was considered as a ‘household’ unit. In the third stage, in each household, a sample of individuals aged 18 years old or more was selected by means of simple random sampling.

The survey population consisted of 1005 individuals out of 1388 fluent speakers of the Greek language, who were initially approached (response rate 72.4%) and were interviewed face to face by trained interviewers. The sample was representative of Greek population in terms of age and residency. The questionnaire included 146 questions or subquestions and 13 open-ended questions, relating to sociodemographic, health needs and utilization measures.

Measures

Sociodemographic and socioeconomic measures

Sociodemographic questions referred to age groups (i.e. 18–24, 25–34, 35–44, 45–54, 55–64 and >65 years old), sex, marital status (i.e. married, single or divorced/widowed), residency (i.e. urban–rural), educational level (i.e. low, middle and high), type of health insurance coverage (i.e. private, public or both) and socioeconomic classes. Socioeconomic classification was based on the Esomar guidelines.¹² More specifically, social classes were assessed on the basis of the job category of the family's main income earner and its level of education, and summed up into three social categories: A/B-C1 (upper to upper middle), C2 (lower middle) and D/E (lower).

Health care needs proxies

Chronic diseases

Respondents were asked whether they suffered from any chronic diseases such as diabetes, hypertension, hypercholesterolaemia, etc. The number of chronic diseases of the responders was evaluated.

Health-related quality of life

The generic SF-36 health-related quality-of-life questionnaire was used to evaluate general subjective health of the respondents.¹³ The physical (PCS-36) and mental (MCS-36) component summary scores, were deviation scores from a mean of 50. The SF-36 has been found to have good construct validity and reliability for the Greek population.¹⁴

Health service utilization

The questions concerning the patterns of health services utilization, constituted the dependent variables and covered both primary and secondary level of health services. The probability of utilization of primary health care was analysed with the use of questions concerning: (1) existence of a family doctor (i.e. ‘does your family have a doctor who can provide you with medical advice when in need?’, (2) visits to health professionals over the past four weeks (i.e. ‘over the past 4 weeks did you see or speak to a health professional about your health? exclude dental services or hospital admissions’), and (3) visits to health professionals over the past 12 months (i.e.‘over the past 12 months did you see or speak to a health professional about your health? exclude dental services or hospital admissions’). The outcomes constituted binary (yes–no) variables. Respondents who answered, ‘yes’ in the two previous questions, were then asked of the number of medical visits over the respective period. The questions assessing the frequency of utilization were open ended and the corresponding variables were numeric. The utilization of hospital services was analysed via questions concerning the probability of hospitalization and the number of hospitalizations over the past 12 months.

In addition, respondents were asked on any out-of-pocket payments during any of their previous visits to health care professionals (i.e. ‘in any of these visits did you have to pay a non-refundable amount out of your pocket?’). Out-of-pocket payments were defined here as (a) payments for services not covered by any form of prepayment or insurance, (b) unofficial, illegal payments for services that should be fully funded (i.e. usually referred as ‘envelope’ or under-the-table payments), and (c) cost sharing/user charges (i.e. payments of part of the cost of care received).

Respondents were then asked about the type of health services they usually refer to, when in need of medical care (i.e. ‘where do you usually turn to when in need of medical care?). The answer options were grouped as follows: All rural health care centres and provincial surgeries belonging to national health service (ESY), primary health centres belonging to IKA and other social insurance funds, outpatient departments of public hospitals and private doctors contracted by insurance funds were categorized as public health services. Private doctors not contracted by social insurance funds and outpatient departments of private hospitals were categorized as private health services. Finally, respondents were asked about any hospital admissions in the last 12 months (i.e. ‘were you admitted to any hospital in the last 12 months?’)

Statistical analysis

Analyses were conducted on full data without missing values. Continuous variables are presented with mean and standard deviation, while quantitative variables are presented with absolute and relative frequencies. For the comparisons of proportions, chi-square tests were used. Student's t-tests were computed for the comparison of mean values. Multiple logistic regression analysis was performed in order to investigate the independent effect of factors with dependent binary variables, presenting health services utilization (i.e. consultation with a health-care professional over the past 4 weeks, family doctor, hospitalizations in the past 12 months). In case of dependent continuous variables (i.e. number of consultations with a health care professional over the past 4 weeks or the past 12 months) multiple linear regression analysis was used using their ranks, as they were not normally distributed according to the Kolmogorov–Smirnov criterion (P < 0.05). The regression equation included terms for sex, age, family status, socioeconomic status, educational level, residence, chronic disease, type of health insurance and self-perceived health status. From the results of the logistic regression analyses, adjusted odds ratios with 95% confidence intervals were computed, while from the results of linear regression analyses regression coefficients (β) with their standard errors were computed. All P values reported are two-tailed. Statistical significance was set at 0.05 and analyses were conducted using SPSS statistical software (version 13.0).

Results

Data from 958 participants were analysed. Sample characteristics are presented in Table 1. A 50.2% of the responders had a family doctor. A 68.7% of individuals belonging to the higher social class (A/B-C1) reported having a family doctor, versus 45.1% and 48% belonging to the lower social classes C2 and D/E, respectively. Of the subjects with a private coverage, 68.5% had a family doctor while the proportion for those without a private coverage was 47.8%. In univariate analyses (Table 2), the presence of a family doctor was significantly associated with higher educational level, higher socioeconomic status, chronic disease and private health insurance. Adjusting for all factors, educational level was no longer a significant predictor (Table 3). On the other hand, subjects belonging to the lowest social classes (i.e. C2 or D/E) had a significantly lower likelihood to refer to a family doctor in multivariate analysis. Furthermore, those with private insurance coverage had greater odds to refer to a family doctor with odds ratios equal to 2.18 (95% confidence interval [CI]: 1.40–3.41).

Table 1

Sample characteristics

	N	%
Sex
Men	460	48.0
Women	498	52.0
Age (years)
18–24	111	11.6
25–34	177	18.5
35–44	171	17.8
45–54	147	15.3
55–64	143	14.9
>65	209	21.8
Educational status
Low	315	32.9
Middle	404	42.2
High	239	24.9
Family status
Married	619	64.6
Single	232	24.2
Divorced/widowed	107	11.2
Socioeconomic status
A/B-C1	163	17.0
C2	443	46.2
D/E	352	36.7
Residence
Urban	719	75.1
Rural	239	24.9
Chronic disease
No	620	64.7
Yes	338	35.3
Private health insurance
No	847	88.4
Yes	111	11.6
PCS, mean (SD)	49.9 (10.8)
MCS, mean (SD)	47.5 (9.41)

PCS = physical component summary score; MCS= mental component summary score

Table 2

Proportions of subjects referring to a family doctor, having a consultation with a health care professional during the past four weeks or a hospitalization during the past 12 months and usual source of care (private versus public) when in need, according to sociodemographics and self-perceived health status

	Family doctor		Consultation with a health care professional during the past four weeks		Hospitalizations in the past 12 months		Private doctor or clinic as usual type of service
	N (%)	P	N (%)	P	N (%)	P	N (%)	P
Sex
Men	230 (50.0)	0.901	115 (25)	0.006	49 (10.7)	0.439	115 (25.4)	0.331
Women	251 (50.4)		165 (33.1)		61 (12.2)		139 (28.3)
Age (years)
18–24	53 (47.7)	0.627	27 (24.3)	<0.001	9 (8.1)	0.003	27 (24.8)	0.018
25–34	91 (51.4)		38 (21.5)		12 (6.8)		54 (30.9)
35–44	78 (45.6)		29 (17.0)		15 (8.8)		59 (35.1)
45–54	71 (48.3)		40 (27.2)		14 (9.5)		40 (27.6)
55–64	76 (53.1)		50 (35.0)		22 (15.4)		33 (23.7)
>65	112 (53.6)		96 (45.9)		38 (18.2)		41 (19.7)
Educational level
Low	146 (46.3)	0.031	116 (36.8)	0.001	48 (15.2)	0.012	64 (20.6)	0.002
Middle	198 (49)		111 (27.5)		45 (11.1)		109 (27.5)
High	137 (57.3)		53 (22.2)		17 (7.1)		81 (34.3)
Family status
Married	313 (50.6)	0.745	185 (29.9)	0.003	80 (12.9)	0.164	173 (28.4)	0.049
Single	118 (50.9)		52 (22.4)		20 (8.6)		63 (27.6)
Divorced/widowed	50 (46.7)		43 (40.2)		10 (9.3)		18 (17)
Socioeconomic status
A/B-C1	112 (68.7)	<0.001	45 (27.6)	0.322	15 (9.2)	0.318	59 (36.6)	0.006
C2	200 (45.1)		140 (31.6)		58 (13.1)		115 (26.3)
D/E	169 (48.0)		95 (27.0)		37 (10.5)		80 (23.1)
Residence
Urban	371 (51.6)	0.135	224 (31.2)	0.023	81 (11.3)	0.715	175 (24.6)	0.006
Rural	110 (46)		56 (23.4)		29 (12.1)		79 (33.9)
Chronic disease
No	295 (47.6)	0.028	129 (20.8)	<0.001	55 (8.9)	0.001	173 (28.4)	0.161
Yes	186 (55)		151 (44.7)		55 (16.3)		81 (24.2)
Private health insurance
No	405 (47.8)	<0.001	245 (28.9)	0.570	103 (12.2)	0.069	200 (24)	<0.001
Yes	76 (68.5)		35 (31.5)		7 (6.3)		54 (49.5)
PCS, mean (SD)*	50.1 (10.4)versus49.6 (11.2)	0.539	52.5 (8.6)versus43.5 (12.9)	<0.001	50.8 (10.1)versus42.9 (13.5)	<0.001	49.5 (11.0)versus50.8 (10.3)	0.115
MCS, mean (SD)*	47.3 (9.3)versus47.7 (9.4)	0.500	48.5 (8.6)versus45.3 (10.8)	<0.001	47.9 (9.1)versus44.8 (11.2)	0.001	47.5 (9.6)versus47.6 (8.7)	0.877

PCS = physical component summary score; MCS = mental component summary score

*Mean values (standard deviation) of PCS and MCS for subjects who did not use the aforementioned health services versus those who did

Table 3

Results from multiple logistic regression analyses concerning sociodemographics and self-perceived health status factors associated with having a family doctor, having a consultation with a health care professional during the past four weeks, hospitalization during the past 12 months and usual source of care (private versus public) when in need

	Family doctor		Consultation with a health care professional during the past four weeks		Hospitalizations in the past 12 months		Private doctor or clinic as usual type of service
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Sex
Men	1.0*		1.0*		1.0*		1.0	–
Women	1.12 (0.85–1.47)	0.436	1.25 (0.9–1.73)	0.182	1.13 (0.73–1.75)	0.587	1.29 (0.96–1.75)	0.095
Age (years)
18–24	1.0		1.0		1.0		1.0	–
25–34	1.19 (0.7–2.04)	0.520	0.76 (0.4–1.46)	0.416	0.96 (0.35–2.59)	0.931	1.22 (0.67–2.24)	0.514
35–44	1 (0.55–1.82)	0.997	0.45 (0.19–1.16)	0.140	1.1 (0.38–3.16)	0.866	1.62 (0.83–3.16)	0.160
45–54	1.04 (0.55–1.97)	0.910	0.64 (0.3–1.37)	0.252	1.07 (0.35–3.25)	0.906	1.04 (0.51–2.13)	0.916
55–64	1.42 (0.72–2.78)	0.309	0.63 (0.28–1.4)	0.258	1.46 (0.48–4.45)	0.509	1.12 (0.52–2.42)	0.774
> 65	1.61 (0.81–3.2)	0.170	0.81 (0.37–1.81)	0.613	1.44 (0.47–4.46)	0.525	1 (0.45–2.2)	0.999
Educational level
Low	1.0		1.0		1.0		1.0	–
Middle	1.29 (0.90–1.85)	0.167	1.38 (0.89–2.13)	0.151	1.27 (0.73–2.21)	0.405	1.33 (0.87–2.04)	0.192
High	1.18 (0.75–1.86)	0.469	0.92 (0.53–1.6)	0.769	0.76 (0.35–1.64)	0.483	1.59 (0.95–2.67)	0.077
Family status
Married	1.0		1.0		1.0		1.0	–
Single	1.08 (0.71–1.66)	0.719	0.84 (0.5–1.43)	0.531	1.02 (0.51–2.03)	0.958	0.92 (0.58–1.47)	0.73
Divorced/widowed	0.82 (0.52–1.28)	0.383	0.84 (0.51–1.41)	0.517	0.54 (0.26–1.08)	0.083	0.58 (0.32–1.02)	0.061
Socioeconomic status
A/B-C1	1.0		1.0		1.0		1.0	–
C2	0.40 (0.26–0.61)	<0.001	0.9 (0.55–1.48)	0.682	0.90 (0.44–1.83)	0.778	0.70 (0.45–1.1)	0.121
D/E	0.46 (0.29–0.74)	0.001	0.77 (0.45–1.32)	0.341	0.66 (0.3–1.44)	0.294	0.58 (0.35–0.95)	0.032
Residence
Urban	1.0		1.0		1.0		1.0	–
Rural	0.93 (0.68–1.28)	0.669	0.50 (0.33–0.75)	0.001	0.93 (0.56–1.54)	0.769	2.23 (1.55–3.2)	<0.001
Chronic disease
No	1.0		1.0		1.0		1.0	–
Yes	1.37 (0.98–1.91)	0.069	1.51 (1.04–2.21)	0.032	1.31 (0.81–2.11)	0.280	0.98 (0.67–1.44)	0.909
Private health insurance
No	1.0		1.0		1.0		1.0	–
Yes	2.18 (1.40–3.41)	0.001	1.41 (0.87–2.31)	0.165	0.53 (0.23–1.23)	0.140	2.76 (1.78–4.27)	<0.001
PCS	1.00 (0.99–1.02)	0.902	0.93 (0.92–0.95)	<0.001	0.95 (0.93–0.97)	<0.001	1.00 (0.98–1.02)	0.914
MCS	1.00 (0.99–1.02)	0.655	0.98 (0.96–0.99)	0.003	0.98 (0.96–1.00)	0.050	0.99 (0.98–1.01)	0.517

OR = odds ratio; CI = 95% confidence interval; PCS = physical component summary score; MCS = mental component summary score

*Indicates reference category

Bold values indicate statistical significant results (P = 0.05)

Approximately, 29.2% of the participants reported at least one consultation with a health care professional over the past four weeks. In univariate analysis (Table 2), female gender, older age, lower educational status, being divorced or widowed, urban residence, having a chronic disease and lower self-perceived health status – both PCS and MCS – were associated with an increase in reporting a consultation. When multivariate analysis was conducted, only residence, chronic disease and the PCS, MCS dimensions were independently associated with consultation with a health care professional over the past four weeks. The likelihood for consulting a health care professional over the past four weeks was 50% lower for rural areas compared with urban areas, while subjects with a chronic disease had 1.51 times greater odds for having a consultation than subjects with no chronic disease. Also, greater PCS and MCS scores were independently associated with a lower likelihood of consultation with a health care professional during the past four weeks.

The number of consultations with a health care professional over the past four weeks had a mean of 1.0 (SD = 6.5) and a median equal to 0 (interquartile range: 0–1), while the number of consultations with a health-care professional during the past 12 months had a mean of 5.6 (SD = 15.0) and a median equal to 2 (interquartile range: 0–4). When multiple analyses were conducted with a dependent variable, the number of consultations with a health care professional over the past four weeks (data not shown), the factors significantly associated were chronic disease (β = 32.3, SE = 16.41, P = 0.049), residence (β = −53.29, SE = 15.82, P = 0.001), PCS score (β = −7.76, SE = 0.74, P < 0.001) and MCS score (β = −2.74, SE = 0.72, P < 0.001). The mean number of consultations with a health care professional over the past four weeks for those with chronic disease was 1.9 (SD = 9.6) and the median equal to 0 (interquartile range: 0 to 1), while for those without chronic disease it was 0.5 (SD = 3.7) and the median equal to 0 (interquartile range: 0–0). Additionally, the mean number of consultations over the past four weeks for those from urban areas was 1.1 (SD = 6.7) and the median equal to 0 (interquartile range: 0–1), while for those from rural areas was 0.8 (SD = 5.7) and the median equal to 0 (interquartile range: 0–0). Additionally, greater PCS and MSC scores were related to a lower number of consultations during the past four weeks.

The results of multiple linear regression analysis with dependent variable the number of consultations with a health care professional over the past 12 months (data not shown) indicated that women (β = 35.22, SE = 11.07, P = 0.039), those of high educational level compared with those of low educational level (β = 64.54, SE = 27.98, P = 0.021), those with chronic disease(s) (β = 74.45, SE = 20.76, P < 0.001) and those with private insurance coverage (β = 60.96, SE = 26.26, P = 0.021) had a greater number of consultations. The mean annual number of medical visits for women was 5.7 (SD = 14.1) and the median equal to 2 (interquartile range: 0–5) and for men it was 5.4 (SD = 15.9) and the median equal to 1 (interquartile range: 0–3.5). The mean annual rate for those with a chronic disease was 8.6 (SD = 17.5) and the median equal to 3 (interquartile range: 1–10), while for those without chronic disease it was 3.9 (SD = 13.1) and the median equal to 1 (interquartile range: 0–3). The mean annual rate for those having private coverage was 7.6 (SD = 18.7) and the median equal to 2 (interquartile range: 0–5) and for those without private coverage it was 5.3 (SD = 14.4) and the median equal to 2 (interquartile range: 0–4). PCS and MCS dimensions revealed an inverse significant association with the number of consultations over the past 12 months (β = −8.54, SE = 0.93, P < 0.001 and β = −4.54, SE = 0.90, P < 0.001, respectively).

In terms of out-of-pocket payments during any of the above visits to health professionals, 39% of the respondents declared having paid a non-refundable amount, while 61% reported no out-of-pocket payments. Individuals aged 34–44 years, inhabitants of rural areas and those with private insurance reported significantly more out-of-pocket payments during consultations with health care professionals (Table 4).

Table 4

Percentages of respondents declaring out-of-pocket non-refundable payments during any of their previous visits to health professionals

	Out-of-pocket payments during health care service visits
	N (%)	P
Sex
Men	44 (38.3)	0.834
Women	64 (39.5)
Age (years)
18–24	10 (37)	0.044
25–34	19 (51.4)
35–44	17 (58.6)
45–54	10 (25)
55–64	20 (40)
>65	32 (34)
Educational status
Low	43 (37.7)	0.716
Middle	46 (41.8)
High	19 (35.8)
Family status
Married	72 (39.3)	0.886
Single	21 (40.4)
Divorced/widowed	15 (35.7)
Socioeconomic status
A/B-C1	19 (42.2)	0.178
C2	46 (33.6)
D/E	43 (45.3)
Residence
Urban	79 (35.7)	0.028
Rural	29 (51.8)
Chronic disease
No	53 (41.4)	0.445
Yes	55 (36.9)
Private health insurance
No	89 (36.8)	0.047
Yes	19 (54.3)
PCS*	43.4 (12.4) versus 43.5 (15.8)	0.970
MCS*	46.2 (11.0) versus 44.5 (10.5)	0.193

PCS = physical component summary score; MCS = mental component summary score

*Mean values (standard deviation) of PCS and MCS for subjects who did not declare out-of-pocket non-refundable payments versus those who did

Bold values indicate statistical significant results (P = 0.05)

In terms of the type of usual source of health care (i.e. public or private) when in need, 26.9% usually turns to private health services (i.e. 26.7% visit private doctors not contracted by their insurance fund and 0.2% visit outpatient departments of private hospitals). The majority (73.1%) of respondents report public services as the usual provider of health care when in need, with 24.5% of them visiting primary health care centres belonging to IKA or other social insurance funds, 25.6% visiting contracted private doctors, 12.7% outpatient departments of public hospitals and 9.6% ESY rural health centres or provincial surgeries. Multiple analyses revealed that individuals belonging to the lower social class (D/E), were about 42% less likely to consult a private doctor or private clinic/hospital compared with individuals belonging to the higher social class (A/B-C1). Finally, greater likelihood to consult a private doctor or clinic was found for individuals living in rural areas and for individuals having private health insurance (Table 3).

An 11.5% of the responders had been admitted to a hospital one or more times in the past 12 months. Educational level, age and chronic disease were associated with hospitalizations in bivariate but not in multiple analyses. PCS and MCS dimensions predicted significant hospitalization in multiple analyses with odds ratios equal to 0.95 (95% CI: 0.93–0.97) and 0.98 (95% CI: 0.96–1.00), respectively.

Discussion

The present study examined the presence of demographic and socioeconomic-related differences in need-adjusted use of health services at both primary and hospital level of care, with results revealing demographic and socially related inequities in terms of utilization of primary health-care services. Another study using Eurobarometer data from 1996, found that income was a determinant of the utilization of primary services in Greece.¹⁵ These are in contrast to many other studies originated from other countries – Australia, Ireland, New Zealand, the UK and the majority of the OECD countries in van Doorslaer and colleagues' papers – where there was little or no evidence of any variation in use of primary care by socioeconomic group, after adjusting for differential need.^16–21

With regard to the presence of a family doctor, almost half of the participants (50.2%) answered that they have a doctor whom they consult on a regular basis. This proportion is relatively small compared with other European countries. In Britain, 98% of total population is registered with a GP through the national health system.²² In the 2002 Swiss Health Survey, 90% of the 35–70-year-old respondents had a personal doctor.²³ The presence of a family doctor is believed to be associated with more personal, continuous and comprehensive care and a better coordination of specialists services.²⁴ It should be stated that in Greece, the institution of the family doctor is unofficial and not well established. Greek people usually refer to different specialists for their health problems, according to perceived need, but none of them bear responsibility for the patient as a whole. In some cases, people consult a single provider–specialist– regularly and they consider him/her as their ‘personal’ or ‘family’ doctor.

The analysis showed that the highest social classes, those with chronic disease(s) and those possessing private insurance, reported presence of a family doctor in a significantly higher degree. Those who belong to the middle and lower classes are significantly less probable to have a family doctor compared with those who belong to the higher social class, stressing a pro-rich inequality. This result is in accordance to the results of earlier studies, which point out that those who belong in the lower social classes are less probable to have a personal doctor²⁴ and receive less often a doctor's advice.^25–27 This is mainly explained by the fact that the duties of family doctors in Greece are usually carried out by private doctors – most specialized in internal medicine. So, the higher social classes, who can either make direct payments or possess private insurance, can afford to have a family doctor.⁸ Also, 68.5% of those having private insurance reported having a family doctor, versus 47.8% of those without private insurance. It is assumed that this is because individuals with private insurance are often free to choose the doctor they prefer with no extra cost.^4,5 Although inequalities observed between different social insurance funds is out of the scope of this paper, it should be noted that the beneficiaries of the more prosperous social insurance funds receive more and better services than the ones of the less privileged insurance funds, including access to a family doctor.²⁸

Around 29.2% of the Greek adults in the present study had come in contact with a health professional over the past four weeks. Similar percentages (26%) have been reported in neighbouring countries, such as Italy.²⁹ In the present study, residence and level of health needs (i.e. chronic disease and low self-assessed general health), were shown to be the most important factors that influenced contact with a health professional over a period of four weeks. Specifically, 31.2% of urban residents versus 23.4% of rural residents had visited a doctor over the previous month. In Greek rural areas, the absence of adequate health care infrastructures and human resources, the fewer options for access to health care – the rural population insured by the Organismos Georgikon Asfaliseon (Agriculture Insurance Organisation; OGA) does not have its own health care network or free access to private doctors and clinics as the beneficiaries of other social funds – and the larger time and travelling costs for accessing health care are translated into important geographical differences in utilization of the primary health care services in relation to the urban centres.^3,4,30 On the contrary, another study on health care access in Greece based on mailed surveys in 2001, found no regional differences in access to health services.¹¹

The average number of consultations with a health professional over the past 12 months was 5.6, and for the last four weeks was 1, whereas prior studies reported slightly different values compared with the present one; lower in terms of mean annual number of visits (i.e. 3.9) and higher in terms of per month number of visits to outpatient services (i.e. 2.5).^11,30,31 Similar mean annual numbers of total medical visits have been reported in most European Union (EU) countries.^3,32 It has been estimated that the number of annual visits to primary health care services per person in developed countries is 4–8.³²

In the present study, the number of annual visits to health care professionals was greater for those with higher educational level (i.e. high and middle) versus those with low educational level, for women, for those with chronic disease(s) and for those with private insurance. Higher use of primary services by women has also been described in other studies.³³ It is argued that women have more health needs and are more aware of health matters than men.^27,34,35 The final observed positive influence of the existence of private insurance on the frequency of medical visits over a year (mean number 7.6 as opposed to 5.3 when a private coverage did not exist) probably correlates with the reimbursement of the immediate payments. It is widely acknowledged that private insurance coverage increases the frequency of health care consultations by richer people.^25,32,36 The lower use of primary health services by people of low educational level could be attributed to the lack of information on matters concerning their health.

The results did not show, however, any difference in the monthly or annual number of visits to health care professionals between people of different socioeconomic classes. This can be interpreted as an indication that even the poor do share the same use of primary care services as the rich, although the care they receive is not always of the same quality, as is shown by the lower percentage of presence of a family doctor among lower social classes. The study could not distinguish between GP and specialist consultations as most European studies of health services utilization do, since GP consultations in Greece are rare.

Out-of-pocket payments during any of the visits to health professionals were reported by 39% of the respondents. Inhabitants of rural areas, individuals aged 34–44 years and those possessing private insurance, reported significantly more out-of-pocket payments. Out-of-pocket payments (mainly comprising direct or cost sharing for dental or primary care) and informal payments (for NHS hospital care) have been previously estimated to account for 47.5% of the total health expenditure, a percentage that is the highest in the EU and among the OECD countries.^4,5 They reflect the inability of the state to achieve a comprehensive coverage of the population and offer high-class services.^37,38 Although there is limited evidence on how informal payments affect the access and utilization of health services, it is clear that those patients who do not choose or cannot afford to pay, cannot access the same level of services or have to wait longer for care.³⁹

The higher probability of paying out-of-pocket for individuals aged 34–44 years old (58.6%) can be explained by their increased payment ability, since they are in a high productive phase. The increased out-of-pocket payments by the rural residents (i.e. 51.8% compared with 35.7% of inhabitants of urban areas), however, represents a different case. It does not reflect a bigger purchasing power, but an effort of surpassing the organizational obstacles in the access to public primary health services in rural areas, either by making consultations to private doctors or by paying money to doctors of public rural posts for some services (i.e. home visits) that are supposed to be free, but not so easily offered.^4,5 Also, the proportion of privately covered individuals that paid out-of-pocket sums was higher (54.3%) than the corresponding proportion of individuals without a private insurance (36.8%). The acceptance of a correlation between private insurance and high income, reinforces the belief that individuals who select private health insurance have a high purchasing power that makes them capable of paying out-of-pocket in both the public and private health sector.^40,41

Findings regarding the usual type of service (i.e. public or private) when in need for care, imply certain socioeconomic inequalities. Of the respondents, 26.9% consult a private provider of primary health care compared with 73.1% that consult a public one. A similar proportion (22%) of private visits in Europe has been noted only in Finland.³⁷ It is noted that Greece has the most ‘privatized’ health care system among all EU countries.⁴² The extensive use of private health services can be explained by the chronic under funding and low quality of public primary care, the fragmentation of health services and the lack of organization and regulation in both the public and private sectors.⁶

In the present study, greater likelihood to refer to private doctors or clinics was found for individuals living in rural areas and those being of a higher socio-educational level, versus individuals belonging to the lowest social class. The finding that private health care services are more popular among the rural residents, seems to be a consequence of the inadequate provision of the public primary health services in rural areas. Actually, as it was previously explained, the rural population is forced to use more services from the private sector, thus burdening their rather limited family budget. The higher probability of visiting a private doctor for those with higher socioeconomic levels is in accordance with another Greek study.¹⁰ The socially privileged individuals who enjoy high standard of living conditions and a high purchasing power, usually consume private health services in order to assure the quality of provided services and save time. Indeed, private visits are concentrated in high-income groups in many European countries.^31,40,43 However, in Greece, not only the higher socioeconomic classes but even the lower ones have often no option but to use private health services (23% of individuals of the lowest class usually consults a private provider when in need), thus rising issues of horizontal and vertical inequity. Educational level constitutes a factor that substitutes the public health services with the private ones in many studies.^10,29,36,44

Finally, with regard to hospitalizations, 11.5% of the respondents reported hospitalization over the past 12 months. In other European studies, such as ones in Belgium and Spain, the corresponding proportion was 13%²⁶ and 12%, respectively.³⁴ In an analysis of health services utilization in the wider region of Athens, the mean annual rate of hospital admissions was found to be 12.2%.¹⁰ Although in the present study, hospital admissions were related to educational level, age and health needs in the univariate analyses, only low self-assessed general health remained a significant prognostic factor for higher hospitalization in the multivariate analyses. This result shows the absence of any socioeconomic barriers in the use of ESY hospital services and lies in agreement with the conclusions of other Greek¹⁰ and European surveys.^25,27,30,44 This can be interpreted as an indication that secondary and tertiary care in Greece is provided according to the health needs of the population and that ESY hospitals do not discriminate between people in need. It is noted that the ESY provides free access to all insured population and offers an adequate number of beds (4.7/1000 persons).⁴⁵ In contrast, in a systematic review of equity in the use of curative health services in various universal systems, the authors found a pro-rich bias in the use of many specialist hospital services.¹

Conclusion

All in all, the above results revealed a number of demographically and socially related inequities in the use of primary health care services by people with equal needs. After adjustment for need, the residents of rural areas reported less contact with health care professionals over the past four weeks and more out-of-pocket payments during these visits than the residents of urban areas. Additionally, rural residents reported significantly higher use of private services (i.e. private doctors or clinics) as their usual source of care. Social class in need-adjusted use of health services was associated with usual source of health care (i.e. private or public) and the presence of a family doctor. The higher social class (i.e. high/middle) was significantly associated with higher use of private services as usual source of health care, compared with the lower social class. Additionally, the higher social class was associated with the presence of a family doctor. After controlling for health care needs, educational level remained a significant determinant, only in terms of the intensity of use of primary health care services, with those of higher educational level (i.e. high and middle) reporting more visits to health professionals over the past year than those with low educational level. Having private insurance was found to be a significant factor associated with health service utilization, after controlling for health care needs. Specifically, those with private insurance reported more contacts with health care professionals in a year, higher use of private services and presence of a family doctor in a higher degree. Gender was associated with the intensity of health services use, with women reporting more visits to health professionals over the past year. Individuals aged 34–44 years reported more out-of-pocket payments (58.6%).

The above findings reflect that the patterns of health services utilization in Greece are quite different between primary and hospital care. On the one hand, secondary and tertiary care health needs are covered on a rather fair and equitable basis. On the other hand, primary care health services are characterized by demographic and socioeconomic inequities to a significant degree, despite the fact that the establishment of the ESY in 1983 was based on the principles of equity in the delivery of health care services, primary health care development and decentralization.

We believe that the results of the study make clear that we are far from the goal of providing equal primary care for equal need. The only way to tackle the inequalities observed, is through a series of reforms that aim at the improvement of the public primary care sector. The institution of the family doctor who can develop a personal, patient-centred relationship with the patient and offer a broad range of services – including disease prevention and health promotion activities – could ensure continuity of care and reasonable use of specialized services. The strengthening of primary health care services in rural areas, through the provision of more human, financial and technological resources and the implementation of governance changes, will reduce the pro-urban in equalities. The control of the vast corruption in both public and private primary care services and the unification of today's fragmented system of urban primary care services, could reduce the waste of public and private resources, which burdens the whole Greek society, especially the unprivileged. If these reforms lead to better public primary services, the private sector will dissolve gradually and the pro-rich inequalities will be diminished. The constant and permanent recording and monitoring of utilization patterns is a necessary tool to evaluate the success of any reforms taken, and so we need to reinforce our efforts towards this direction.

Conflict of interest and sources of funding statement

On behalf of my co-authors, I submit the manuscript with the title ‘Sociodemographic and socioeconomic determinants of health services utilization in Greece: the Hellas Health I study’ for consideration by the Journal of Health Services Management Research. We authors declare that no external funding, apart from the support of the authors' institution, was available for this study. We also declare that there are no conflicts of interest in this study.

References

Hanratty

, Zhang

, Whitehead

. How close have universal health systems come to achieving equity in use of curative services? A systematic review. Int J Health Serv 2007;37:89–109

Schoen

, Osborn

, Huynh

, Primary care and health system performance: adults' experiences in five countries. Health Aff (Millwood) 2004(Suppl):W4-487-503

Van Doorslaer

, Koolman

, Jones

. Explaining income-related inequalities in doctor utilization in Europe. Health Econ 2004;13:629–47

Mossialos

, Allin

, Davaki

. Analysing the Greek health system: a tale of fragmentation and inertia. Health Econ 2005;14 (S1):151–68

Tountas

, Karnaki

, Pavi

, Souliotis

. The “unexpected” growth of the private health sector in Greece. Health Policy 2005;74:167–80

Oikonomou

, Mariolis

. Three trends that undermine the Greek health system: is there a way out? Eur J Gen Pract 2009;15:67–8

Mossialos

. Citizens' views on health care systems in the 15 member states of the European Union. Health Econ 1997;6:109–16

Tountas

, Karnaki

, Pavi

. Reforming the reform: the Greek national health system in transition. Health Policy 2002;62:15–29

Niakas

, Gnardellis

, Theodorou

. Is there a problem with quality in the Greek hospital sector? Preliminary results from a patient satisfaction survey. Health Serv Manage Res 2004;17:62–9

10.

Pappa

, Niakas

. Assessment of health care needs and utilization in a mixed public–private system: the case of the Athens area. BMC Health Serv Res 2006;6:146

11.

Geitona

, Zavras

, Kyriopoulos

. Determinants of healthcare utilization in Greece: implications for decision-making. Eur J Gen Pract 2007;13:144–50

12.

Esomar. Standard demographic segmentation, a system of international socio-economic classification of respondents to survey research, 1994. See www.esomar.org/index.php/glossary-s.html (last accessed 23 August 2010)

13.

Ware

, Kosinski

, Keller

. SF-36 Physical and Mental Health Summary Scales: A User's Manual. Boston, MA: The Health Institute, 1994

14.

Pappa

, Kontodimopoulos

, Niakas

. Validating and norming of the Greek SF-36 Health Survey. Qual Life Res 2005;14:1433–8

15.

Mergoupis

. Access to health care in Greece: a micro-level analysis. Proceedings of the First Conference of the Hellenic Social Policy Association, Komotini, 2001

16.

Nolan

. Economic incentives, health status and health services utilization. J Health Econ 1993;12:151–69

17.

Boniface

, Denham

. Factors influencing the use of community health and social services by those aged 65 and over. Health Soc Care Communiy 1997;5:48–54

18.

Peacock

, Devlin

, McGee

. The horizontal equity of care in New Zealand. Aust NZJ Public Health 1999;23:126–30

19.

Young

, Dobson

, Byles

. Determinants of general practitioner use among women in Australia. Soc Sci Med 2001;53:1641–51

20.

Morris

, Sutton

, Gravelle

. Inequity and inequality in the use of health care in England: an empirical investigation. Soc Sci Med 2005;60:1251–66

21.

Van Doorsaler

, Masseria

, Koolman

. Inequalities in access to medical care by income in developed countries. CMAJ 2006;174:177–83

22.

Truswell

. Family physicians and patients: is effective nutrition interaction possible? Am J Clin Nutr 2000;71:6–12

23.

Schneider

, Bouvier Gallacchi

, Goehring

, Kunzi

, Bovier

. Personal use of medical care and drugs among Swiss primary care physicians. Swiss Med Wkly 2007;137:121–6

24.

Mendoza-Sassi

, Béria

. Prevalence of having a regular doctor, associated factors, and the effect on health services utilization: a population-based study in SouthernBrazil. Cadernos de Saúde Pública 2003;19:1257–66

25.

Borrell

, Fernandez

, Schiaffino

Social class inequalities in the use of and access to health services in Catalonia, Spain: what is the influence of supplemental private health insurance?

Int J Qual Health Care 2001;13:117–25

26.

Van der Heyden

, Demarest

, Tafforeau

, Van Oyen

. Socio-economic differences in the utilization of health services in Belgium. Health Policy 2003;65:153–65

27.

Habicht

, Kunst

. Social inequalities in health care services utilization after eight years of health care reforms: a cross-sectional study of Estonia, 1999. Social Sci Med 2005;60:777–87

28.

Souliotis

, Lionis

. Creating an integrated health care system in Greece: a primary care perspective. J Med Syst 2005;29:187–96

29.

Atella

, Brindisi

, Deb

, Rosati

. Determinants of access to physician services in Italy: a latent class seemingly unrelated probit approach. Health Econ 2004;13:657–68

30.

Van Doorslaer

, Masseria

. Income-Related Inequality in the Use of Medical Care in 21 OECD Countries. No 14, OECD Health Working Papers. OECD Directorate for Employment, Labour and Social Affairs, 2004

31.

Van Doorslaer

, Koolman

, Puffer

. Equity in the use of Physician Visits in OECD Countries: has Equal Treatment for Equal Need been Achieved? In Measuring Up: Improving Health Systems Performance in OECD Countries. Paris: OECD, 2002

32.

Starfield

. Primary Care: Balancing Health Needs, Services and Technology. New York: Oxford University Press, 1998

33.

Bertakis

, Azaris

, Helms

, Callahan

, Robbins

. Gender differences in the utilization of health services. J Fam Pract 2000;49:147–52

34.

Wagner

, Guendelman

. Healthcare utilization among hispanics: findings from the 1994 minority health survey. Am J Manag Care 2000;6:355–64

35.

Redondo-Sendino

, Guallar-Castillón

, Banegas

, Rodríguez-Artalejo

. Gender differences in the utilization of health-care services among the older adult population of Spain. BMC Public Health 2006;6:155

36.

Rodríguez

, Stoyanova

. The effect of private insurance access on the choice of GP/specialist and public/private provider in Spain. Health Econ 2004;13:689–703

37.

Siskou

, Kaitelidou

, Papakonstantinou

, Liaropoulos

. Private health expenditure in the Greek health care system: where truth ends and the myth begins. Health Policy, 2008;88:282–93, doi: 10.1016/j.healthpol.2008.03.016

38.

Liaropoulos

, Siskou

, Kaitelidou

, Theodorou

, Katostaras

. Informal payments in public hospitals in Greece. Health Policy, 2008;87:72–81, doi: 10.1016/j.healthpol.2007.12.005

39.

Abel-Smith

, Calltorp

, Dixon

, Report on the Greek Health Services. Ministry of Health and Social Welfare of Greece; Pharmetrica, Athens, 1994

40.

Hakkinen

. Change in determinants of use of physician services in Finland between 1987 and 1996. Soc Sci Med 2002;55:1523–37

41.

Propper

, Eachus

, Chan

, Pearson

, Smith

. Access to health care resources in the UK: the case of care for arthritis. Health Econ 2005;14:391–406

42.

Ovaskainen

, Rautana

, Ojantlava

, Pakkila

, Paivarinta

. Analysis of primary health care utilization in south-western Finland-a tool for management. Health Policy 2003;66:229–38

43.

Mastilica

, Božikov

. Out-of-pocket payments for health care in croatia: implications for equity. Croat Med J 1999;40:152–9

44.

Iversen

, Kopperud

. Regulation versus practice – the impact of accessibility on the use of specialist health care in Norway. Health Econ 2005;14:1231–8

45.

OECD Health Data. 2006. See www.oecd.org/health/healthdata (last accessed 23 August 2010)