• Open Access

Confidence in receiving medical care when seriously ill: a seven-country comparison of the impact of cost barriers

Authors


Claus Wendt
Chair, Sociology of Health and Healthcare Systems
Department of Sociology
University of Siegen
Adolf-Reichwein-Straße 2
D-57068 Siegen
Germany
and
External Fellow, Project Director
Mannheim Center for European Social Research
University of Mannheim
A5, 6
68159 Mannheim, Germany
E-mail: wendt@soziologie.uni-siegen.de

Abstract

Objective  This paper examines how negative experiences with the health-care system create a lack of confidence in receiving medical care in seven countries: Australia, Canada, Germany, the Netherlands, New Zealand, the United Kingdom, and the United States.

Methods  The empirical analysis is based on data from the Commonwealth Fund International Health Policy Survey 2007, with nationally representative samples of adults aged 18 and over. For the analysis of the experience of cost barriers and confidence in receiving medical care, we conducted pairwise comparisons of group percentages as well as country-wise multivariate logistic regression models.

Results  Individuals who have experienced cost barriers show a significantly lower level of confidence in receiving safe and quality medical care than those who have not. This effect is most pronounced in the United States, where people who have foregone necessary treatment because of costs are four times as likely to lack confidence as individuals without the experience of cost barriers (adjusted odds ratio 4.00). In New Zealand, Germany, and Canada, individuals with the experience of cost barriers are twice as likely to report low confidence compared with those without this experience (adjusted odds ratios of 1.95, 2.19 and 2.24, respectively). In the Netherlands and UK, cost barriers are only a marginal phenomenon.

Conclusions  The fact that the experience of financial barriers considerably lowers confidence indicates that financial incentives, such as private co-payments, have a negative effect on overall public support and therefore on the legitimacy of health-care systems.

Introduction

The ability to guarantee security against life risks is a major accomplishment of modern health-care systems. As the availability of health services can be a matter of life and death, people need the security that necessary care will be provided in the case of serious illness or injury. Cost pressures and economists’ promotion of the idea of moral hazard1 have fostered the introduction and expansion of cost-sharing instruments. While we have observed a general trend of using cost-sharing measures over the last decades, substantial differences among countries remain in the form (e.g. co-payment, co-insurance, deductibles) and level of cost-sharing, as well as in the applied protection mechanisms (e.g. exemptions, reduced rates, annual caps).2,3 While many studies have investigated the effects of cost-sharing measures, both in terms of their effectiveness in reducing utilization and costs and in terms of their unintended consequences in the form of inequalities and deterioration of health status,3–5 the impact of cost barriers on trust in the health-care system has, to our knowledge, not yet been investigated.

This paper analyses the effect of cost barriers on people’s confidence in receiving safe and quality medical care when falling seriously ill and is based on data from the 2007 Commonwealth Fund International Health Policy Survey. More specifically, we investigate whether there are

  • • differences in cost barriers among countries;
  • • differences in trust levels because of cost barriers among countries; and
  • • inequalities among groups with different levels of income, health and insurance arrangements in both experienced cost barriers and trust levels.

Theoretical background

The level of confidence in receiving medical care is an important indicator of trust that people have in the medical system.6–9 Trust, which, according to Mechanic,6 is ‘the expectation that individuals and institutions will meet their responsibilities to us,’ has been analysed from different angles. Studies have focused on satisfaction with the health-care system,9–12 on trust in one’s own medical doctor,6,8,13–15 on patient preferences16–18 and on past experiences6,13 when analysing trust in the medical system. These studies have enhanced our understanding of people’s perception of health care and health-care systems in various ways. It has been shown that not only people in the United States but also those in other developed nations are increasingly dissatisfied with their health-care systems.9–12 At the same time, however, they express great trust in their personal physicians.6,8,11–13 When analysing the factors influencing trust, delayed care and unmet needs have demonstrated a negative effect on patients’ trust in a physician,14 but studies have also pointed towards trust-enhancing factors, such as good access to health care, patient-centredness and continuity of care.13,17,19

When analysing people’s confidence in receiving safe and quality medical care when falling seriously ill, we focus on questions that, to our knowledge, have been neglected in previous comparative research. Generally speaking, we can expect that those with higher educational levels and higher incomes are more likely to trust in the capabilities of health-care systems, for they gain access to general practitioners and specialists more easily and have fewer difficulties in meeting co-payments. Moreover, we expect that this ‘feeling of security’ is influenced by past experiences with the health-care system.16,20,21 Because ‘confidence in receiving medical care’ is presumably related to the overall health-care system and, in particular, co-payments, we expect that institutional structures matter. Consequently, the impact of individual socio-economic status on confidence in receiving health care should vary across health-care systems.8 We therefore compare citizens’ levels of trust in the health-care systems of Australia, Canada, Germany, the Netherlands, New Zealand, the United Kingdom, and the United States.

The selection of countries, although partly guided by data availability, has the advantage that different types of health-care systems are included in the analysis. Health-care systems have been differentiated as national health systems, social health-insurance systems and private health-insurance systems;22,23 however, this distinction provides little insight in terms of cost barriers and other issues concerning patients’ access to health-care services. Table 1 gives an overview of existing cost barriers in the seven countries under study.

Table 1. Cost barriers
CountryCoverage (%)1Cost-sharing1Out-of-pocket payments2
General practitionerSpecialistIn-patient carePharmaceuticals% of total population with no out-of-pocket payments% who pay more than 1000 USD (Commonwealth Survey)
  1. 1The OECD Health Project (2004): Towards High-Performing Health Systems, Paris: OECD.

  2. 2Calculations based on Commonwealth Survey.

Australia100For 25% of bills, average of USD 5. General patient reimbursement85% of schedule fee if not bulk billedFor 71% of bills, average of USD 8. Patient reimbursement 85% of schedule fee if referredNoneMaximum AUD 23.70 (around USD 18) per prescription for general patients for drugs on the PBS Scheme.1319
Canada100NoneNoneNoneDiscretion of provinces2112
Germany89.4 (public), 10.4 (private)Fee of 10 EUR covers all visits during the quarter. Preventive measures are exempt from practice feesPatients who are referred by one doctor to another pay no additional practice fees as long as the referral falls within the same quarterCo-payment of EUR10 per day, limited to a maximum of28 days in a calendar yearCo-payment amounting to 10% of the price, but no less than EUR 5 and no more than EUR 10 per medication910
Netherlands98.6NoneNoneNoneGenerics covered. Non-generics covered if no alternatives available385
New Zealand100Extra billingOutpatients USD 3–USD 17NoneUSD 2 – USD 8 with stop loss1210
United Kingdom100NoneNoneNoneUSD 9 per prescription; free with a ‘season ticket’ of USD 130. Many persons exempt124
United States27.4 (public) 57.9 (private)20% in excess of 100 USD deductible. Also a 66.60 USD monthly premium for coverage of physician services20% in excess of 100 USD deductible. Also a 66.60 USD monthly premium for coverage of physician services876 USD deductible for first 60 hospital days100%1030

Cost barriers consist of coverage levels and cost-sharing arrangements. Coverage represents a cost barrier because those individuals not covered by a public or private health-care scheme need to pay the total costs of their utilized health care. We can thereby see clear country differences with regard to cost barriers. Canada, the Netherlands and the UK not only provide full coverage to their citizens, but they also have no cost-sharing except for pharmaceuticals. Australia, Germany and New Zealand have moderate cost barriers. While these countries provide universal coverage (quasi-universal in Germany), they have introduced cost-sharing for out-patient care and pharmaceuticals (and also for in-patient care in Germany). While cost-sharing in these countries can represent a substantial cost barrier to access, cost-sharing levels are still significantly lower than in the United States, which has a deductible of 100 USD and a co-insurance level of 20% for outpatient care. In addition, many American citizens are without health insurance and therefore face tremendous cost barriers.

The regulations are only partly matched by out-of-pocket expenditure data. The Netherlands and Canada have a particularly high number of persons (in per cent of total population) without out-of-pocket expenditures: 21 and 38%, respectively. The other countries range between 9% (in Germany) and 13% (in Australia). When looking at the percentage of persons with a particularly high level of out-of-pocket expenditures (more than 1000 USD per year), the United States stands out with 30%, followed by Australia with 19%. The UK and the Netherlands, with 5 and 4%, respectively, have a particularly low number of individuals facing high out-of-pocket expenditures.

Based on these differences in terms of cost barriers, we would expect the Netherlands and the UK in particular to show a low number of skipped visits because of costs and therefore also high levels of confidence.23,24 Beyond mere country differences, individual characteristics should be important for the relevance of cost barriers and confidence in receiving health care. In general, it is assumed that the higher cost barriers are, the higher their regressive effect will be, meaning that persons with low levels of income might be especially prone to skipping visits.3,5 We can thus expect income to be an important predictor of confidence and skipped visits because of cost barriers in countries that have high cost barriers, such as in the United States and probably also in Australia and New Zealand. In contrast, income should play a minor role in the Netherlands and the UK. The same relationships should apply to people with poor health status.4 In systems with low cost barriers, people in poor health might not skip a doctor’s visit owing to costs, but when cost barriers are high, visits to the doctor might no longer be affordable, particularly for those with chronic conditions.

Our main theoretical argument was that earlier experiences have a strong influence on people’s trust in the future.6,13,16,20 An important negative event is represented by the experience of not going to a doctor because of costs despite having a medical condition. In a first step, we investigate the extent to which various social groups actually experience cost barriers that keep them from visiting a doctor in different institutional settings. In a second step, we analyse differences in trust levels across countries.

Methods

Sample

The empirical analysis is based on data from the Commonwealth Fund International Health Policy Survey 2007, which was conducted by Harris Interactive from March to May 2007, in seven countries with nationally representative samples of adults aged 18 and over. In Australia, 1009 people were interviewed, 1000 in New Zealand, 1407 in Germany, 1557 in the Netherlands, 1434 in the UK and 2500 in the United States.25 The final samples were adjusted to expected population distributions using country-specific weights provided by Harris Interactive. These weights included age, sex and additional variables following standards for each country. The samples thus reflect the demographic distribution of the adult population in the countries under analysis.

As we expect that the experience of financial barriers to accessing health care has a strong effect on confidence, we first examine which population groups are particularly prone to having this negative experience within their respective national health system. Second, we investigate which factors impact the lack of confidence in receiving future medical care in the seven countries. In this part, we focus on socio-economic characteristics, such as income and education, as well as on the respondents’ own experiences within the health-care system, namely financial barriers to service access.

Dependent variables

The aspect of financial barriers to the access of necessary health care was measured using the following question: ‘During the past 12 months, was there a time when you had a medical problem but did not visit a doctor because of cost?’ The lack of confidence in receiving good medical care in the future was measured with the following question: ‘How confident are you that if you become seriously ill, you will get quality and safe medical care?’ In our analysis, we focused on those respondents who reported to be ‘not very’ or ‘not at all confident.’

Statistical analysis

We conducted pairwise comparisons of group percentages using the Scheffé method for both dependent variables.26 We also conducted country-wise multivariate logistic regression models to examine the net effects of the individual level characteristics.

Study results

The experience of not going to the doctor because of cost barriers is mainly dependent on the setup of the health-care system (public vs. private insurance, level of cost-sharing) and can therefore be directly influenced by health-policy measures.

Not going to the doctor because of costs: national averages

As shown in Fig. 1, only a small minority of the population has had the experience of not going to the doctor because of cost barriers in three countries (the Netherlands, the UK and Canada). In Australia and Germany, in contrast, more than 10% of respondents have experienced cost barriers, a figure that stands at 20 and 25% in New Zealand and the United States, respectively. The analysis of inequalities with respect to cost barriers is therefore restricted to the four latter countries.

Figure 1.

 Experience of not going to the doctor because of costs, by countries, 2007. Source: Authors’ calculations using the 2007 Commonwealth Fund International Health Policy Survey.

Socio-economic factors hardly matter

In Australia and Germany, socio-economic factors do not influence the decision to not go to the doctor because of costs. In the United States, 37% of low-income earners skip a doctor’s visit because of costs compared with 15% of the high-income earners (P < 0.001). Furthermore, the United States is the only country in which less educated people skip a doctor’s visit significantly more often than those with a higher level of education (29 and 15%, respectively; P < 0.001; see Table 2).

Table 2. Percentage (N) not going to doctor because of costs, by countries and population subgroups, 2007
 AustraliaGermanyNew ZealandUnited States
  1. Source: Authors’ calculations using the 2007 Commonwealth Fund International Health Policy Survey: Data weighted. P-value in brackets.

  2. *Significantly different from reference group (P ≤ 0.05).

Total13 (952)12 (1253)20 (948)25 (2347)
Gender
 Women15 (609)12 (652)22 (573)*27 (1452)*
 (0.086)(0.532)(0.043)(0.013)
 Men (ref.)11 (343)11 (601)17 (375)22 (895)
Age
 Age 18–4918 (455)*15 (725)*26 (519)*31 (1081)*
 (<0.001)(<0.001)(<0.001)(<0.001)
 Age 50+ (ref.)5 (497)7 (528)8 (429)15 (1266)
Education
 High school13 (455)11 (781)20 (418)29 (730)*
 (0.927)(0.991)(0.697)(<0.001)
 College13 (177)15 (316)22 (253)27 (768)*
 (0.895)(0.365)(0.391)(<0.001)
 Tertiary (ref.)14 (320)10 (156)17 (277)15 (849)
Income
 Below average16 (302)14 (442)24 (240)*37 (766)*
 (0.093)(0.087)(0.037)(<0.001)
 Average15 (188)11 (289)25 (167)23 (459)*
 (0.294)(0.730)(0.057)(0.005)
 Above average (ref.)11 (462)10 (522)16 (541)15 (1122)
Health status
 Fair-poor17 (149)15 (207)22 (90)37 (417)*
 (0.163)(0.123)(0.646)(<0.001)
 Good-excellent (ref.)13 (803)11 (1046)20 (858)22 (1930)
Insurance status (Australia, Germany, New Zealand)
 Standard17 (365)*13 (1015)*27 (470)* 
 (0.003)(0.013)(<0.001) 
 Private (ref.)10 (587)7 (238)12 (478) 
Insurance status (United States)
 Employer   18 (1306)*
    (0.023)
 Other private   24 (175)*
    (0.009)
 Medicaid   37 (158)*
    (<0.001)
 Uninsured   53 (221)*
    (<0.001)
 Medicare (ref.)   11 (496)

People in poor health skip a doctor’s visit more often in the United States

The United States is the only country in which people who consider their own health to be fair or poor skip a doctor’s visit significantly more often than healthier patients (37 and 22%, respectively; P < 0.001).

Not going to the doctor because of costs is related to insurance status

The way people are covered by the health-care system matters in all countries: In Australia, Germany, and New Zealand, patients covered by the standard system skip a visit to the doctor significantly more often because of costs than those with (supplementary) private health insurance (Australia: 17 vs. 10%, P = 0.003; Germany: 13 vs. 7%, P = 0.013; New Zealand: 27 vs. 12%, P < 0.001). In the United States, we distinguished between employer-based private insurance, other private insurance, Medicare, Medicaid and being uninsured. The difference compared with those who are covered by Medicare (11% of Medicare recipients skipped a doctor’s visit because of costs) is especially large for those with Medicaid coverage (37%, P < 0.001) and the uninsured (53%, P < 0.001).

Multivariate analysis

Controlling for gender and age, we applied logistic regression analysis including income, education, health status, and insurance status (Table 3). Considering respondents’ socio-economic background, the probability of skipping a doctor’s visit does not vary with respect to education. Low-income earners skip a visit more often than respondents with a high income (except in Australia), and the difference between both groups is especially pronounced in the United States (adjusted odds ratio of 2.21 [1.52, 3.20]). Considering self-reported health, results point out that people with poor health are more likely to skip a doctor’s visit in the United States (adjusted odds ratio of 1.97 [1.37, 2.85]) and Germany (adjusted odds ratio of 1.7 [1.06, 2.72]). Furthermore, the multivariate analysis confirms that insurance status matters in all countries (not significant in Australia), the most pronounced effects being in the United States. Compared with Medicare patients, all other groups are significantly more likely to skip a visit to the doctor because of costs. Medicaid patients and the uninsured face major barriers when seeking health care (adjusted odds ratios of 2.71 [1.46, 5.05] and 5.54 [3.07, 10.01], respectively).

Table 3. Not going to the doctor because of costs, by countries and respondents’ characteristics, 2007: adjusted odds ratios from logistic regression models1
 AustraliaGermanyNew ZealandUnited States
  1. Authors’ calculations using the 2007 Commonwealth Fund International Health Policy Survey. Data weighted. P-value in brackets.

  2. *P < 0.05 **P < 0.01 ***P < 0.001.

  3. 1Odds ratios [95% confidence interval (CI)] adjusted for all characteristics listed in the left column; comparing two groups, an odds ratio of 1 indicates identical chances for both groups that an event occurs (here, skipping a doctor’s visit). An odds ratio >1 indicates a higher chance compared with the reference group, an odds ratio of <1 a smaller chance.

Female (vs. male)1.381.241.331.22
(0.266) (0.260) (0.195) (0.176)
[0.78, 2.44] [0.85, 1.80] [0.86, 2.05] [0.91, 1.64]
Age 18–49 (vs. age 50+)5.07***2.50***5.49***2.19***
(<0.001) (<0.001) (<0.001) (<0.001)
[2.74, 9.39] [1.62, 3.85][3.38, 8.90] [1.61, 2.97]
Education (vs. tertiary)
 High school0.830.921.261.09
(0.577) (0.795) (0.391) (0.669)
[0.44, 1.58] [0.51, 1.68] [0.75, 2.12] [0.74, 1.61]
 College0.751.291.221.36
(0.454) (0.425) (0.501) (0.074)
[0.36, 1.58] [0.69, 2.40] [0.69, 2.16] [0.97, 1.91]
Income (vs. above average)
 Below average1.861.481.732.21***
(0.117) (0.065) (0.058) (<0.001)
[0.85, 4.06] [0.98, 2.25] [0.98, 3.05] [1.52, 3.20]
 Average1.671.121.441.32
(0.216) (0.652) (0.206) (0.209)
[0.74, 3.79] [0.68, 1.85] [0.82, 2.52] [0.86, 2.04]
Self-reported health status (vs. excellent/good)
 Fair/poor1.551.70*1.371.97***
(0.294) (0.028) (0.342) (<0.001)
[0.69, 3.49] [1.06, 2.72] [0.72, 2.61] [1.37, 2.85]
Insurance status (Australia, New Zealand, Germany)
 Standard (vs. private)1.411.89*2.43*** 
(0.268) (0.028) (<0.001) 
[0.77, 2.58] [1.07, 3.33] [1.54, 3.83] 
Insurance status (United States)
 Employer scheme (vs. medicare)   1.69*
    (0.043)
    [1.02, 2.82]
 Medicaid (vs. medicare)   2.71**
    (0.002)
    [1.46, 5.05]
 Private plan (vs. medicare)   2.15*
    (0.027)
    [1.09, 4.22]
 Not insured (vs. medicare)   5.54***
    (<0.001)
    [3.07, 10.01]

In a second step, the lack of confidence in receiving safe and quality medical care when seriously ill is analysed.

Lack of confidence: national average

Only 5% of Dutch respondents lack confidence in the health-care system. In all other countries, between one-fifth and one-quarter of the population lacks confidence in receiving medical care when in need (Table 4).

Table 4. Lack of confidence in receiving medical care, by countries and population subgroups [in per cent of population, (N)], 2007
 AustraliaCanadaGermanyNetherlandsNew ZealandUnited KingdomUnited States
  1. Authors’ calculations using the 2007 Commonwealth Fund International Health Policy Survey. Data weighted. P-value in brackets.

  2. *Significantly different from reference group (P ≤ 0.05).

Total20 (952)19 (2726)26 (1253)5 (1422)22 (948)27 (1170)21 (2347)
Gender
 Women22 (609)*21 (1463)*32 (652)*5 (825)25 (573)31 (697)*20 (1452)
 (0.039)(0.005)(<0.001)(0.099)(0.015)(0.011)(0.499)
 Men (ref.)17 (343)17 (1263)19 (601)4 (597)18 (375)24 (473)21 (895)
Age
 Age 18–4921 (455)19 (1454)24 (725)4 (522)22 (519)27 (548)22 (1081)*
 (0.215)(0.916)(0.175)(0.664)(0.844)(0.476)(0.040)
 Age 50+ (ref.)18 (497)19 (1272)28 (528)5 (900)21 (429)28 (622)19 (1266)
Education
 High school22 (455)*23 (772)*29 (781)5 (954)26 (418)*29 (551)26 (730)*
 (0.009)(0.004)(0.135)(0.973)(0.035)(0.999)(<0.001)
 College17 (177)19 (1130)20 (316)4 (372)18 (253)24 (312)21 (768)*
 (0.595)(0.351)(0.967)(0.989)(0.999)(0.393)(<0.001)
 Tertiary (ref.)13 (320)16 (824)21 (156)4 (96)18 (277)29 (307)12 (849)
Income
 Below average26 (302)*25 (810)*29 (442)*7 (472)*29 (240)*35 (391)*29 (766)*
 (0.001)(<0.001)(0.008)(0.018)(0.004)(0.001)(<0.001)
 Average18 (188)19 (583)31 (289)*4 (248)23 (167)23 (331)19 (459)
 (0.777)(0.197)(0.004)(0.832)(0.423)(0.945)(0.190)
 Above average (ref.)15 (462)16 (1333)20 (522)3 (702)18 (541)24 (448)15 (1122)
Health status
 Fair-poor35 (149)*30 (328)*38 (207)*8 (570)*40 (90)*41 (241)*39 (417)*
 (<0.001)(<0.001)(<0.001)(<0.001)(<0.001)(<0.001)(<0.001)
 Good-excellent (ref.)17 (803)18 (2398)23 (1046)3 (852)20 (858)24 (929)17 (1930)
Experience
 Cost barriers29 (103)*38 (100)*41 (143)*16 (30)*33 (158)*52 (22)*43 (464)*
 (0.009)(<0.001)(<0.001)(0.003)(<0.001)(0.009)(<0.001)
 No cost barriers (ref.)18 (849)18 (2626)24 (1110)4 (1392)19 (790)27 (1148)14 (863)

Education and income matter

People with lower levels of education (compared with people with tertiary education) show less confidence in receiving safe and quality care (in four countries). In all countries, people with below-average incomes are less confident in receiving good medical care than those with above-average incomes.

Poor health reduces confidence

Those in poor health show much lower levels of confidence in receiving medical care than their healthier counterparts in all countries. In Germany, New Zealand, the UK and the United States, about 40% of those who consider their own health to be fair or poor are either not very or not at all confident that they will receive safe and quality medical care when seriously ill (see Table 4). The gap between groups with better and poorer health is greatest in the United States (39 and 17%, respectively, P < 0.001), followed by New Zealand (40 and 20%; P < 0.001).

Cost barriers have a substantial effect

In all countries, the experience of not going to the doctor because of costs has a strong negative effect on confidence. This factor has the most striking effect in five of seven countries. In the United States, the gap between patients with and without the experience of cost barriers is almost 30 percentage points (43 and 14%, respectively, P < 0.001).

Multivariate analysis

The bivariate results were generally confirmed in logistic regressions that – controlling for gender and age – included income and education, health status and the experience of cost barriers (see Table 5). While education seems to be of lesser importance (except in the United States), income still matters after controlling for other factors (in Canada, Germany, and the UK). Beside an inferior health status, the experience of not going to the doctor has a substantial negative effect on confidence in the overall system’s capacity to deliver health-care services when necessary (not significant in Australia and the Netherlands).

Table 5. Lack of confidence in receiving medical care, by countries and respondents’ characteristics, 2007: adjusted odds ratios from logistic regression models1
 AustraliaCanadaGermanyNetherlandsNew ZealandUnited KingdomUnited States
  1. Authors’ calculations using the 2007 Commonwealth Fund International Health Policy Survey. Data weighted. P-value in brackets. *P < 0.05 **P < 0.01 ***P < 0.001.

  2. 1Odds ratios [95% confidence interval (CI)] adjusted for all characteristics listed in the left column; comparing two groups, an odds ratio of 1 indicates identical chances for both groups that an event occurs (here: the lack of confidence). An odds ratio >1 indicates a higher chance compared with the reference group, an odds ratio of <1 a smaller chance.

Female (vs. male)1.431.29*1.83***1.421.431.360.83
 (0.157) (0.014)(<0.001) (0.429) (0.072) (0.066) (0.191)
 [0.87, 2.36] [1.05, 1.58] [1.39, 2.41] [0.60, 3.36] [0.97, 2.10] [0.98, 1.87] [0.62, 1.10]
Age 18–49 (vs. age 50+)1.511.131.031.091.181.081.13
 (0.098) (0.247) (0.844) (0.813) (0.422) (0.651) (0.407)
 [0.93, 2.46] [0.92, 1.38] [0.77, 1.37] [0.52, 2.30] [0.79, 1.75] [0.77, 1.52] [0.84, 1.52]
Education (vs. tertiary)
 High school1.601.231.130.891.410.741.79**
 (0.106) (0.147) (0.573) (0.872) (0.174) (0.168) (0.003)
 [0.90, 2.84] [0.93, 1.64] [0.73, 1.75] [0.20, 3.90] [0.86, 2.32] [0.48, 1.14] [1.23, 2.62]
 College1.341.070.771.020.870.651.54*
 (0.376) (0.623) (0.297) (0.983) (0.622) (0.055) (0.018)
 [0.70, 2.53] [0.83, 1.37] [0.48, 1.25] [0.25, 4.17] [0.50, 1.51] [0.42, 1.01] [1.08, 2.22]
Income (vs. above average)
 Below average1.501.45**1.42*1.591.401.70*1.20
 (0.185) (0.004) (0.025) (0.308) (0.154) (0.019) (0.324)
 [0.82, 2.73] [1.13, 1.86] [1.05, 1.94] [0.65, 3.87] [0.88, 2.22] [1.09, 2.65] [0.84, 1.72]
 Average1.101.131.61**1.061.230.960.97
 (0.799) (0.389) (0.007) (0.906) (0.483) (0.852) (0.874)
 [0.54, 2.21] [0.86, 1.48] [1.14, 2.28] [0.38, 3.02] [0.69, 2.19] [0.63, 1.47] [0.63, 1.47]
Self-reported health status (vs. Excellent/good)
 Fair/poor2.48**1.74***1.75**2.51*2.38**2.04***2.38***
 (0.002) (<0.001) (0.001) (0.011) (0.003) (<0.001) (<0.001)
 [1.41, 4.36] [1.32, 2.29] [1.25, 2.46] [1.23, 5.10] [1.35, 4.19] [1.40, 2.97] [1.69, 3.36]
Experience (vs. no cost barrier)
 Cost barrier1.522.24***2.19***2.221.95**2.324.00***
 (0.201) (<0.001) (<0.001) (0.166) (0.005) (0.059) (<0.001)
 [0.80, 2.91] [1.45, 3.47] [1.49, 3.20] [0.72, 6.87] [1.22, 3.12] [0.97, 5.54] [2.91, 5.49]
N 95227261253142294811702347
Pseudo R20.0510.0250.0510.0450.0470.0370.120

Discussion

Our results indicate a good match between cost-sharing regulations in different countries and experienced cost barriers. In the Netherlands, the UK and Canada, only a very small percentage skipped a doctor’s visit because of costs, whereas a quarter of the respondents in the United States did so. The experienced cost barriers not only affect the utilization of health care but also people’s trust in the system. We found a substantial effect of experienced cost barriers on people’s confidence in receiving medical care. While the overall level of confidence in medical care is rather high (between 75 and 95%), cost barriers seem to play an important role in country variation in confidence levels.

Furthermore, it was hypothesized that the different institutional arrangements, particularly cost-sharing instruments of the respective health-care systems, affect socio-economic differences in reaction to cost barriers and confidence in receiving care. We find that a low income significantly reduces confidence in receiving medical care. Those who consider their own health as fair or poor are more pessimistic about receiving health care when in need compared with people in better health. This group difference was most pronounced in the United States and New Zealand, countries with especially high co-payments, and might prove a barrier to care for those who need it most often. Because people in all countries who have skipped a doctor’s visit because of costs show lower confidence, it can be argued that financial-barrier experiences have a strong negative influence on trust in the overall health-care system.

Controlling for other factors, lower-income groups in New Zealand, the United States and Germany face more cost barriers than do higher-income groups, and patients with a poorer health status skip a doctor’s visit more often in the United States and Germany. Furthermore, insurance status matters in all countries. People without (supplementary) private insurance skip a visit more often in New Zealand and Germany. In the United States, Medicaid patients and the uninsured are particularly prone to not go to the doctor. Compared with Medicare patients, those with employer plans and other private plans are significantly more likely to not go to a doctor because of costs despite having a medical condition.

Our analysis contributes to the discussion of ‘trust in the medical system’6–9 in the following respect: Patients need to feel secure that they will receive health care when in need. This does not always occur, even if people trust their personal physician. In certain circumstances, patients must rely on doctors whom they have not met before. The question of confidence in receiving quality and safe medical care when one becomes seriously ill might therefore represent the missing link between ‘trust in one’s personal doctors’ and ‘satisfaction with the health-care system’ and seems to be especially suited to capturing trust in the health-care system’s ability to guarantee security in the case of illness.

Limitations

In this study, we cannot control for the severity of the condition despite which a person decides not to see a doctor because of costs. A study by Schoen and colleagues, however, provides some indication of the impact of severe conditions, for it focuses on people with serious chronic diseases. According to the results of the study, a substantial portion of this specific group reports access difficulties.21 A further limitation to our study is that we cannot be certain about the causal direction of the relationship between confidence and skipped visits. Even though costs are the major reason for the skipped visits, people with low levels of confidence in the system might react especially strongly to co-payments. With regard to institutional factors, we have only used insurance status to explain which people did not go to the doctor because of costs. Other factors, such as the perceived quality of care or waiting times, which we did not enter into our equation, might also influence the propensity to not see a doctor in spite of a health condition. In the UK, for instance, existing socio-economic differences in the lack of confidence might be more strongly related to quality issues than to cost barriers.

Policy implications

Our analysis shows major disparities in all included countries among social groups regarding the confidence in receiving medical care when in need. Given the high emphasis of equality, especially in the British NHS, these differences in confidence require further elaboration. Our findings cannot provide direct lessons as to how the disparities between education and health groups could be reduced; however, the fact that the experience of financial barriers considerably lowers confidence has important health-policy implications.

Although those who did not go to the doctor because of costs have less confidence in all countries, we could show that in Canada, and particularly in the UK and the Netherlands, only a few people have experienced cost barriers at all. This phenomenon could be partly related to the fact that visits to a doctor are free and that other private out-of-pocket payments have also been particularly low for many years in both countries while existing at a high level in the United States, Australia and New Zealand.27 Canada’s health-care system, however, creates low cost barriers, although out-of-pocket payments are higher than in Germany. The type of co-payment, as well as the time of the introduction of co-payments, might play a role in establishing cost barriers. In Germany, for instance, a payment of ten Euros per quarter for a doctor’s visit was introduced in 2004. This fee seems to have a stronger impact on patients’ decision to not see a doctor than do co-payments that were introduced earlier in other countries.

When analysing group-specific risks of facing cost barriers in more detail, we find that socio-economic factors hardly matter. But insurance status does. In Australia, Germany and New Zealand, the chances of experiencing cost barriers are significantly reduced when patients subscribe to a supplementary private insurance. Even if the difference between patients with and without private insurance in Australia and Germany is modest, the inequalities between people with and without (supplementary) private insurance might indicate that health-care systems with universal coverage also face problems in guaranteeing equal access to necessary health care.

The greatest inequalities across groups with different insurance statuses can be found in the United States. While the deprived position of the uninsured has been shown elsewhere,20 our results indicate that Medicare is better suited for protecting patients from access problems than are employer-based plans, other private insurance plans or Medicaid. Although cost barriers should be especially low for Medicaid patients, our results show that this is not the case. The experience of financial barriers might be related to the fact that medical doctors often do not accept Medicaid patients, and this group might therefore face additional costs (e.g. travel costs) when visiting a doctor. Concerning the ‘promise of security’ against major life risks, Medicare seems to be closer to other nations’ public health systems than alternative US plans are. In contrast to Medicare, however, the public schemes of other countries provide coverage for the entire population, not only for a small subgroup.

Our findings indicate that the trend towards increased cost-sharing of patients has more unintended consequences than might have been previously realized. Beside the risk of an increased burden on low-income groups and seriously ill persons, experienced cost barriers also affect trust in the health-care system. As trust in the health-care system is not only important for the legitimacy of the system but also for treatment outcomes, we would suggest that policy makers consider the negative impact on trust when dealing with cost-sharing. Guaranteeing access to necessary health-care services for those who fall seriously ill can be considered the single most important task of health-care systems in modern societies, and cost barriers therefore represent a major health-policy issue.

Acknowledgements

The research reported here has received financial support from the Harkness Program of Health Policy & Practice of the Commonwealth Fund and the Bosch Foundation. A first version of this article was presented at The Commonwealth Fund 2008–09 Final Conference in 2009, and we gratefully acknowledge the helpful comments and criticism by the participants, particularly Brad Gray, Rubin Minhas, and Jako Burger as well as by Ted Marmor, Bob Blendon, Jason Beckfield, and three anonymous reviewers.

Ancillary