# copy the linklink copied!3. Inequalities in the utilisation of health care services

This chapter turns to the question of whether health systems treat people with comparable needs equally irrespective of their income. It measures income-related inequalities in health care services utilisation, adjusted for needs where relevant, based on national health survey data for 33 EU and OECD countries carried out between 2014 and 2017. It investigates inequalities in doctor visits, hospital admissions, as well as preventive care such as cancer screening, flu vaccination, and dental care. Summary measures of inequality are derived to compare results across the various health care services and countries.

Note by Turkey: The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union: The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

One of the pathways to improving health and overcoming inequalities is for health systems to ensure access to quality services irrespective of people’s socio-economic circumstances. Measuring within and across health systems whether the care received by patients is commensurate to need and of quality presents considerable methodological challenges. As people typically turn to the health system when sick, the volume of services used gives an indication of their ability to access it. However, a more important question is whether people’s utilisation is commensurate to their needs. One strategy to answer this question consists in examining whether, for a given health status – a proxy of need – people have a comparable level of utilisation of the health system.

Recognising that no measure of access is perfect, the main question of interest in this chapter is whether systematic differences across socio-economic groups in the level of utilisation of care can be detected within EU and OECD countries. Using micro-level data, Section 3.2 measures the extent to which, within countries, access to health care services (physician and hospital care) varies across people with comparable needs but different income. Section 3.3 explores whether utilisation of preventive services – such as cancer screening, dental care and vaccination- differs across income groups. Section 3.4 concludes with key findings and presents a summary measure of inequalities in health care services utilisation by clustering countries into groups that display comparable levels of inequalities in service utilisation.

The following two chapters (Chapters 4 and 5) will look into complementary aspects of access to care, namely whether inequalities across socio-economic groups exist for unmet needs and the financial protection against the costs of care. Together, these three chapters allow for a comprehensive analysis of the extent to which health systems treat all patient equally regardless of income.

## copy the linklink copied!3.2. Income-related inequalities exist in the utilisation of some – but not all – health care services across EU and OECD countries

Determining whether access to health care services is similarly distributed within a population irrespective of income requires factoring in health care needs. As shown in Chapter 2, poor health is concentrated among the least well-off in all countries. Consequently, a simple comparison of utilisation across income groups could obscure or underestimate inequalities. Using individual-level data, this section therefore seeks to determine whether, once need is taken into account, within-country patterns of access to care are comparable across income levels, according to a well-established methodology.

More specifically, the analysis reviews patterns of access to GP and specialist consultations as well as hospital admissions. National health survey data from 33 EU and OECD countries are used for this analysis (the 2014 European Health Interview Survey wave 2 for European countries, the Canadian Community Health Survey 2015-16, the Chilean National Socio-Economic Characterization Survey 2017, and the US Medical Expenditure Panel Survey 2016) (see Box 2.2 in Chapter 2 for a brief description). Table 3.1 provides more information about the variables used and the method is detailed in Box 3.1, which also provides a numerical example of the difference between simple and needs-adjusted probabilities highlighting why such an adjustment is necessary.

Table 3.1. Variables used and coverage of the population to analyse differences in the utilisation of health care services

Description of variables and coverage of the population

Dependent variables

Doctor visit: Dummy variable describing whether people have visited a doctor in the past 12 months or not (in EHIS this variable is constructed by combining responses on GP and specialists visits).

GP visit: Dummy variable describing whether people have visited a GP in the past 12 months or not.

Frequency of GP visits: Number of GP visits in the past 4 weeks (for only those who visited a GP in the past 12 months)

Specialist visit: Dummy variable describing whether people have visited a specialist in the past 12 months or not.

Frequency of specialist visits: Number of specialist visits in the past 4 weeks (for only those who visited a specialist in the past 12 months)

Hospitalisation: Dummy variable identifying people who had an inpatient hospital admission in the past 12 months, versus those who did not.

Cervical cancer screening: Dummy variable identifying women aged 20-69 who had a Pap Smear test in the past 3 years versus those who did not.

Breast cancer screening: Dummy variable identifying women aged 50-69 who had a mammography in the past 2 years versus those who did not.

Colorectal cancer screening: Dummy variable identifying persons aged 50-74 who had a Faecal Occult Blood Test in the past 2 years or a colonoscopy in the past 10 years versus those who did not.

Dental visit: Dummy variable identifying people who had a dentist visit in the past 12 months versus those who did not.

Flu vaccination: Dummy variable to identify people aged 65 and over who had flu vaccination in the past 12 months versus those who did not.

Explanatory variables

Income level: Categorical variable with income quintiles from the lowest income Q1 to the highest income Q5 (accounting for household size).

Needs variables

Gender, Age group

Self-assessed health status: Categorical variable with very poor, poor, fair, good or very good health status.

Limitations in daily activities: Categorical variable with severely limited, limited, or not limited.

Countries included

30 European countries (Belgium did not enquire about the number of physician visits in the previous month), Canada, Chile, and the United States.

Covered population

The age was restricted to the population over 18 years of age, unless otherwise mentioned for preventive care.

Note: For general information about national surveys used, please refer to Box 2.2 in Chapter 2. For descriptive statistics, see Annex Table 3.A.1 and Annex Table 3.A.8-9. The description of variables in the above table refers to EHIS used for 30 of the 33 countries studied. The main differences with other national surveys were as follows: the United States survey does not enquire separately about GP and specialists. In Chile, the recall period of physician and dentist visits is 3 months for the probability and the number of visits. In Canada, the recall period for the number of visits is 12 months.

Source: Authors.

Box 3.1. Estimating need-adjusted utilisation across income groups: Methodology

The method used in this chapter to measure inequalities in health care services utilisation is well established (van Doorslaer and Masseria, 2004[1]; OECD, 2003[2]; O’Donnell et al., 2008[3]).

## Indirect standardisation for health care needs

Visits to doctors, GPs and specialists, and inpatient hospital admissions are standardised for health care needs, while the probabilities of using preventive care (cancer screening, dental care and flu vaccination) are not, as explained in Section 3.3.

The need for medical care is proxied by age, gender, and health status variables (self-assessed health and activity limitation).

An indirect standardisation is used to predict the probability of use of medical services, adjusted for health care needs in a series of country-specific regressions (O’Donnell et al., 2008[3]). A logistic regression model is used to estimate the probability of the use of each type of health care services and a linear model to estimate the frequency of visit to GPs and specialists. The health regression estimates the following:

[1]

where Y denotes the dependent variable (e.g. doctor visits of individual in a given period), X a set of need indicator variables including demographic and morbidity variables (age, gender, self-assessed health and activity limitation), and Z a set of non-need control variables (education, marital status, occupational status, income, size of household, urbanisation level, variables used to control for, in order to estimate partial correlations with the need variables), α, β and δ are parameters vectors, and ε an error term.

Estimations are produced for each country. Equation 1 can be used to generate need-predicted, or X-expected, values of Y. ${\mathbit{Y}}_{\mathbit{i}}^{\mathbit{X}}$ represents the amount of medical care an individual i would have received if she/he had been treated as others with the same need characteristics, on average:

[2]

where $\stackrel{-}{\mathbit{Z}}$ refers to the sample mean values.

Estimates of the indirectly needs-standardised utilisation, ${\mathbit{Y}}_{\mathbit{i}}^{\mathbit{I}\mathbit{S}}$, are then obtained as the difference between actual and x-expected utilisation, plus the sample mean $\stackrel{-}{\mathbit{Y}}$ :

[3]

For each level of income, the value of ${\mathbit{Y}}_{\mathbit{i}}^{\mathbit{I}\mathbit{S}}$ can be interpreted as the level of health care utilisation one would find if needs were equally distributed across income groups.

## Concentration index

The concentration index (CI) of health care utilisation measures the degree of inequality across the income distribution. The concentration index of a variable Y can be computed using a simple “convenient covariance” formula:

[4]

where μ is the weighted sample mean of Y, covw denotes the weighted covariance and Ri is the (representatively positioned) relative fractional rank of the ith individual in the income distribution. The Stata command concindc is used to calculate the CI and its confidence interval (Chen, 2007[4]).

If the concentration index is significantly above (or below) 0, high-income people are more (or less) likely to access medical care services than low-income people. If the 95% confidence intervals of the concentration cross the 0 line, there is no significant inequality.

For preventive services and for unmet needs variables presented in the next chapter, all binary variables, the generalised concentration index (GCI) is used in order to measure absolute inequalities taking into account the overall level of the variable of interest (Yi). The GCI is derived from the standard concentration index by multiplying it with the mean of Yi. As a result, for instance, if two countries have the same level of relative inequality in cancer screening (measured by the CI), the inequality between rich and poor will be deemed higher in the country with the higher prevalence. Using an analogy with the RII and SII discussion in earlier chapters, a similar ratio between the prevalence of the low and high income groups translates into larger absolute differences between these groups when the average is higher. The GCI captures absolute inequalities and also leads to the same ranking of countries irrespective of whether the inequality in having received the service or not having received the service is measured. The Stata command conindex is used to calculate the GCI and its confidence interval (O’Donnell et al., 2016[5]).

## The importance of adjusting for need: a numerical example

The example below illustrates the effect of the needs-standardisation procedure. Figure 3.1 shows the proportion of the population visiting a specialist in Estonia and the calculated probability after standardisation for health care needs. The observed probabilities of a low or high income person visiting a specialist are virtually identical. However, once the differences in health care needs are taken into account, the probability of a visit increases by 8 percentage points in the highest quintile (Q5), and reduces by 3 percentage point in the lowest quintile (Q1). So while high and low-income people see specialists as frequently, the latter’s health status is worse on average. The standardisation erases this difference and shows that, with equal health, a person in the top income quintile has a 12% higher likelihood to see a specialist than one in the bottom quintile. As a result, the concentration index (CI) -which was small and not significant before standardisation- becomes larger, positive and significant after needs-standardisation indicating a distribution in favour of the better-off.

### 3.2.1. In most countries, for a given level of needs, access to the doctor increases with income level but less so for GPs

#### Higher income translates into a higher needs-adjusted probability of seeing a doctor

In the vast majority of countries, for a comparable level of needs, low-income people are less likely to have seen a physician in the past 12 months than high-income people. The varying levels of predicted probabilities across countries confirm that differences in utilisation patterns remain large even when need is taken into account. The need-standardised probability of reporting a doctor’s visit in the past 12 months in the total adult population ranges from 45% in Romania to 89% in France (Chile’s much lower average corresponds to a 3 months reporting period). More importantly, the probability of seeing a doctor, for a given level of needs, is higher for those in the highest quintile than in the lowest in all countries except in Denmark, and Slovak Republic (Figure 3.2).

Overall, statistically significant pro-rich inequalities in access to physicians exist in three quarters of EU and OECD countries. The concentration index provides a summary measure of inequalities across the entire population in a country (Figure 3.3). The income-related gradient of inequality is positive and significant in 25 out of 33 countries. In these countries, the higher people’s income is, the more likely they are to visit a doctor for the same level of need. However, the gradient is reversed in Denmark. Only seven countries provide the same level of access to their population irrespective of income (the Slovak Republic, Malta, Sweden, Ireland, the United Kingdom, the Netherlands and Luxembourg) as differences are not statistically significant.

#### Once access to a GP is secured, low-income patients have at least as many if not more visits to the GP than the rich in all but one country

Access to primary care services varies by country, and more importantly across income levels. Taking into account differences in needs, the probability of having visited a General Practitioner (GP) in the last 12 months among the total adult population varies across countries, from 14% in Cyprus to 86% in France. It also varies with income: 67% of people with lower-income have seen a GP in the past 12 months compared to 72% in the higher-income group, on average across EU and OECD countries (Figure 3.4, Panel A). The difference between the needs-adjusted probabilities of seeing a GP is more than 10 percentage points in seven countries (Bulgaria, Croatia, Finland, Greece, Latvia, Poland and Romania). The overall gradient of inequality (which takes into account the full distribution of GP visits across income levels, measured by the concentration index) is positive and significant in 17 out of 32 countries. In these countries, the higher people’s income is, the more likely they are to visit a GP for the same level of needs. In contrast, Denmark and Spain display inequalities in favour of people with lower income. The 13 remaining countries show no significant inequalities in the probability of having a GP visit across income levels (see concentration index in Annex Table 3.A.2).

In most countries, once they get a first contact with a GP, the poor have the same number of GP visits as the rich, after health care needs are taken into account. Figure 3.4 (Panel B) presents the frequency of GP visits in the past four weeks for people who had a first contact with a GP during the year. In nine countries, for a given level of need, the number of visits to the GP is disproportionally concentrated among those with lower income (the Czech Republic, Denmark, France, Germany, Italy, Portugal, Slovenia, Spain and the United Kingdom) (Annex Table 3.A.3). Conversely, in Romania, the higher the income of people, the greater their number of GP consultations. No significant inequalities in the number of GP visits across income levels can be observed in the remaining 21 countries.

Results on the probability of a medical visit and the conditional number of visits provide complementary information. It is interesting to differentiate the probability of at least one visit in the past year and the number of visits (conditional on the first medical contact), under the assumption that the decision of initiating use is more patient-driven and the decision about continued use more doctor-driven as suggested by Van Doorslaer and Masseria (2004[1]). Results in that study show that the patterns are not identical for these two parts of the utilisation. Overall, in the majority of countries, richer patients are more likely to turn to the system and seek out the services of a physician. On the other hand, the fairly equal distribution of the number of visits across income groups suggests that (if the doctor-driven decision hypothesis was confirmed) the doctor see richer patients as often as poorer patients based on their medical need (and not based on financial conditions). However, this interpretation is only tentative1.

#### The higher people’s income, the more likely they are to see a specialist, for a given level of needs, in 29 out of 32 countries

The probability of visiting a specialist varies across countries, and more importantly, there are inequalities across income levels within countries. The needs-adjusted probability of having visited a specialist in the last 12 months among the total adult population ranges between 18% in Romania and 64% in Germany. In the vast majority of countries, for the same level of needs, the better-off have a higher probability of visiting a specialist than lower-income people (Figure 3.5, Panel A). The overall gradient of inequalities (measured by the concentration index) confirms this finding, and shows significant pro-rich inequalities in 29 countries. Only in three countries (Denmark, Ireland, and the Slovak Republic), access to specialist care is effectively irrespective of people’s income level (see concentration index in Annex Table 3.A.4 and Annex Figure 3.A.3). In comparison with GP visits, the probability of a specialist visit shows much larger degrees of income-related inequality in most countries.

Once people get access to specialist care, the poor and the rich have –for the same level of need– the same number of specialist visits in the majority of countries. Figure 3.5 (Panel B) presents the number of specialist visits in the past four weeks for people who had a first contact with a specialist in the past year, once health care needs are taken into account. At population level, in 18 countries, the number of specialists visits is not linked to income once need is factored in. Eleven countries show pro-rich inequalities, where the higher the income, the greater the number of specialist visits for the same level of need. In contrast, in Bulgaria and Sweden the inequality pattern is in favour of the poor (Annex Table 3.A.5, Annex Table 3.A.4).

### 3.2.2. Access to hospital services does not depend on income in most countries

The frequency of hospital admissions also varies considerably across countries. On average, one in ten adults in EU and OECD countries were hospitalised in the 12 months prior to the survey. The needs-standardised probability of having an inpatient hospital admission in the past 12 months ranges between 4% in Romania and 16% in Ireland (Annex Table 3.A.7).

The majority of countries provide equal access to hospital services irrespective of income. On average across EU and OECD countries, the probability of hospital admissions are almost identical for the highest and lowest income quintiles. A handful of countries display fairly large gaps in the predicted probability of admission between the lowest and highest income quintile, which in most cases suggests that people belonging to the lowest income group are more frequently hospitalised. However, across the entire income distribution, the overall gradients of inequality (measured with the concentration index presented in Annex Table 3.A.7) show no income-related differences in hospitalisations in 25 out of 33 countries. In six countries (Canada, Estonia, Luxembourg, the United States, Latvia and Germany) however, the probability of being hospitalised decreases when income rises. Conversely, the better-off are more likely to be admitted into a hospital in Romania and Italy.

### 3.2.3. Summary of inequalities in utilisation of curative services

Overall, once adjusted for health care needs, the differences in probabilities of having a doctor, GP or specialist visit for various income groups point to inequalities in access to care in favour of high-income people in most EU and OECD countries although the pattern is less frequent for GP visits (Table 3.2). When it comes to the number of visits to a GP or a specialist and the probability of having a hospital admission, differences between income groups are much less pronounced. For the number of GP consultations, there are more countries where the inequalities in utilisation appear to be pro-poor than countries where the opposite is the case.

Table 3.2. Summary of inequalities in doctor visits and hospitalisations

Once need is taken into account

Increases as your income becomes higher

Does not differ across income groups

Decreases as your income becomes higher

Probability of doctor visit… ➜

CZE, DEU, ISL, BEL, NOR, FRA, ESP, HUN, AUT, LTU, ITA, SVN, CAN, PRT, HRV, GRC, EST, POL, LVA, CYP, FIN, USA, BGR, ROU, CHL

SVK, MLT, SWE, IRL, GBR, NLD, LUX

DNK

Probability of visit to the GP ➜

FRA, AUT, NOR, BEL, ITA, LTU, SVN, HRV, EST, CAN, LVA, POL, GRC, CHL, FIN, CYP, BGR, ROU

SWE, SVK, MLT, NLD, DEU, CZE, PRT, GBR, HUN, IRL, LUX, ISL

ESP, DNK

Number of visits to the GP ➜

ROU

LVA, LUX, BGR, SWE, LTU, FIN, HUN, IRL, CHL, NLD, POL, ISL, CYP, AUT, EST, SVK, MLT, GRC, NOR, HRV, CAN

PRT, FRA, CZE, ESP, ITA, DEU, GBR, DNK, SVN

Probability of visits to a specialist ➜

BEL, NLD, LUX, DEU, HUN, LTU, AUT, CZE, CYP, EST, GRC, GBR, NOR, MLT, ISL, CAN, FRA, ITA, SWE, SVN, LVA, ESP, POL, HRV, FIN, PRT, ROU, BGR, CHL

DNK, IRL, SVK

Number of visits to a specialist ➜

AUT, POL, SVK, NLD, IRL, GBR, FIN, ROU, PRT, LUX, CHL

EST, HRV, ISL, LTU, SVN, MLT, DEU, FRA, DNK, HUN, ITA, CZE, GRC, CYP, ESP, LVA, NOR, CAN

SWE, BGR

Probability of hospitalisation ➜

ITA, ROU

DNK, LTU, BEL, ISL, SVN, PRT, GBR, FRA, CHL, MLT, FIN, IRL, BGR, NOR, HUN, SWE, AUT, POL, CZE, NLD, CYP, ESP, SVK, HRV, GRC

DEU, LVA, USA, LUX, EST, CAN

Note: Countries are ranked from lowest to highest degree of inequality (based on the Concentration Index).

Source: OECD estimates based on national surveys data.

The findings presented here are generally consistent with previous studies on inequalities in health care utilisation. Previous work by the ECuity project group, which mostly covered Western and Southern Europe, generally found no difference by income level in the probability of a GP visit, for the same level of needs, whereas in the present study the probability of a GP visit disproportionally favours the better off in more than half of the countries. But, aligned with the results of this chapter, the ECuity project group also found that when they started to see a GP, low-income people had more GP visits than high-income people in some countries. High-income people were more likely to see a specialist than low-income people (for the same level of health care needs), and they were often visiting these specialists more (Doorslaer, Koolman and Jones, 2004[6]; van Doorslaer and Masseria, 2004[1]). Two other studies of European countries also showed pro-rich inequalities in specialist visits once standardised for needs, whereas the picture for GP visits was less clear-cut with some evidence for pro-poor inequalities (Or, Jusot and Yilmaz, 2008[7]; Bago d’Uva, Jones and van Doorslaer, 2009[8]). A range of studies in Latin America found more systematic evidence of a distribution of utilisation favouring the better-off. In Mexico, higher-income people used more curative and hospital care then lower-income people both in 2000 and 2006, with a slight decrease in inequalities in curative care over time (Barraza-Lloréns, Panopoulou and Díaz, 2013[9]). A study on Colombia and Brazil showed that 20 years after the introduction of reforms to improve access, income-related inequalities persisted in both countries (for specialist care in Brazil, and for primary, secondary and emergency care in Colombia) (Garcia-Subirats et al., 2014[10]). In Chile, once adjusted for needs, high-income people had a higher probability and intensity of use of GPs and specialists. Low-income people used emergency care more frequently, partly because of the cost associated with GP and specialist visits. For hospitalisation, high-income people tended to use the system more frequently in Chile, but low-income people stayed longer once they used it. This suggests that people with low income are hospitalised at a later stage, when the medical condition is already more critical. These patterns of inequality appeared to have become more pronounced over time (Vásquez, Paraje and Estay, 2013[11]).

The results presented here suggest that inequalities in the probability of a doctor visit have remained stable since the latest OECD study (Devaux and de Looper, 2012[12]). Comparing 16 countries that are included in both analyses, results show that, for GP visits, the degree of inequality remains constant for 13 countries. Only Finland displays increasing inequalities2, while they decreased in Denmark and the Slovak Republic. Regarding the probability of a specialist visit, most countries show no change in the degree of inequality over time. However, inequalities decreased in four countries (Belgium, Canada, Demark, and France), whereas Finland, Slovenia, and the United Kingdom display rising inequalities. The present analysis adds 17 new countries to the 2012 study, of which 14 show pro-rich inequalities in GP visits. When it comes to specialist visits, data for all new countries display pro-rich inequalities.

There are various explanations for income-related inequalities in health care services utilisation. First, these inequalities can be driven by financial barriers to health care access, in particular in countries where the depth, breadth and height of coverage are low, or where private health care services play an important role in the health care system (see Chapter 5). Although most EU and OECD countries have achieved universal coverage for at least a core set of services, there are caveats (such as partial set of services covered, role of private health insurance, high cost-sharing) that makes access dependent on income. Second, health literacy and information about health care, such as awareness about availability and efficacy of health care services, may vary across population groups (Goddard and Smith, 2001[13]). People with a better understanding of the health care options and pathways can navigate the health care system more easily. Third, availability and quality of care may contribute to these inequalities as well (Goddard and Smith, 2001[13]). Availability of health care services may vary across population groups, or clinicians may have different propensities to offer treatment to patients from different population groups, even where they have identical needs.

## copy the linklink copied!3.3. Lower income people use preventive services less frequently

Few would argue against providing preventive and screening services (in relevant target groups) to detect the early onset of diseases to patients irrespective of income. Moreover, given again the higher burden of diseases of disadvantaged people, providing universal access to these services, especially cancer screening, is particularly important for health systems to effectively reduce inequalities in health outcomes.

In most countries, national authorities determine who should have access to screening services (see Box 3.2). They also have dedicated programmes to raise awareness about them. Service use may be free of cost to the people in the target group and service providers may even be incentivised to ensure their patients receive them. Immunisation against seasonal flu for the elderly and regular dental check-ups are slightly different in nature. Although they may be desirable and are certainly seen as preventive services, they may not be as actively promoted or available free of charge.

This section examines inequalities in the take-up of preventive services across income groups using generalised concentration indices as described in Box 3.1.

Box 3.2. Recommendations for preventive care

Preventive care such as cancer screening, annual dental check-ups, and flu vaccination are recommended by national health authorities to all people in particular target age groups. National recommendations may vary across countries, but the following rule applies in a number of countries where, most often, cancer screening programmes are delivered free of charge.

• For cervical cancer screening, Pap Smear tests are recommended every 3 years in women aged 20-69.

• For breast cancer screening, mammography is recommended every 2 years in women aged 50-69.

• For colorectal cancer screening, Faecal Occult Blood Tests are recommended every 2 years in adults aged 50-74.

• Flu vaccination is recommended every year for people aged over 65.

• Dental examination is recommended once a year for the general population but more frequently among high-risk individuals (Kay, 1999[14]; ADA, 2013[15]).

As preventive services are mostly targeted to specific populations determined by their age and sex and take up should not depend on health status, inequalities in this section are measured on observed value (and not standardised as for curative services analysed in the previous section).

Although most countries have adopted population-based cancer screening programmes as an effective way for detecting diseases early, some questions emerge about the effectiveness of increasing coverage of mammography screening. This debate is related to recent progress in treatment outcomes and concerns about false-positive results, over-diagnosis and overtreatment. In high-income countries, WHO now recommends organised population-based mammography screening for women aged between 50 and 69, if specific criteria are met such as whether women are able to make an informed decision based on the benefits and risks of mammography screening (WHO, 2014[16]). Hence, increasing coverage of breast cancer screening is not by itself an objective for health care systems.

### 3.3.1. In virtually all countries, cancer screening is less frequently availed by people with lower income

There are large variations in cancer screening rates across EU and OECD countries and across different types of cancer. The national health surveys inquire whether people in the target groups have received cervical, breast or colorectal cancer screenings. The data on screening rates show:

• Colorectal cancer screening is less common than other types of cancer screening. On average in EU and OECD countries, 71% of targeted women were screened for cervical cancer and 66% for breast cancer, while only 38% people in the target group underwent colorectal cancer screening in the recommended time period.

• The variations in cancer screening rates across countries are large. The rates of cervical cancer screening range from 27% in Romania to 87% in the Czech Republic. For breast cancer, they vary from 7% in Romania to 91% in Sweden, and they stand between 6% in Bulgaria and 74% in Germany for colorectal cancer.

• Screening rates are generally correlated within countries but in some countries coverage performance depends on the type of cancer. Romania, Bulgaria, Estonia, Latvia, Cyprus, Malta and Poland are characterised by fairly low cancer screening rates across the three types of cancer. The opposite is true for France, Austria, Germany, the United States, Luxembourg and the Czech Republic. There are no clear patterns in other countries: Finland has one of the highest screening rate for breast and cervical cancer, but a low one for colon cancer. On the other hand, Denmark’s screening rate is much higher than average for colon and breast cancer but below average for cervical cancer.

Overall, across EU and OECD countries, the less well-off have a lower probability of screening for all three types of cancer. For instance, only 61% of poor women had cervical cancer screening compared to 78% of women with high income. Figure 3.7 presents the rate of cervical cancer screening, showing large income-related inequalities in screening uptake in many countries.

When comparing the richest and poorest income quintiles, only few countries manage to ensure that high- and low-income groups have similar access to cancer screening. The distribution of cervical cancer screening is biased towards higher-income women who are more likely to have cervical cancer screening than lower-income women in virtually all countries. Only in Ireland is the uptake of cervical cancer screening similar across income groups. Comparable patterns of inequality are found in breast and colorectal cancer screening, with richer people having greater access to screening services in the vast majority of countries.

Taking account of the distribution across the entire population, pro-rich inequalities in access to cancer screening exist in the vast majority of EU and OECD countries. Inequalities in favour of people with higher income, measured across the entire population using generalised concentration indices, are significant in 31 countries for cervical cancer screening, and in 26 countries for breast cancer (out of 33). Colorectal cancer screening is somewhat less unequally distributed: the GCIs are positive and significant in 18 countries (out of 32), not significant in 12 and the distribution favours lower income groups in Sweden and the United Kingdom (Annex Table 3.A.8, Annex Table 3.A.9-11). Overall, the less well-off display a lower cancer screening uptake, despite the fact that these programmes are organised and provided at no cost in the majority of countries (Ponti A et al., 2017[17]). Limitations in health literacy among people with lower income may partly explain these inequalities. However, countries could also investigate the design of these programmes; whether the existing communication strategies are effective or whether they need to be tailored to populations groups with different socio-economic backgrounds.

There are no discernible patterns of inequalities in cancer screening, suggesting that efforts to improve utilisation need to be designed on a country-by-country and cancer-by-cancer basis. The correlation between countries’ screening rates and inequalities levels for any given cancer is weak. In other words, for a given screening rate, inequalities can be high or low in two different countries3. For instance, Denmark and the United States have comparable and relatively high levels of breast cancer screening (82% and 80%) but in the former country, inequalities are very low while the United States belong to the high inequality group. In Spain and Sweden, colorectal cancer screening is mediocre (around 25% for an average across countries of 38%), but the respective levels of inequalities are high in Spain and low in Sweden. Additionally, inequalities in coverage vary across types of cancer within countries. In Greece, for instance, inequalities are high for breast cancer screening, low for cervical and intermediate for colorectal cancers. In Portugal, they are respectively low (breast), intermediate (cervical) and high (colorectal). Overall, for all three types of cancer screening, Cyprus and Hungary display very high inequalities while they are very low in the Netherlands and Ireland.

### 3.3.2. All countries, except Ireland, show inequalities in dentist visits in favour of people with higher income

The proportion of people who see a dentist annually varies by country, and across socio-economic groups within countries. On average, 60% of adults in EU and OECD countries had at least one dental consultation in the 12 months prior to the survey, with this share ranging from 15% in Romania to 93% in Ireland (Figure 3.8). Inequalities across income levels are apparent in all countries and the average difference between low and high income groups is close to 20 percentage points.

Income-related inequalities for dental consultations exist in all but one country and they favour the rich. With the exception of Ireland, the overall gradient of inequality taking into account the full distribution of dental visits across incomes groups (measured with the generalised concentration index) is significant in all countries (Annex Table 3.A.9 and Annex Table 3.A.7). Inequalities are also generally higher than for other preventive services.

These high levels of inequalities may be partially related to the benefit-design of collectively-financed health care goods and services. Unlike hospital care or visits to GP and specialists, the costs of dental care are much less well protected in many EU and OECD countries (see Chapter 5). As a result, poorer people may not be able to afford these services.

### 3.3.3. Flu vaccination among the elderly seems more evenly distributed among income groups than other preventives services

The vaccination rates against the flu among people aged 65 and over vary greatly across countries. On average, 39% of the elderly were vaccinated against the flu in the year preceding the survey across 30 EU and OECD countries (Figure 3.9). The variation between countries was more than 90-fold: this share stood as high as 91% in Finland but was as low as 1% in Estonia.

Immunisation against the flu does appear to be more evenly distributed across income levels than other preventive services. Compared to those for other preventive services Figure 3.9 displays relatively small gaps in vaccination rates between the higher and lower income groups, however, due to the relatively small sample size the probabilities of being immunised represented on the chart are those of the two lowest and two highest income quintiles. The GCIs are generally lower for flu immunisation than other preventive services and they are only significant in 9 out of 30 countries. However, this result is also likely to be partly driven by the relatively small number of people 65 and over in the surveys (which leads to large confidence intervals pictured in Annex Table 3.A.8). The distribution is significantly biased towards the better off in 8 countries while in the Netherlands immunisation is more concentrated among the lower income groups (Annex Table 3.A.9).

### 3.3.4. Summary of inequalities in utilisation of preventive services

Overall, the observed differences in the probabilities of cancer screening, dental care and flu vaccination across income groups suggest inequalities in favour of the better-off in the vast majority of EU and OECD countries. Table 3.3 illustrates whether countries display inequalities in favour of the rich or the poor, or whether no socio-economic differences in access to preventive services can be detected.

Evidence from the health service research literature confirms the findings of mostly pro-rich inequalities in preventive care use. Socio-economic inequalities in breast and cervical cancer screening have been documented in a number of European countries (Devaux and de Looper, 2012[12]; Sirven and Or, 2011[18]). Interestingly, countries which offer nationwide population-based screening programmes have more equal access to these services, irrespective of income, compared to those countries where cancer screening happens in an opportunistic (and less organised) manner (Palencia et al., 2010[19]). Higher use of dental care utilisation with rising income is also confirmed in a number of studies (Palència et al., 2014[20]; Devaux and de Looper, 2012[12]; Listl, 2011[21]; Tchicaya and Lorentz, 2014[22]).

Table 3.3. Summary of inequalities in cancer screening, dental care and vaccination

Looking at preventive care

Increases as your income becomes higher

No significant difference across income groups

Decreases as your income becomes higher

Probability of cervical cancer screening… ➜

CHL, NLD, USA, CAN, AUT, SVK, GBR, GRC, DEU, BEL, PRT, SVN, LUX, LTU, CZE, HUN, LVA, FRA, MLT, FIN, ITA, EST, HRV, CYP, POL, DNK, ESP, ROU, SWE, NOR, BGR

IRL, ISL

Probability of breast cancer screening… ➜

IRL, SWE, ROU, FIN, LTU, FRA, GBR, SVN, CHL, CAN, AUT, NOR, SVK, BEL, ESP, POL, USA, CZE, LVA, ITA, HUN, GRC, MLT, HRV, BGR, CYP

DNK, LUX, EST, PRT, DEU, NLD, ISL

Probability of colorectal cancer screening… ➜

ROU, BGR, DEU, POL, GRC, AUT, LVA, CYP, CAN, LUX, PRT, ISL, FRA, HRV, ESP, SVN, ITA, USA

NOR, EST, NLD, LTU, HUN, IRL, SVK, FIN, DNK, MLT, CZE, BEL

GBR, SWE

Probability of dentist visit… ➜

CHL, DEU, SWE, LUX, DNK, LTU, AUT, FRA, ROU, GRC, CZE, NOR, GBR, SVK, EST, NLD, FIN, MLT, ISL, ITA, BEL, SVN, ESP, HRV, LVA, POL, CYP, HUN, PRT, USA, CAN, BGR

IRL

Probability of flu vaccination… ➜

LVA, ROU, GRC, HUN, CAN, USA, AUT, POL

MLT, DNK, BEL, DEU, GBR, PRT, FRA, FIN, IRL, ITA, EST, HRV, LTU, NOR, CZE, LUX, ISL, SWE, SVN, SVK, CYP

NLD

Note: Countries are ranked from lowest to highest degree of inequality using the generalised concentration index.

Source: OECD estimates based on national health survey data.

## copy the linklink copied!3.4. Synthesis and conclusion

### 3.4.1. With differences in needs factored in, the utilisation of curative and especially preventive services is generally more concentrated among high income groups

This chapter provides new evidence on the degree of income-related inequalities in health care services utilisation across 33 EU and OECD countries, based on national health survey data. Results clearly show that, consistently, but to an extent that varies across countries, people with different incomes for a given level of needs are not treated equally.

Income-related inequalities exist in the utilisation of some – but not all – curative health care services across EU and OECD countries:

• When controlling for differences in health care needs, people with lower income are less likely to visit a doctor in three quarters of EU and OECD countries.

• Inequalities in access to doctors are for a large part driven by access to specialist care: corrective for need, a person with low-income is 12 percentage points less likely than a person with high income to see a specialist. The summary measures of inequality show that the probability of a specialist visits is disproportionately concentrated among the better-off in all but three counties and the inequalities are larger than for GPs.

• Inequalities are somewhat less pronounced for access to a GP. On average, low-income people are 5 percentage point less likely to see a generalist than high-income people but the gap in the needs-adjusted probability is more than 10 percentage points in seven countries (Bulgaria, Croatia, Finland, Greece, Latvia, Poland and Romania). The summary measure of inequality shows that in a bit more than half of the countries (18 out of 32) the probability of seeing a GP in a year is more concentrated among higher income groups. The reverse is true in Denmark and Spain. In the 12 remaining countries, the summary measure of inequalities is not significant.

• Once access to a GP is secured, low-income patients have at least as many if not more visits to the GP than the rich in all but one country. Similarly, the number of visits to the specialists is equally distributed in the majority of countries.

• An intuitive – if not perfectly rigorous – interpretation of these findings is that lower-income people struggle more to reach the system but that once they have, they are in general likely to receive the same level of care as their higher-income counterparts.

• In the majority of countries, the probability of hospital admission is not associated with income levels.

Despite fairly low inequalities in access to a GP, lower-income people consistently have a lower utilisation of preventive services in virtually all countries suggesting problems in the provision of comprehensive primary care for the entire population.

• In general, access to preventive services varies greatly across countries. For instance, among the three categories of cancer screening reviewed, cervical cancer screening has the highest coverage on average (71%) - yet it ranges from 27% in Romania to 87% in the Czech Republic. Moreover, across OECD and EU countries, four in ten persons above 65 are immunised against the flu but this ratio is as low as 1% in Estonia and as high as 90% in Finland.

• For cervical, breast and colorectal cancers, the probabilities that the low-income people in the target population will have undergone screening in the recommended period is 17, 13 and 6 percentage points lower than that of the high-income people. The summary measures show that inequalities in favour of the rich prevail in the majority of countries but are slightly less systematic or marked for colorectal cancer (which has the lowest coverage rate – 38% of the target population). The data also show that for a given level of screening rate, inequalities can be high or low in two different countries. Within one country, the level of inequality in screening rates for different types of cancer can also vary considerably. Any policy aimed at improving utilisation should therefore factor in the inequality dimension to the extent it is relevant for a particular service in a given country.

• People in the lowest income quintile are nearly 20 percentage points less likely to have seen a dentist in the year. Summary inequalities are very high and detrimental to the poor in all but one country (Ireland). On the other hand, flu immunisation among the elderly is the least unequally distributed service among the preventive activities considered in this chapter, but given the small size of the samples for that population in various countries, these results would need to be confirmed.

The large differences in cancer screening rates between the rich and the poor in some countries suggest that current screening programmes and primary health care models are not succeeding in delivering recommended preventive care to the entire population. A reconfiguration of primary health care delivery towards more patient-centred models may be needed to better reach out to population groups of lower socio-economic status who frequently live in disadvantaged areas (OECD, forthcoming[23]).

### 3.4.2. Some countries are better at ensuring a more equal distribution of various types of care than others

Considering jointly the degree of income-related inequalities for all types of health care services described above, some general patterns emerge. A clustering aimed at dividing countries in three comparable-sized groups of high, low and intermediate inequalities was elaborated. It is based on the average rank of each country’s inequality index across seven services (GP visit, specialist visit, dentist visit, hospitalisation, as well as cervical, breast and colorectal cancer screenings). Countries are clustered into groups reflecting the overall level of inequalities in the utilisation of these preventive and curative services4:

• the lowest levels of inequalities are found in Denmark, Estonia, Germany, Ireland, Lithuania, Luxembourg, the Netherlands, the Slovak Republic, Sweden, and the United Kingdom.

• the highest levels of inequalities are observed in Bulgaria, Croatia, Cyprus, Finland, Greece, Italy, Latvia, Poland, Romania, Slovenia, Spain and the United States.

• The intermediate group comprises Austria, Belgium, Canada, the Czech Republic, France, Hungary, Iceland, Malta, Norway and Portugal.

### 3.4.3. Inequalities in the utilisation of care is only one aspect of the access question

This publication underlines the importance of addressing inequalities in health but this should obviously be done in conjunction with other policies aimed at improving people’s health.

To start with, in some circumstances, improving utilisation may be more of a priority than reducing inequalities. This can be particularly relevant for preventive services. It may be more desirable for a country to have high overall cancer screening rates with moderate inequality across socio-economic groups than displaying low screening rates where uptake is distributed evenly. Taking the example of preventive services, Estonia presents a good example: it belongs to the group of countries where inequalities in access to care are lowest – but at the same time, after Romania and Bulgaria and generally on par with Latvia, it has very low coverage for the preventive services analysed in this chapter. In other words, in the context of a health system performance assessment, inequalities and utilisation rates both need to be analysed jointly.

Furthermore, ensuring equality in the utilisation of services is good but not sufficient for modern, person-centred health systems. The fact that people can reach care providers is only one part of the equation: a contact with the system can only translate into better health outcomes if service provision is aligned with best practice. Similarly, if patients are harmed in the process of care, which is not rare, additional health system resources have to be mobilised and, very frequently, both the harm and the induced cost could have been avoided. In other words, for health systems to deliver results, high quality of care is paramount.

Yet, the data used to analyse inequalities in the utilisation of curative services in this chapter does not allow to assess whether the services provided have all been of good quality. Patients reporting a visit to the GP are typically not in a position to evaluate whether the treatment was in line with best practice. So it may be the case that there is an additional dimension of inequality across income groups pertaining to the quality of curative treatment. For preventive service, the situation is a bit different. Service such as cancer screening and flu vaccination are based on clinical recommendation and thus can be used to some extent to measure the quality of health systems. They contribute effectively to better health outcomes. Hence, by looking at access to preventive services, this chapter touches upon the question of health care quality. However, this approach is not comprehensive since quality of care is a multi-dimensional concept5 that is difficult to assess and country’s health systems deliver very unevenly on all dimensions. Although beyond the scope of this report, a more detailed analysis of differences in the quality of the care people received is warranted.

From a policy perspective, the results presented in this chapter open the question of what countries should do to ensure that health care services are accessible by everyone who has needs for health care, independently of the income level. At the same time, utilisation of services is only one dimension when it comes to measuring access to care and related inequalities. The following two chapters which explore differences in unmet need for care and in the financial protection for health care costs help understanding barriers in access. For example, geographical barriers to care are a reason for higher unmet need among the poor but can also lead to reduced service utilisation. The same is true if lacks in financial coverage exist for disadvantaged population groups. Taken together, these elements provide the ground for a more complete discussion of the policy options to redress inequalities in access to care.

## References

[15] ADA (2013), American Dental Association Statement on Regular Dental Visits, https://www.ada.org/en/press-room/news-releases/2013-archive/june/american-dental-association-statement-on-regular-dental-visits (accessed on 11 October 2018).

[8] Bago d’Uva, T., A. Jones and E. van Doorslaer (2009), “Measurement of horizontal inequity in health care utilisation using European panel data”, Journal of Health Economics, Vol. 28/2, pp. 280-289, http://dx.doi.org/10.1016/j.jhealeco.2008.09.008.

[9] Barraza-Lloréns, M., G. Panopoulou and B. Díaz (2013), “Income-related inequalities and inequities in health and health care utilization in Mexico, 2000 - 2006”, Revista Panamericana de Salud Pública, Vol. 33/2, pp. 122-130, http://dx.doi.org/10.1590/S1020-49892013000200007.

[24] Carinci, F. et al. (2015), “Towards actionable international comparisons of health system performance: Expert revision of the OECD framework and quality indicators”, International Journal for Quality in Health Care, http://dx.doi.org/10.1093/intqhc/mzv004.

[4] Chen, Z. (2007), CONCINDC: Stata module to calculate concentration index with both individual and grouped data, Boston College Department of Economics, https://ideas.repec.org/c/boc/bocode/s456802.html (accessed on 27 November 2018).

[12] Devaux, M. and M. de Looper (2012), “Income-Related Inequalities in Health Service Utilisation in 19 OECD Countries, 2008-2009”, OECD Health Working Papers, No. 58, OECD Publishing, Paris, https://dx.doi.org/10.1787/5k95xd6stnxt-en.

[6] Doorslaer, E., X. Koolman and A. Jones (2004), “Explaining income-related inequalities in doctor utilisation in Europe”, Health Economics, Vol. 13/7, pp. 629-647, http://dx.doi.org/10.1002/hec.919.

[10] Garcia-Subirats, I. et al. (2014), “Inequities in access to health care in different health systems: a study in municipalities of central Colombia and north-eastern Brazil.”, International journal for equity in health, Vol. 13, p. 10, http://dx.doi.org/10.1186/1475-9276-13-10.

[13] Goddard, M. and P. Smith (2001), “Equity of access to health care services: Theory and evidence from the UK”, Social Science & Medicine, Vol. 53/9, pp. 1149-1162, http://dx.doi.org/10.1016/S0277-9536(00)00415-9.

[14] Kay, E. (1999), “How often should we go to the dentist?”, BMJ (Clinical research ed.), Vol. 319/7204, pp. 204-5, http://www.ncbi.nlm.nih.gov/pubmed/10417062 (accessed on 11 October 2018).

[21] Listl, S. (2011), “Income-related inequalities in dental service utilization by Europeans aged 50+.”, Journal of dental research, Vol. 90/6, pp. 717-23, http://dx.doi.org/10.1177/0022034511399907.

[5] O’Donnell, O. et al. (2016), Conindex: Estimation of concentration indices, https://folk.uib.no/secaa/Public/Undervisning/ECON220/Kap05_ci_Lorenz.pdf (accessed on 27 November 2018).

[3] O’Donnell, O. et al. (2008), Analyzing Health Equity Using Household Survey Data A Guide to Techniques and Their Implementation Analyzing Health Equity Using Household Survey Data, World Bank Group, Washington D.C, http://dx.doi.org/10.1596/978-0-8213-6933-3.

[2] OECD (2003), “Equity in the Use of Physician Visits in OECD Countries: Has Equal Treatment for Equal Need Been Achieved?”, in Measuring Up: Improving Health System Performance in OECD Countries, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264195950-13-en.

[23] OECD (forthcoming), Doing Things Differently: Towards Better Primary Health Care in the 21st Century, OECD, Paris.

[7] Or, Z., F. Jusot and E. Yilmaz (2008), “Impact of Health Care System on Socioeconomic Inequalities in Doctor Use Impact of health care system on socioeconomic inequalities in doctor use”, Working Paper, No. DT n° 17, IRDES, http://www.irdes.fr (accessed on 7 September 2018).

[20] Palència, L. et al. (2014), “Socio-economic inequalities in the use of dental care services in Europe: What is the role of public coverage?”, Community Dent Oral Epidemiol., Vol. 42/2, pp. 97-105, http://dx.doi.org/10.1111/cdoe.12056.

[19] Palencia, L. et al. (2010), “Socio-economic inequalities in breast and cervical cancer screening practices in Europe: influence of the type of screening program”, International Journal of Epidemiology, Vol. 39/3, pp. 757-765, http://dx.doi.org/10.1093/ije/dyq003.

[17] Ponti A et al. (2017), Cancer Screening in the European Union - Report on the implementation of the Council Recommendation on cancer screening, https://ec.europa.eu/health/sites/health/files/major_chronic_diseases/docs/2017_cancerscreening_2ndreportimplementation_en.pdf (accessed on 14 September 2018).

[18] Sirven, N. and Z. Or (2011), “Disparities in Regular Health Care Utilisation in Europe”, in The Individual and the Welfare State, Springer Berlin Heidelberg, Berlin, Heidelberg, http://dx.doi.org/10.1007/978-3-642-17472-8_22.

[22] Tchicaya, A. and N. Lorentz (2014), Socioeconomic inequalities in the non-use of dental care in Europe, http://dx.doi.org/10.1186/1475-9276-13-7.

[1] van Doorslaer, E. and C. Masseria (2004), “Income-Related Inequality in the Use of Medical Care in 21 OECD Countries”, OECD Health Working Papers, No. 14, OECD Publishing, Paris, https://dx.doi.org/10.1787/687501760705.

[11] Vásquez, F., G. Paraje and M. Estay (2013), “Income-related inequality in health and health care utilization in Chile, 2000 - 2009”, Revista Panamericana de Salud Pública, Vol. 33/2, pp. 98-106, http://dx.doi.org/10.1590/S1020-49892013000200004.

[16] WHO (2014), WHO position paper on mammography screening, http://www.who. (accessed on 31 January 2019).

copy the linklink copied!Annex 3.A. Additional results on inequalities in utilisation of care

## copy the linklink copied!Utilisation of physician and hospital care

Annex Table 3.A.1. Descriptive statistics: Physician and hospital care

Visited any doctor

Visited any GP

Visited any specialist

Number of GP visits

Number of specialist visits

Hospitalised as inpatient

EU28/27

78%

71%

47%

0.70

0.65

10%

OECD

79%

73%

47%

0.69

0.66

10%

Total

78%

70%

46%

0.68

0.63

10%

Austria

86%

76%

63%

0.65

0.53

15%

Belgium

84%

79%

48%

-

-

10%

Bulgaria

77%

74%

32%

0.75

0.56

10%

76%

71%

32%

2.48

1.06

8%

Chile

22%

16%

10%

0.69

0.31

5%

Croatia

77%

72%

48%

0.83

0.59

10%

Cyprus

66%

14%

61%

0.41

0.55

8%

Czech Republic

86%

75%

62%

0.57

0.60

12%

Denmark

83%

80%

35%

0.75

0.72

8%

Estonia

76%

66%

51%

0.47

0.71

10%

Finland

75%

68%

42%

0.52

0.53

9%

France

90%

88%

49%

0.78

0.70

12%

Germany

87%

79%

65%

0.90

0.91

15%

Greece

77%

59%

47%

0.67

0.61

9%

Hungary

84%

76%

62%

0.71

0.76

13%

Iceland

76%

68%

37%

0.45

0.50

8%

Ireland

78%

74%

35%

0.78

0.59

16%

Italy

81%

75%

55%

1.23

0.84

8%

Latvia

77%

71%

55%

0.47

0.39

11%

Lithuania

76%

74%

38%

0.66

0.50

13%

Luxembourg

88%

82%

54%

0.70

0.81

11%

Malta

79%

76%

34%

0.64

0.50

8%

Netherlands

76%

70%

42%

0.59

0.64

8%

Norway

77%

74%

33%

0.47

0.30

9%

Poland

82%

77%

56%

0.65

0.64

13%

Portugal

84%

75%

48%

0.40

0.45

9%

Romania

46%

45%

17%

0.53

0.31

4%

Slovak Republic

75%

69%

44%

0.54

0.63

12%

Slovenia

72%

66%

43%

0.86

0.77

11%

Spain

85%

77%

55%

0.50

0.37

8%

Sweden

66%

60%

34%

1.68

1.79

9%

United Kingdom

77%

74%

34%

0.59

0.46

9%

United States

65%

-

-

-

-

7%

Note: Proportion of adults who had a medical visit in the past 12 months, except in Chile (in the past 3 months). Number of visits in the past 4 weeks for people who had a medical visit in the past year, except in Canada (number of visits in the past 12 months). They are excluded from the averages as relevant.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.2. Quintile distribution of the probability of a GP visit after needs-standardisation, inequality index

Poorest

Quintile 2

Quintile 3

Quintile 4

Richest

Total

GS

CI

EU28

67

69

70

71

72

70

OECD

70

72

73

74

74

73

Total

67

69

70

71

72

70

Austria

73

75

75

78

76

76

*

0.011*

(71-74)

(74-77)

(74-77)

(77-80)

(75-78)

(75-76)

Belgium

79

77

82

84

82

81

*

0.012*

(76-81)

(74-80)

(80-85)

(81-86)

(80-84)

(80-82)

Bulgaria

57

70

74

75

80

72

*

0.061*

(54-59)

(67-73)

(72-77)

(73-78)

(78-83)

(71-73)

66

68

70

74

75

70

*

0.026*

(65-68)

(66-69)

(68-71)

(72-76)

(73-76)

(69-71)

Chile

15

16

16

17

19

17

*

0.043*

(15-15)

(16-16)

(16-17)

(17-18)

(19-19)

(17-17)

Croatia

66

72

70

71

77

71

0.025*

(63-70)

(68-75)

(67-73)

(68-75)

(73-80)

(70-73)

Cyprus

12

13

15

14

16

14

0.050*

(10-14)

(11-16)

(13-18)

(12-16)

(14-18)

(13-15)

Czech Republic

72

74

77

76

73

74

0.000

(70-75)

(72-76)

(74-79)

(73-78)

(70-75)

(73-75)

Denmark

81

81

80

79

74

79

-0.017*

(79-84)

(79-84)

(78-83)

(77-82)

(71-77)

(78-80)

Estonia

64

62

65

65

72

66

0.025*

(60-67)

(59-65)

(61-68)

(62-69)

(68-75)

(64-67)

Finland

57

66

69

72

74

68

*

0.049*

(54-59)

(63-69)

(67-72)

(70-75)

(72-77)

(67-69)

France

83

86

87

87

88

86

*

0.011*

(81-84)

(85-87)

(85-88)

(86-89)

(87-89)

(86-87)

Germany

77

79

80

79

78

79

*

0.000

(76-78)

(78-80)

(79-81)

(78-80)

(76-79)

(78-79)

Greece

55

57

57

59

66

58

0.032*

(52-57)

(55-59)

(55-60)

(56-61)

(63-68)

(57-59)

Hungary

74

76

74

75

75

75

0.003

(71-76)

(74-78)

(71-76)

(73-78)

(73-78)

(74-76)

Iceland

65

69

66

70

69

68

0.010

(62-68)

(66-72)

(62-69)

(66-73)

(66-72)

(66-69)

Ireland

71

72

73

72

73

72

0.004

(69-73)

(70-74)

(71-75)

(70-74)

(71-75)

(71-73)

Italy

70

72

75

76

75

74

*

0.015

(69-71)

(71-74)

(74-76)

(74-77)

(74-76)

(73-74)

Latvia

62

70

70

72

73

69

*

0.027*

(60-65)

(68-72)

(68-73)

(69-74)

(71-76)

(68-71)

Lithuania

74

69

72

77

77

74

0.016*

(71-76)

(67-72)

(70-75)

(74-79)

(75-80)

(73-75)

Luxembourg

81

80

83

81

83

82

0.005

(77-84)

(76-83)

(80-86)

(77-84)

(79-86)

(80-83)

Malta

76

78

74

78

76

76

-0.001

(74-78)

(75-80)

(71-77)

(74-81)

(72-80)

(75-78)

Netherlands

69

70

69

69

69

69

0.000

(66-72)

(67-72)

(67-72)

(67-72)

(67-71)

(68-70)

Norway

70

74

73

76

75

74

*

0.011*

(68-73)

(72-77)

(71-75)

(74-78)

(73-77)

(73-75)

Poland

69

74

76

78

82

76

*

0.032*

(67-70)

(73-75)

(74-77)

(77-80)

(80-83)

(75-76)

Portugal

76

78

80

78

77

78

*

0.002

(75-77)

(76-79)

(78-81)

(77-80)

(75-78)

(77-78)

Romania

35

40

44

49

50

44

*

0.068*

(34-37)

(39-42)

(43-46)

(47-50)

(48-51)

(43-45)

Slovak Republic

68

70

71

69

66

69

-0.008

(66-71)

(68-73)

(69-74)

(66-71)

(63-69)

(68-70)

Slovenia

63

65

67

70

68

66

0.019*

(60-66)

(62-67)

(64-70)

(67-73)

(65-71)

(65-68)

Spain

76

77

78

76

73

76

*

-0.008*

(74-77)

(76-79)

(76-79)

(75-77)

(71-74)

(75-76)

Sweden

59

61

60

56

57

59

-0.012

(56-62)

(58-63)

(57-63)

(53-59)

(55-60)

(57-60)

United Kingdom

74

73

72

75

74

74

0.002

(73-76)

(72-74)

(71-74)

(74-77)

(73-76)

(73-75)

Notes: Probabilities are expressed in percentages and indirectly standardised for need controlling for marital status, income, education, occupational status, income size of household and urbanisation level. GS refers to the “Global Significance” of the “income” variable in the regression used for the indirect standardisation. The star (*) denotes significant at 5%. CI means concentration index.

In Chile, visits refer to the past 3 months, Chile is not included in average.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.3. Annex Table 3.A.3. Quintile distribution of the number of GP visits after needs-standardisation, inequality index

Poorest

Quintile 2

Quintile 3

Quintile 4

Richest

Total

GS

CI

EU27

0.73

0.70

0.70

0.66

0.66

0.69

OECD

0.72

0.70

0.69

0.65

0.65

0.68

Total

0.71

0.68

0.68

0.65

0.64

0.67

Austria

0.67

0.66

0.65

0.69

0.65

0.66

0.000

(0.61-0.73)

(0.61-0.7)

(0.61-0.69)

(0.65-0.74)

(0.6-0.7)

(0.64-0.69)

Bulgaria

0.77

0.82

0.76

0.69

0.72

0.74

-0.022

(0.67-0.86)

(0.73-0.9)

(0.69-0.83)

(0.63-0.75)

(0.66-0.79)

(0.71-0.78)

2.44

2.46

2.40

2.38

2.49

2.43

-0.000

(2.28-2.6)

(2.32-2.6)

(2.26-2.54)

(2.27-2.5)

(2.36-2.61)

(2.37-2.5)

Chile

0.68

0.79

0.70

0.68

0.68

0.71

-0.012

(0.61-0.75)

(0.71-0.86)

(0.63-0.77)

(0.62-0.75)

(0.62-0.74)

(0.68-0.74)

Croatia

0.84

0.69

0.93

0.92

0.83

0.85

0.018

(0.68-1)

(0.6-0.78)

(0.77-1.09)

(0.79-1.05)

(0.73-0.93)

(0.79-0.9)

Cyprus

0.44

0.4

0.34

0.43

0.4

0.4

-0.006

(0.33-0.55)

(0.31-0.49)

(0.26-0.43)

(0.34-0.53)

(0.32-0.49)

(0.36-0.44)

Czech Republic

0.6

0.58

0.61

0.5

0.54

0.56

*

-0.030*

(0.53-0.68)

(0.52-0.64)

(0.54-0.69)

(0.45-0.55)

(0.49-0.59)

(0.53-0.59)

Denmark

0.82

0.8

0.68

0.65

0.66

0.71

-0.051*

(0.65-0.99)

(0.7-0.89)

(0.61-0.76)

(0.59-0.72)

(0.59-0.72)

(0.67-0.76)

Estonia

0.39

0.49

0.47

0.41

0.43

0.44

0.001

(0.32-0.46)

(0.41-0.57)

(0.4-0.55)

(0.34-0.47)

(0.35-0.52)

(0.4-0.47)

Finland

0.52

0.55

0.53

0.5

0.49

0.51

-0.017

(0.44-0.59)

(0.46-0.63)

(0.47-0.58)

(0.44-0.56)

(0.43-0.54)

(0.49-0.54)

France

0.91

0.68

0.76

0.75

0.74

0.76

*

-0.025*

(0.85-0.97)

(0.63-0.72)

(0.71-0.81)

(0.66-0.84)

(0.67-0.81)

(0.73-0.79)

Germany

1.00

0.92

0.86

0.83

0.81

0.88

*

-0.043*

(0.95-1.05)

(0.88-0.97)

(0.83-0.9)

(0.8-0.87)

(0.77-0.85)

(0.87-0.9)

Greece

0.62

0.67

0.77

0.65

0.65

0.66

0.007

(0.54-0.7)

(0.61-0.74)

(0.7-0.84)

(0.59-0.71)

(0.59-0.7)

(0.63-0.69)

Hungary

0.74

0.68

0.72

0.66

0.68

0.7

-0.016

(0.67-0.81)

(0.62-0.74)

(0.64-0.79)

(0.58-0.73)

(0.6-0.76)

(0.66-0.73)

Iceland

0.41

0.5

0.45

0.44

0.42

0.45

-0.007

(0.33-0.5)

(0.43-0.57)

(0.38-0.53)

(0.38-0.5)

(0.37-0.48)

(0.41-0.48)

Ireland

0.79

0.9

0.81

0.77

0.78

0.81

-0.015

(0.71-0.87)

(0.82-0.98)

(0.75-0.88)

(0.7-0.84)

(0.72-0.84)

(0.78-0.84)

Italy

1.33

1.31

1.18

1.12

1.14

1.21

*

-0.036*

(1.27-1.39)

(1.25-1.36)

(1.13-1.23)

(1.08-1.17)

(1.09-1.19)

(1.19-1.24)

Latvia

0.49

0.46

0.45

0.39

0.46

0.45

-0.024

(0.44-0.54)

(0.42-0.5)

(0.41-0.5)

(0.36-0.43)

(0.41-0.5)

(0.43-0.47)

Lithuania

0.65

0.65

0.65

0.63

0.59

0.63

-0.018

(0.6-0.71)

(0.58-0.71)

(0.58-0.71)

(0.56-0.69)

(0.54-0.64)

(0.61-0.66)

Luxembourg

0.72

0.63

0.66

0.61

0.63

0.65

-0.023

(0.59-0.84)

(0.53-0.72)

(0.58-0.75)

(0.53-0.7)

(0.55-0.7)

(0.61-0.69)

Malta

0.69

0.59

0.58

0.70

0.67

0.64

0.003

(0.63-0.75)

(0.53-0.66)

(0.52-0.65)

(0.6-0.8)

(0.57-0.77)

(0.61-0.68)

Netherlands

0.65

0.52

0.6

0.6

0.56

0.58

-0.010

(0.54-0.75)

(0.46-0.59)

(0.51-0.69)

(0.55-0.65)

(0.51-0.61)

(0.55-0.62)

Norway

0.43

0.5

0.46

0.48

0.49

0.47

0.017

(0.38-0.48)

(0.44-0.55)

(0.41-0.5)

(0.43-0.52)

(0.45-0.54)

(0.45-0.49)

Poland

0.66

0.67

0.69

0.7

0.62

0.67

*

-0.008

(0.62-0.69)

(0.64-0.71)

(0.65-0.72)

(0.67-0.73)

(0.59-0.65)

(0.65-0.68)

Portugal

0.42

0.41

0.4

0.4

0.38

0.4

-0.019*

(0.39-0.45)

(0.39-0.44)

(0.37-0.43)

(0.38-0.43)

(0.35-0.4)

(0.39-0.41)

Romania

0.5

0.51

0.55

0.51

0.55

0.53

*

0.017*

(0.46-0.53)

(0.48-0.54)

(0.52-0.58)

(0.49-0.54)

(0.53-0.58)

(0.51-0.54)

Slovak Republic

0.52

0.53

0.56

0.5

0.55

0.53

0.002

(0.46-0.57)

(0.48-0.58)

(0.51-0.62)

(0.44-0.55)

(0.46-0.63)

(0.5-0.56)

Slovenia

1.02

0.97

0.8

0.68

0.73

0.85

*

-0.082*

(0.84-1.2)

(0.84-1.1)

(0.7-0.9)

(0.6-0.75)

(0.65-0.81)

(0.8-0.91)

Spain

0.55

0.51

0.51

0.47

0.47

0.5

*

-0.032*

(0.51-0.59)

(0.48-0.54)

(0.48-0.54)

(0.44-0.49)

(0.45-0.5)

(0.49-0.52)

Sweden

1.8

1.62

1.75

1.65

1.5

1.66

-0.020

(1.61-1.99)

(1.43-1.82)

(1.55-1.95)

(1.51-1.79)

(1.37-1.63)

(1.58-1.74)

United Kingdom

0.67

0.61

0.57

0.54

0.54

0.58

-0.044*

(0.63-0.71)

(0.57-0.65)

(0.54-0.61)

(0.51-0.57)

(0.51-0.57)

(0.57-0.6)

Notes: The number of visits over the past 4 weeks is indirectly standardised for need controlling for marital status, income, education, occupational status, income, size of household and urbanisation level. GS refers to the “Global Significance” of the “income” variable in the regression used for the indirect standardisation. The star (*) denotes significant at 5%. CI means concentration index.

In Chile (Canada), visits refer to the past 3 (12) months, Chile (Canada) is not included in average.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.4. Quintile distribution of the probability of a specialist visit after needs-standardisation, inequality index

Poorest

Quintile 2

Quintile 3

Quintile 4

Richest

Total

GS

CI

EU28

41

44

47

49

52

47

OECD

41

44

47

49

52

47

Total

39

43

45

47

51

45

Austria

57

61

64

65

68

63

*

0.032*

(55-59)

(59-63)

(62-65)

(64-67)

(66-69)

(62-64)

Belgium

47

41

46

50

48

47

*

0.019*

(43-50)

(37-44)

(43-49)

(47-53)

(45-50)

(45-48)

Bulgaria

20

30

32

38

43

33

*

0.137*

(18-23)

(26-33)

(29-35)

(35-40)

(41-46)

(32-35)

26

30

32

36

35

32

*

0.056*

(25-28)

(29-32)

(30-34)

(34-37)

(33-37)

(31-32)

Chile

8

8

10

12

18

11

*

0.176*

(7-8)

(8-9)

(9-10)

(11-12)

(18-18)

(11-11)

Croatia

38

45

49

48

59

49

*

0.076*

(35-42)

(41-49)

(45-53)

(45-52)

(55-63)

(47-50)

Cyprus

53

60

58

63

67

60

*

0.041*

(50-56)

(57-63)

(55-61)

(61-66)

(64-69)

(59-61)

Czech Republic

54

63

61

63

67

62

*

0.036*

(51-57)

(61-66)

(58-64)

(61-66)

(65-69)

(61-63)

Denmark

33

33

36

38

33

35

0.005

(30-37)

(30-37)

(33-39)

(35-41)

(30-36)

(33-36)

Estonia

47

48

56

52

59

53

*

0.041*

(43-51)

(45-52)

(53-59)

(48-55)

(56-63)

(51-54)

Finland

32

38

41

44

51

42

*

0.083*

(30-35)

(35-41)

(38-44)

(42-47)

(48-54)

(40-43)

France

42

43

48

51

56

48

*

0.058*

(41-44)

(41-45)

(46-50)

(49-52)

(54-58)

(47-49)

Germany

60

62

66

66

69

64

*

0.029*

(58-61)

(60-63)

(64-67)

(65-68)

(68-70)

(64-65)

Greece

42

45

45

49

53

46

*

0.044*

(40-45)

(42-47)

(42-48)

(46-51)

(51-55)

(45-47)

Hungary

55

60

62

63

65

61

0.030*

(52-58)

(57-62)

(59-64)

(61-66)

(62-68)

(60-62)

Iceland

31

36

37

41

40

37

*

0.051*

(27-34)

(33-40)

(34-40)

(38-44)

(37-43)

(36-39)

Ireland

35

34

33

34

36

35

0.008

(33-37)

(32-36)

(31-36)

(32-36)

(34-39)

(34-36)

Italy

45

51

55

58

62

54

*

0.060*

(44-46)

(49-52)

(54-56)

(57-59)

(60-63)

(54-55)

Latvia

45

51

55

59

63

55

*

0.066*

(42-47)

(49-54)

(52-57)

(56-61)

(61-66)

(54-56)

Lithuania

40

32

32

40

42

37

*

0.031*

(37-43)

(29-35)

(29-35)

(37-43)

(40-45)

(36-39)

Luxembourg

52

55

56

56

60

56

0.025*

(48-57)

(50-59)

(52-60)

(52-60)

(56-65)

(54-58)

Malta

29

33

34

38

36

34

*

0.048*

(26-32)

(30-36)

(31-38)

(35-42)

(32-40)

(32-35)

Netherlands

38

38

44

40

43

41

*

0.023*

(35-41)

(36-41)

(41-46)

(38-42)

(41-45)

(40-42)

Norway

30

32

35

35

38

34

*

0.047*

(27-32)

(29-34)

(33-37)

(33-37)

(36-40)

(33-35)

Poland

44

51

56

57

67

55

*

0.075*

(43-46)

(49-52)

(55-58)

(55-58)

(65-69)

(55-56)

Portugal

39

43

46

54

66

50

*

0.105*

(37-40)

(41-45)

(45-48)

(52-55)

(64-67)

(49-50)

Romania

12

15

18

21

21

18

*

0.109*

(11-13)

(13-16)

(17-19)

(20-22)

(20-22)

(17-18)

Slovak Republic

41

41

45

44

43

43

0.012

(39-44)

(38-44)

(42-48)

(41-46)

(41-46)

(42-44)

Slovenia

36

42

43

45

52

43

*

0.064*

(32-39)

(39-44)

(40-46)

(41-48)

(49-55)

(42-45)

Spain

45

49

54

57

65

54

*

0.071*

(44-47)

(47-50)

(53-56)

(56-59)

(63-66)

(54-55)

Sweden

27

29

34

33

39

33

*

0.064*

(24-30)

(26-32)

(31-36)

(31-36)

(36-41)

(32-34)

United Kingdom

27

32

34

34

36

33

*

0.047*

(26-29)

(31-34)

(32-35)

(33-36)

(35-38)

(32-34)

Notes: Probabilities are expressed in percentages and indirectly standardised for need controlling for marital status, income, education, occupational status, income, size of household and urbanisation level. GS refers to the “Global Significance” of the “income” variable in the regression used for the indirect standardisation. The star (*) denotes significant at 5%. CI means concentration index.

In Chile, visits refer to the past 3 months, Chile is not included in average.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.5. Quintile distribution of number of specialist visits after needs-standardisation, inequality index

Poorest

Quintile 2

Quintile 3

Quintile 4

Richest

Total

GS

CI

EU27

0.65

0.66

0.65

0.64

0.66

0.65

OECD

0.66

0.66

0.65

0.65

0.67

0.66

Total

0.63

0.64

0.63

0.62

0.64

0.63

Austria

0.51

0.51

0.55

0.58

0.55

0.54

0.024*

(0.46-0.56)

(0.46-0.55)

(0.5-0.61)

(0.53-0.63)

(0.51-0.59)

(0.52-0.56)

Bulgaria

0.65

0.59

0.72

0.5

0.42

0.55

-0.094*

(0.51-0.79)

(0.43-0.75)

(0.53-0.92)

(0.4-0.61)

(0.34-0.5)

(0.49-0.61)

0.93

0.91

0.94

1.03

1.02

0.97

0.024

(0.81-1.06)

(0.82-1)

(0.84-1.03)

(0.94-1.13)

(0.92-1.12)

(0.92-1.01)

Chile

0.27

0.26

0.26

0.33

0.46

0.32

*

0.113*

(0.23-0.31)

(0.22-0.3)

(0.23-0.3)

(0.29-0.37)

(0.42-0.5)

(0.30-0.34)

Croatia

0.56

0.64

0.63

0.56

0.49

0.57

-0.041

(0.43-0.7)

(0.49-0.79)

(0.52-0.73)

(0.48-0.64)

(0.42-0.56)

(0.52-0.62)

Cyprus

0.52

0.55

0.53

0.57

0.56

0.55

0.014

(0.46-0.59)

(0.48-0.62)

(0.48-0.59)

(0.52-0.63)

(0.51-0.62)

(0.52-0.57)

Czech Republic

0.58

0.66

0.62

0.54

0.64

0.61

0.002

(0.5-0.67)

(0.58-0.73)

(0.54-0.69)

(0.49-0.6)

(0.57-0.71)

(0.57-0.64)

Denmark

0.62

0.75

0.84

0.69

0.65

0.71

-0.006

(0.51-0.73)

(0.61-0.89)

(0.66-1.02)

(0.58-0.81)

(0.52-0.77)

(0.65-0.77)

Estonia

0.86

0.88

0.71

0.72

0.72

0.77

-0.044

(0.62-1.09)

(0.65-1.11)

(0.56-0.87)

(0.57-0.87)

(0.6-0.85)

(0.69-0.86)

Finland

0.48

0.49

0.53

0.56

0.6

0.54

0.046*

(0.39-0.58)

(0.4-0.58)

(0.44-0.61)

(0.47-0.65)

(0.52-0.67)

(0.5-0.58)

France

0.76

0.64

0.73

0.67

0.69

0.7

-0.011

(0.68-0.84)

(0.57-0.71)

(0.64-0.82)

(0.55-0.79)

(0.64-0.75)

(0.66-0.74)

Germany

0.97

0.86

0.86

0.88

0.88

0.89

-0.013

(0.92-1.03)

(0.81-0.9)

(0.82-0.91)

(0.83-0.92)

(0.84-0.93)

(0.87-0.91)

Greece

0.63

0.6

0.63

0.6

0.66

0.62

0.006

(0.54-0.72)

(0.53-0.66)

(0.56-0.69)

(0.53-0.67)

(0.58-0.74)

(0.59-0.66)

Hungary

0.72

0.81

0.81

0.69

0.75

0.76

-0.005

(0.56-0.87)

(0.67-0.94)

(0.71-0.91)

(0.61-0.77)

(0.66-0.84)

(0.71-0.81)

Iceland

0.55

0.56

0.46

0.5

0.47

0.51

-0.037

(0.4-0.7)

(0.45-0.67)

(0.38-0.55)

(0.41-0.58)

(0.37-0.56)

(0.46-0.55)

Ireland

0.55

0.64

0.56

0.59

0.71

0.61

0.039*

(0.48-0.63)

(0.54-0.73)

(0.48-0.64)

(0.51-0.67)

(0.62-0.8)

(0.58-0.65)

Italy

0.83

0.83

0.8

0.82

0.82

0.82

-0.001

(0.76-0.9)

(0.76-0.89)

(0.75-0.85)

(0.78-0.87)

(0.78-0.87)

(0.8-0.85)

Latvia

0.41

0.37

0.34

0.4

0.42

0.39

0.016

(0.36-0.46)

(0.31-0.42)

(0.29-0.38)

(0.35-0.45)

(0.38-0.47)

(0.37-0.41)

Lithuania

0.54

0.5

0.46

0.45

0.48

0.49

-0.029

(0.47-0.62)

(0.41-0.59)

(0.38-0.53)

(0.38-0.53)

(0.4-0.56)

(0.45-0.52)

Luxembourg

0.67

0.69

0.8

0.82

0.95

0.79

0.070*

(0.53-0.81)

(0.52-0.85)

(0.62-0.98)

(0.66-0.98)

(0.84-1.07)

(0.72-0.86)

Malta

0.53

0.56

0.49

0.49

0.50

0.52

-0.020

(0.45-0.61)

(0.44-0.68)

(0.39-0.59)

(0.38-0.59)

(0.36-0.65)

(0.47-0.57)

Netherlands

0.53

0.6

0.75

0.66

0.69

0.66

0.038*

(0.43-0.63)

(0.47-0.73)

(0.62-0.88)

(0.58-0.75)

(0.6-0.78)

(0.61-0.71)

Norway

0.32

0.24

0.31

0.27

0.33

0.3

0.021

(0.24-0.41)

(0.17-0.3)

(0.26-0.36)

(0.22-0.33)

(0.28-0.39)

(0.27-0.32)

Poland

0.62

0.65

0.64

0.68

0.71

0.66

0.028*

(0.57-0.67)

(0.6-0.69)

(0.6-0.68)

(0.64-0.72)

(0.67-0.76)

(0.64-0.68)

Portugal

0.49

0.43

0.41

0.43

0.59

0.48

0.048*

(0.41-0.57)

(0.39-0.47)

(0.37-0.45)

(0.39-0.46)

(0.56-0.63)

(0.45-0.5)

Romania

0.29

0.31

0.28

0.32

0.37

0.32

0.048*

(0.24-0.34)

(0.26-0.35)

(0.24-0.32)

(0.27-0.36)

(0.32-0.42)

(0.3-0.34)

Slovak Republic

0.58

0.62

0.67

0.66

0.7

0.65

0.035*

(0.5-0.65)

(0.55-0.7)

(0.59-0.74)

(0.58-0.74)

(0.61-0.79)

(0.61-0.68)

Slovenia

0.79

0.86

0.72

0.83

0.69

0.78

-0.027

(0.64-0.94)

(0.71-1.02)

(0.63-0.82)

(0.7-0.95)

(0.59-0.79)

(0.72-0.84)

Spain

0.41

0.4

0.37

0.41

0.44

0.41

0.015

(0.36-0.47)

(0.36-0.44)

(0.34-0.4)

(0.37-0.45)

(0.41-0.47)

(0.39-0.42)

Sweden

2

1.9

1.62

1.69

1.51

1.73

-0.049*

(1.39-2.61)

(1.58-2.22)

(1.42-1.81)

(1.46-1.93)

(1.3-1.73)

(1.6-1.87)

United Kingdom

0.42

0.45

0.43

0.48

0.52

0.46

0.042*

(0.36-0.47)

(0.38-0.52)

(0.39-0.47)

(0.43-0.53)

(0.47-0.57)

(0.44-0.49)

Notes: The number of visits over the past 4 weeks is indirectly standardised for need controlling for marital status, income, education, occupational status, income size of household and urbanisation level. GS refers to the “Global Significance” of the “income” variable in the regression used for the indirect standardisation. The star (*) denotes significant at 5%. CI means concentration index.

In Chile (Canada), visits refer to the past 3 (12) months, Chile (Canada) is not included in average.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.6. Quintile distribution of the probability of a doctor visit after needs-standardisation, inequality index

Poorest

Quintile 2

Quintile 3

Quintile 4

Richest

Total

GS

CI

EU28

74

76

78

79

81

77

OECD

75

77

79

80

81

78

Total

73

76

77

78

80

77

Austria

82

85

86

88

88

86

*

0.014*

(80-83)

(84-87)

(85-87)

(87-89)

(87-89)

(85-86)

Belgium

83

81

87

88

86

86

*

0.011*

(80-85)

(78-83)

(85-89)

(86-90)

(84-88)

(85-86)

Bulgaria

59

74

76

78

83

75

*

0.059*

(57-62)

(71-76)

(74-78)

(76-81)

(81-86)

(74-76)

72

73

74

79

79

75

*

0.022*

(70-73)

(71-74)

(73-76)

(77-80)

(77-80)

(74-76)

Chile♦

19

21

22

24

31

24

*

0.089*

(19-20)

(21-21)

(21-22)

(24-25)

(30-31)

(23-24)

Croatia

71

76

76

76

83

77

0.026*

(68-74)

(72-79)

(73-79)

(73-80)

(80-86)

(75-78)

Cyprus

58

64

63

67

72

65

*

0.039*

(55-61)

(61-67)

(61-66)

(65-70)

(69-74)

(63-66)

Czech Republic

83

85

87

85

87

85

0.008*

(81-85)

(83-87)

(85-88)

(84-87)

(85-89)

(85-86)

Denmark

82

82

83

81

76

81

-0.013*

(80-85)

(79-84)

(80-85)

(79-84)

(74-79)

(80-82)

Estonia

72

71

74

76

83

75

*

0.028*

(69-75)

(69-74)

(71-77)

(73-79)

(80-86)

(74-77)

Finland

63

72

75

78

82

74

*

0.045*

(61-66)

(70-75)

(73-78)

(76-80)

(79-84)

(73-75)

France

86

88

89

90

91

89

*

0.011*

(85-87)

(87-89)

(88-90)

(89-91)

(90-92)

(88-89)

Germany

84

86

87

87

88

86

*

0.008*

(83-85)

(85-87)

(86-88)

(86-88)

(87-89)

(86-87)

Greece

72

75

76

77

84

76

*

0.028*

(70-74)

(73-77)

(74-78)

(75-79)

(82-86)

(75-77)

Hungary

81

83

84

85

87

84

0.013*

(79-83)

(81-85)

(82-86)

(82-87)

(85-89)

(83-85)

Iceland

72

76

73

78

76

75

0.011*

(69-75)

(74-79)

(70-76)

(75-81)

(73-79)

(74-77)

Ireland

75

75

76

75

76

75

0.003

(73-77)

(73-77)

(74-78)

(73-77)

(74-78)

(74-76)

Italy

75

78

82

82

83

80

*

0.019*

(74-76)

(77-79)

(80-83)

(81-83)

(82-84)

(79-80)

Latvia

67

76

77

79

82

76

*

0.035*

(65-70)

(74-78)

(75-79)

(77-81)

(80-84)

(75-77)

Lithuania

76

71

75

78

79

76

0.016*

(73-78)

(69-74)

(72-77)

(76-81)

(77-82)

(75-77)

Luxembourg

87

87

89

87

91

88

0.008

(84-90)

(84-90)

(86-91)

(84-90)

(89-94)

(87-90)

Malta

79

80

77

81

80

79

0.002

(76-81)

(78-83)

(74-80)

(78-84)

(77-84)

(78-81)

Netherlands

73

74

75

75

75

75

0.005

(70-76)

(72-76)

(73-78)

(73-77)

(73-77)

(74-76)

Norway

74

77

76

78

79

77

*

0.011*

(72-76)

(75-79)

(74-78)

(76-80)

(77-81)

(76-78)

Poland

73

78

80

82

87

80

*

0.032*

(72-74)

(77-79)

(78-81)

(81-83)

(86-88)

(80-81)

Portugal

81

83

86

87

91

86

*

0.023*

(80-82)

(82-84)

(85-87)

(86-88)

(90-92)

(85-86)

Romania

36

41

46

50

51

45

*

0.068*

(34-38)

(40-43)

(44-47)

(49-52)

(49-52)

(44-46)

Slovak Republic

73

75

77

75

72

74

-0.003

(70-75)

(73-77)

(74-79)

(72-77)

(69-74)

(73-75)

Slovenia

68

70

71

75

74

71

0.019*

(65-71)

(67-73)

(68-74)

(72-78)

(71-77)

(70-73)

Spain

81

83

84

85

86

84

*

0.012*

(79-82)

(82-84)

(83-86)

(84-86)

(85-87)

(83-84)

Sweden

62

65

66

61

65

64

0.003

(59-65)

(62-68)

(64-69)

(59-64)

(63-68)

(63-65)

United Kingdom

76

76

75

78

77

76

0.004

(75-77)

(74-77)

(74-77)

(77-79)

(76-78)

(76-77)

United States

59

57

61

67

74

65

*

0.053*

(57-60)

(56-59)

(60-62)

(66-68)

(73-75)

(65-66)

Notes: Probabilities are expressed in percentages and indirectly standardised for need controlling for marital status, income, education, occupational status, income, size of household and urbanisation level. GS refers to the “Global Significance” of the “income” variable in the regression used for the indirect standardisation. The star (*) denotes significant at 5%. CI means concentration index.

In Chile, visits refer to the past 3 months, Chile is not included in average.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.7. Quintile distribution of the probability of inpatient hospitalisation after needs-standardisation, inequality index

Poorest

Quintile 2

Quintile 3

Quintile 4

Richest

Total

GS

CI

EU28

11

11

11

11

10

11

OECD

11

11

11

11

10

11

Total

10

10

10

10

10

10

Austria

14

14

15

14

15

14

0.016

(12-15)

(12-15)

(13-16)

(13-15)

(14-16)

(14-15)

Belgium

11

11

11

11

9

11

-0.027

(9-13)

(8-13)

(9-13)

(10-13)

(8-11)

(10-11)

Bulgaria

9

10

12

11

10

10

0.001

(8-11)

(8-12)

(10-14)

(9-12)

(8-11)

(10-11)

10

10

10

7

7

9

*

-0.068*

(9-11)

(9-11)

(9-11)

(7-8)

(6-8)

(8-9)

Chile

7

6

6

5

6

6

*

-0.008

(6-7)

(6-6)

(5-6)

(5-6)

(6-7)

(6-6)

Croatia

10

11

11

10

12

11

0.027

(8-13)

(8-13)

(9-14)

(8-12)

(10-15)

(10-12)

Cyprus

8

7

8

9

7

8

0.025

(6-9)

(5-9)

(6-10)

(8-11)

(6-9)

(7-9)

Czech Republic

10

11

13

12

12

11

0.023

(8-12)

(9-13)

(11-14)

(10-13)

(10-13)

(11-12)

Denmark

10

8

7

8

7

8

-0.046

(8-12)

(6-10)

(5-8)

(6-9)

(6-9)

(7-9)

Estonia

11

12

11

8

9

10

-0.066*

(9-14)

(9-14)

(9-14)

(6-10)

(7-11)

(9-11)

Finland

10

9

9

9

9

9

-0.003

(8-12)

(7-10)

(7-10)

(8-11)

(8-11)

(8-10)

France

13

13

12

12

13

13

-0.008

(12-14)

(12-14)

(10-13)

(11-13)

(12-14)

(12-13)

Germany

17

16

16

14

14

16

-0.034*

(16-18)

(15-17)

(15-17)

(13-15)

(14-15)

(15-16)

Greece

9

8

10

9

10

9

0.028

(8-10)

(7-10)

(8-11)

(8-11)

(9-12)

(9-10)

Hungary

13

14

13

14

14

14

0.007

(11-15)

(12-16)

(11-15)

(12-16)

(12-15)

(13-15)

Iceland

9

10

8

9

8

9

-0.025

(7-11)

(8-12)

(6-10)

(7-11)

(7-10)

(8-10)

Ireland

15

18

16

17

16

16

0.000

(14-17)

(16-20)

(14-18)

(15-19)

(14-18)

(16-17)

Italy

8

8

8

9

9

8

0.024*

(7-9)

(7-9)

(7-9)

(8-9)

(8-10)

(8-9)

Latvia

13

12

11

11

10

11

-0.056*

(11-15)

(10-14)

(9-13)

(9-13)

(8-11)

(10-12)

Lithuania

16

12

11

14

12

13

*

-0.036

(13-18)

(10-14)

(9-13)

(12-16)

(10-13)

(12-14)

Luxembourg

15

13

12

12

10

12

-0.065*

(11-18)

(9-16)

(9-14)

(9-14)

(8-12)

(11-13)

Netherlands

7

8

9

8

8

8

0.024

(5-9)

(6-9)

(8-10)

(7-9)

(7-10)

(7-9)

Malta

8

10

9

10

8

9

-0.005

(6-10)

(8-12)

(7-10)

(8-12)

(6-10)

(8-10)

Norway

10

8

10

11

9

9

*

0.006

(8-11)

(6-9)

(8-11)

(9-12)

(7-10)

(9-10)

Poland

13

13

13

14

13

13

0.018

(11-14)

(12-14)

(12-14)

(13-15)

(12-15)

(13-14)

Portugal

10

10

9

10

9

10

-0.020

(9-11)

(9-11)

(8-10)

(9-10)

(9-10)

(9-10)

Romania

3

4

5

5

4

4

*

0.065*

(2-4)

(3-5)

(4-5)

(4-6)

(4-5)

(4-5)

Slovak Republic

11

11

12

14

11

12

0.027

(9-12)

(9-13)

(10-14)

(12-16)

(9-13)

(11-13)

Slovenia

11

11

10

11

9

10

-0.022

(8-13)

(9-13)

(8-12)

(9-13)

(8-11)

(9-11)

Spain

8

7

9

8

8

8

0.025

(7-9)

(6-8)

(8-10)

(7-9)

(8-9)

(8-8)

Sweden

6

10

10

8

8

9

*

0.012

(5-8)

(8-12)

(8-12)

(7-10)

(7-10)

(8-9)

United Kingdom

8

9

9

7

8

8

-0.011

(7-9)

(8-10)

(8-9)

(6-8)

(8-9)

(8-9)

United States

12

10

9

8

9

9

*

-0.064*

(11-13)

(9-11)

(8-10)

(8-9)

(8-9)

(9-10)

Notes: Probabilities are expressed in percentages and indirectly standardised for need controlling for marital status, income, education, occupational status, income, size of household and urbanisation level. GS refers to the “Global Significance” of the “income” variable in the regression used for the indirect standardisation. The star (*) denotes significant at 5%. CI means concentration index.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.8. Descriptive statistics and generalised concentration indexes: Cancer screening

Cervical cancer screening

Breast cancer screening

Colorectal cancer screening

Q1

Mean

Q5

GCI

Q1

Mean

Q5

GCI

Q1

Mean

Q5

GCI

EU28

60%

70%

77%

58%

65%

71%

34%

37%

39%

OECD

65%

73%

79%

63%

70%

74%

38%

42%

44%

Total

61%

71%

78%

59%

66%

72%

34%

38%

40%

Austria

83%

87%

92%

0.018*

67%

73%

80%

0.025*

67%

71%

74%

0.015*

Belgium

61%

76%

80%

0.025*

73%

75%

83%

0.030*

32%

35%

37%

0.010

Bulgaria

30%

52%

74%

0.088*

20%

32%

47%

0.052*

4%

7%

8%

0.009*

71%

76%

80%

0.018*

65%

74%

78%

0.023*

41%

49%

52%

0.018*

Chile

71%

72%

75%

0.007*

56%

61%

68%

0.022*

-

-

-

Croatia

61%

77%

86%

0.043*

54%

67%

75%

0.046*

24%

31%

36%

0.026*

Cyprus

55%

65%

76%

0.046*

52%

66%

78%

0.054*

10%

18%

23%

0.018*

Czech Republic

78%

87%

94%

0.029*

65%

77%

85%

0.039*

50%

57%

58%

0.008

Denmark

51%

64%

74%

0.047*

83%

82%

80%

-0.007

46%

48%

49%

0.003

Estonia

50%

58%

71%

0.042*

41%

39%

46%

0.005

16%

16%

18%

-0.006

Finland

69%

79%

87%

0.035*

77%

86%

89%

0.019*

30%

30%

30%

0.001

France

72%

82%

89%

0.034*

79%

87%

90%

0.020*

53%

64%

67%

0.025*

Germany

73%

81%

86%

0.024*

70%

74%

72%

0.006

71%

74%

77%

0.012*

Greece

69%

76%

81%

0.024*

51%

60%

70%

0.042*

16%

23%

24%

0.013*

Hungary

60%

71%

77%

0.033*

53%

65%

73%

0.042*

23%

23%

19%

-0.002

Iceland

77%

80%

82%

0.012

60%

66%

69%

0.021

36%

42%

48%

0.021*

Ireland

71%

69%

69%

0.000

69%

68%

71%

0.012*

35%

38%

38%

-0.001

Italy

56%

68%

75%

0.038*

55%

67%

75%

0.042*

31%

43%

50%

0.039*

Latvia

68%

78%

84%

0.033*

38%

47%

56%

0.039*

23%

30%

32%

0.015*

Lithuania

59%

62%

68%

0.029*

45%

46%

51%

0.020*

28%

28%

24%

-0.002

Luxembourg

79%

84%

92%

0.027*

82%

81%

77%

-0.007

55%

57%

65%

0.018*

Malta

49%

64%

67%

0.034*

48%

58%

73%

0.044*

25%

27%

27%

0.005

Netherlands

49%

49%

53%

0.013*

73%

80%

82%

0.013

24%

23%

24%

-0.003

Norway

45%

66%

76%

0.056*

59%

76%

79%

0.026*

29%

31%

28%

-0.008

Poland

59%

72%

84%

0.046*

45%

59%

66%

0.034*

16%

20%

23%

0.013*

Portugal

63%

71%

76%

0.026*

82%

84%

86%

0.006

50%

57%

62%

0.021*

Romania

13%

27%

38%

0.050*

2%

7%

10%

0.018*

4%

6%

8%

0.009*

Slovak Republic

61%

69%

71%

0.023*

48%

54%

59%

0.028*

35%

35%

34%

0.000

Slovenia

69%

78%

85%

0.027*

56%

61%

71%

0.022*

59%

69%

81%

0.037*

Spain

56%

69%

81%

0.050*

68%

80%

86%

0.032*

19%

26%

35%

0.028*

Sweden

56%

81%

89%

0.053*

74%

91%

93%

0.015*

43%

25%

22%

-0.025*

United Kingdom

56%

63%

69%

0.023*

51%

59%

62%

0.021*

50%

49%

45%

-0.012*

United States

78%

80%

84%

0.016*

71%

80%

87%

0.035*

51%

63%

71%

0.042*

Note: GCI means generalised concentration index. Q1 refers to the proportion in the first income quintile (lowest), Q5 in the highest.

Source: OECD calculations based on national health surveys.

Annex Table 3.A.9. Descriptive statistics and generalised concentration indexes: Dental visits and flu vaccination

Visited any dentist

Flu vaccination

Q1

Mean

Q5

GCI

Q1

Mean

Q5

GCI

EU28/26

51%

59%

69%

36%

37%

37%

OECD

54%

63%

72%

40%

40%

41%

Total

51%

60%

70%

38%

39%

39%

Austria

66%

72%

79%

0.026*

17%

20%

24%

0.018*

Belgium

49%

60%

69%

0.044*

60%

59%

55%

-0.010

Bulgaria

26%

45%

61%

0.069*

-

-

-

47%

66%

81%

0.069*

56%

58%

62%

0.017*

Chile

4%

6%

10%

0.010*

-

-

-

Croatia

44%

54%

68%

0.046*

25%

25%

26%

0.001

Cyprus

38%

48%

66%

0.053*

30%

33%

32%

0.012

Czech Republic

68%

76%

81%

0.029*

15%

16%

17%

0.006

Denmark

72%

81%

84%

0.020*

49%

48%

41%

-0.014

Estonia

45%

50%

63%

0.035*

2%

1%

4%

0.000

Finland

46%

57%

66%

0.038*

92%

91%

90%

-0.003

France

49%

55%

62%

0.026*

56%

55%

53%

-0.004

Germany

78%

82%

83%

0.010*

49%

48%

45%

-0.005

Greece

43%

48%

59%

0.029*

48%

52%

54%

0.016*

Hungary

31%

46%

60%

0.055*

26%

28%

29%

0.017*

Iceland

61%

70%

80%

0.041*

53%

53%

52%

0.008

Ireland

93%

93%

94%

0.003

54%

54%

55%

-0.001

Italy

35%

46%

56%

0.044*

39%

41%

39%

-0.001

Latvia

37%

49%

61%

0.046*

3%

4%

6%

0.006*

Lithuania

44%

47%

53%

0.023*

5%

5%

4%

0.001

Luxembourg

75%

79%

82%

0.017*

45%

47%

51%

0.006

Malta

42%

56%

61%

0.041*

54%

53%

49%

-0.015

Netherlands

72%

79%

88%

0.037*

76%

73%

68%

-0.019*

Norway

67%

78%

85%

0.030*

22%

24%

22%

0.006

Poland

42%

53%

68%

0.050*

5%

10%

13%

0.019*

Portugal

36%

49%

67%

0.063*

49%

48%

45%

-0.004

Romania

8%

15%

23%

0.029*

4%

6%

8%

0.010*

Slovak Republic

65%

75%

82%

0.033*

13%

14%

18%

0.009

Slovenia

47%

59%

70%

0.045*

11%

12%

15%

0.009

Spain

34%

47%

58%

0.046*

-

-

-

Sweden

71%

71%

77%

0.013*

37%

38%

39%

0.008

United Kingdom

64%

74%

80%

0.032*

80%

79%

78%

-0.004

United States

26%

41%

57%

0.069*

69%

72%

75%

0.018*

Note: In Chile, dentist visits refer to the past 3 months; Chile is not included in the average. GCI means generalised concentration index. Q1 refers to the proportion in the first income quintile (lowest), Q5 in the highest.

Source: OECD calculations based on national health surveys.

## Notes

← 1. In practice, the surveys do not allow for such a clear-cut distinction because the first visit in a year does not necessarily need to be a patient-initiated visit, and neither do we know whether subsequent visits in the same year are necessarily doctor-initiated (ibid).

← 2. Change over time is measured by computing the difference between the concentration indexes in the present study and the concentration index in the 2012 study. An increase (decrease) in inequality is assumed to be meaningful if the difference is greater (lesser) than 0.02 (-0.02).

← 3. For this discussion, countries were divided in 3 groups of relatively high/low and intermediate level of inequality based on the value of the GCI for each cancer separately.

← 4. Numerous sensitivity analyses were carried out using in particular a range of Principal Component Analyses on different sets of variables (those listed above or subsets of them, as well as adding flu immunisation, the number of GP and specialists visits). Different ways of dealing with missing information (mainly the facts that the Unites States does not distinguish GP and specialist visits) were also tested to produce another range of groupings. Chile was excluded from the final analysis: the results across grouping methods were not very stable and the fact that the reporting period for services was very different from all other countries for all services probably limits comparability. The final grouping distinguishing relatively low, medium and large is based on the average rank. These “convenient thirtiles” are based on the level of this average with minor adjustment of the boundaries to ensure countries more systematically fall into the group other analyses suggested they should fit in.

← 5. The OECD distinguishes three dimensions of quality: (a) effectiveness, which describes the health system’s ability to achieve clinically desirable outcomes; (b) safety, which is about avoiding adverse health outcomes due to health care; and (c) responsiveness, which refers to how a system treats people to meet their legitimate expectations (Carinci et al., 2015[24]).

## Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/3c8385d0-en