- Email: [email protected]
Inequities in access to medical care in five countries: findings from the 2001 Commonwealth Fund International Health Policy Survey
Health Policy 67 (2004) 309–322
Inequities in access to medical care in five countries: findings from the 2001 Commonwealth Fund International Health Policy Survey Cathy Schoen∗ , Michelle M. Doty1 Health Policy, Research and Evaluation, The Commonwealth Fund, One East 75th Street, New York, NY 10021, USA Accepted 16 September 2003
Abstract Objective: To examine across five countries inequities in access to health care and quality of care experiences associated with income, and to determine whether these inequities persist after controlling for the effect of insurance coverage, minority and immigration status, health and other important co-factors. Design: Multivariate analysis of a cross-sectional 2001 random survey of 1400 adults in five countries: Australia, Canada, New Zealand, United Kingdom, and United States. Main outcome measures: Access difficulties and waiting times, cost-related access problems, and ratings of physicians and quality of care. Results: The study finds wide and significant disparities in access and care experience between US adults with above and below-average incomes that persist after controlling for insurance coverage, race/ethnicity, immigration status, and other important factors. In contrast, differences in UK by income were rare. There were also few significant access differences by income in Australia; yet, compared to UK, Australians were more likely to report out of pocket costs. New Zealand and Canada results fell in the mid-range of the five nations, with income gaps most pronounced on services less well covered by national systems. In the four countries with universal coverage, adults with above-average income were more likely to have private supplemental insurance. Having private insurance in Australia, Canada, and New Zealand protects adults from cost-related access problems. In contrast, in UK having supplemental coverage makes little significant difference for access measures. Being uninsured in US has significant negative consequences for access and quality ratings. Conclusions: For policy leaders, the five-nation survey demonstrates that some health systems are better able to minimize among low income adults financial barriers to access and quality care. However, the reliance on private coverage to supplement public coverage in Australia, Canada, and New Zealand can result in access inequities even within health systems that provide basic health coverage for all. If private insurance can circumvent queues or waiting times, low income adults may also be at higher risks for non-financial barriers since they are less likely to have supplemental coverage. Furthermore, greater inequality in care experiences by income is associated with more divided public views of the need for system reform. This finding was particularly striking in Canada where an increased incidence of disparities by income in 2001 compared to a 1998 survey was associated with diverging views in 2001. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Health; Policy; International; Inequality; Access; Insurance
∗
Corresponding author. Tel.: +1-212-606-3864; fax: +1-212-606-3508. E-mail addresses: [email protected] (C. Schoen), [email protected] (M.M. Doty). 1 Tel.: +1-212-606-3860; fax: +1-212-606-3508.
0168-8510/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.healthpol.2003.09.006
310
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
1. Introduction Inequities in access to medical care by income can contribute to and exacerbate disparities in health and quality of life. Conversely, policies that seek to promote equity in access for lower income individuals offer the potential of moderating underlying health differences and, over the longer term, providing more equal opportunities for health and productivity. Countries vary widely in the extent to which public financing of health care and rules governing insurance markets seek to promote equity in access and health care experiences across income classes. Although the United States stands out among developed nations at one end of the spectrum for its lack of universal coverage, countries with universal coverage also differ in benefit design, patient-cost-sharing, and the role of private insurance. Non-financial barriers (such as waiting lists or queues, community shortages, complexity) as well as financial barriers may contribute to inequities in access. If private insurance enables more affluent families to have more ready access to care or to purchase a different standard of care, access inequities by income may emerge even where few financial barriers to health care exist. In an era of budgetary constraints and often rapid changes in public health care policies, cross-national comparisons of access and care experiences by income offer the opportunity for countries to learn from each other and to assess relative performance. Yet, such comparative studies have been relatively rare due to lack of common access measures beyond physician visits or other measures of service use, and the lack of international studies using similar data-bases and time periods [1]. To assess access experiences and variations by income, the 2001 Commonwealth Fund International Health Policy Survey interviewed adults in five countries—Australia, Canada, New Zealand, United Kingdom and United States—using an array of questions on difficulties accessing medical care, financial barriers to obtaining needed health care, physician quality ratings, and overall health care system views. Previous descriptive analysis of this survey explored the extent to which experiences varied by income, without taking into account the effects of important confounding factors such as health insurance coverage, health status, education, race/ethnicity, immigra-
tion status, and residential location [2]. Each of these confounding factors can vary across income groups within each country as well as influence access and quality of care. Thus, failure to control for these effects could potentially bias any observed income-related inequities. In this paper, we re-examine the relative importance of low income and its association with access and care experiences disparities in Australia, Canada, New Zealand, United Kingdom, and the United States controlling for the effects of health insurance, health status, education, race/ethnicity, immigration status, and residential location. In addition, we examine the role of health insurance in minimizing income-related access and quality of care disparities. The analysis focuses on four central questions: How do access, cost, and adults’ perceptions of quality vary between low and higher income adults within each country? To what extent are access inequities by income attenuated after adjusting for insurance, health, and other characteristics likely to influence care experiences? In countries with universal coverage, how do access and care experiences vary depending on whether or not adults have private insurance in addition to public coverage? Do system views vary by income, and if so, are divergent system views more likely in countries with inequities in care experiences?
2. Methods and data Data come from the 2001 Commonwealth Fund International Health Policy Survey, a five-nation survey consisting of interviews with a random sample of approximately 1400 adults age 18 and older in each of five countries: Australia (1412), Canada (1400), New Zealand (1400), United Kingdom (1400), and United States (1401). Conducted by telephone during April and May of 2001 by Harris Interactive, the survey explored problems accessing medical care, quality ratings, and adults’ views of their health care system. Details of the survey methodology are reported elsewhere [3]. The following describes key measures used in the analysis and the general analytical approach. 2.1. Relative income status Income status is the main independent variable of interest. To gauge relative income levels in each coun-
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
try, the survey quoted the national median household income in the specific country (Australia $ 38,000 Australian dollars, Canada $ 40,200 Canadian dollars; New Zealand $ 33,000 NZ dollars, UK £ 23,000, and US$ 37,000) and asked respondents whether their own income was much or somewhat below this amount, about average, or much or somewhat above. In general, this technique for collecting income information resulted in few missing observations; only between 3 and 7% of the sample in each country did not respond to the income question. Respondents who did not provide responses were dropped from the analysis.2 The analysis groups adults into three income categories: “below-average”, “average”, or “above-average” median household income (Table 2). 2.2. Health status and insurance The survey asked two questions about health status, self-reported health status and the presence of a chronic disease, injury, or disability. In the analysis, we used both questions to classify a respondent as “sicker” if he or she rated their health as fair or poor health status or said yes to a serious or chronic illness, injury, or disability that required a lot of medical care. In each of the four countries with universal coverage, the survey asked whether respondents had private health insurance in addition to public or national health coverage. In US, respondents were asked whether they had private or public coverage, or were uninsured. 2.3. Race/ethnicity and foreign-born In US, sample sizes were large enough to categorize respondents by race/ethnicity according to self-reported group membership—Latino/Hispanic, black, white, Asian and other. In the remaining four countries, due to small sample sizes, a dichotomous minority status variable was created using self-reported group membership and an additional question, “do you consider yourself to be a member of minority group?” For each country, respondents 2 This translates into 92 dropped cases in Australia, 66 in Canada, 47 in New Zealand, 57 in UK, and 48 in US. Excluded cases were not significantly different from those reporting their relative income with respect to education, minority and immigration status.
311
who reported that they were a member of a minority group or who indicated they were of Aboriginal, Asian, Central/South American, or of Middle Eastern descent in Australia; Aboriginal, black, Asian, South/Central American or Middle Eastern descent in Canada; Maori or Asian/Pacific Islander in New Zealand; and black, Asian/Pacific Islander in UK were considered minorities. In the absence of having a variable which measures language proficiency, we use foreign-born as a proxy measure. In all countries, respondents who indicated that they were born in another country were categorized as foreign-born. 2.4. Education A four category educational variable was created for each country. In regressions, the equivalent of high school graduate in US was used as the referent category in all country models. 2.5. Access to health care The survey included five measures of access not explicitly related to cost. Three items asked about difficulties getting care when needed: seeing a specialist, getting care in the evenings or weekends, and getting care where they live. Two items asked about waiting times to get care, including days waited to see a doctor when sick and waiting times for elective surgery among respondents with surgery in the past 2 years. 2.6. Cost-related access problems and financial burdens The survey includes four questions about going without needed care within the past 12 months because of costs: (1) not filling a prescription; (2) having a medical problem but not seeking doctor or medical care; (3) having a dental problem but not seeing a dentist; and (4) skipping recommended medical tests, treatment, or follow-up. The survey also asked respondents to estimate actual out of pocket costs during the year. In the analysis, all currency values were converted to US dollars using currency exchange rates current at the time of the survey. We use these out of pocket cost estimates as background information on the extent to which patients are paying directly for care. Respondents also indicated whether
312
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
or not in the past 12 months they had any problems paying their medical bills. 2.7. Quality of care and physician ratings Seven ratings of quality of care (rated as excellent or very good) are included—overall quality of care received in the past 12 months and six aspects of physician care: treating you with dignity and respect; listening to your questions and health concerns; providing you with the information you want; spending enough time with you; being accessible by phone or in person, and being aware of relevant family history as it pertained to your health. 2.8. Health care system views The survey also gauged adults’ views of their health care system by asking about the perceived need for reform: whether they believed their country’s health care system worked well with only minor changes, needed major changes, or required complete rebuilding. 2.9. Analytic approach To provide a country context, we describe and compare for each country insurance coverage and insurance policies, including benefits included in national public systems. Next we present survey findings overall and by income without adjusting for confounders. Then, we present a series of within country logistic regressions which assess the independent and main effects of income, insurance coverage, and health status on access and care experiences, controlling for
race/ethnicity or minority status, immigration status, education, age and residential location (rural, suburban or urban). In the analysis, data are weighted to reflect the demographic composition of each country and are analyzed using STATA statistical software. Tables report Chi-square tests of significance and log-likelihood at P < 0.05.
3. Results 3.1. Country context: benefit coverage and role of private insurance The health insurance systems in the five countries included in the study differ substantially with respect to the role of private insurance, the extent of patient cost-sharing, exposure to out of pocket costs for medical bills, and the range of benefits covered by public insurance systems. US stands out as the one country that does not provide universal coverage for at least a core set of benefits and for its reliance on a mixed system of voluntary private insurance and public coverage for the elderly and some of the poor. Based on national census data, as of 2001 nearly one out of five US adults (18%) under age 65 were uninsured. The four other nations in the survey provide universal public coverage yet differ in the scope of benefits, cost-sharing, and prevalence and the role played by supplemental private insurance (Table 1). Australia and New Zealand both include patient cost-sharing within their basic public systems, including fees for
Table 1 Insurance and cost-sharing policies in four countries with universal coverage Insurance and cost-sharing policies
Australia
Canada
New Zealand
United Kingdom
Private insurance for services covered by public Population with private coverage (%)
Permitted
Prohibited
Permitted
Permitted
33
Majority excluded benefits only None for basic
33
12
Co-payments for many services Covered School children
None for basic
Public plan patient cost-sharing Prescription drugs Dental
On most services balance billing allowed by doctors Covered Covered
Not covered Not covered
Covered Covered
Source: G. Anderson, Multinational Comparisons of Health Care: Expenditures, Coverage, and Outcomes, The Commonwealth Fund (New York, NY): Oct. 1998.
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
physician care and other services. Although Canadian health care benefits vary somewhat by Province, Canadian benefits for physicians and hospitals are the same across provinces with no patient fees at the point of services for doctors or hospital costs. Similarly, UK has few or no patient fees at the point of services for most benefits included in public coverage. However, Canada’s core benefits exclude prescription and dental benefits while UK includes these benefits. Canadian insurance rules restrict use of private insurance to benefits not otherwise covered by the public program or for out of country care [4]. The other three countries with universal coverage allow use of private insurance for access to private physicians or hospitals, more ready access to care, and in the case of Australia and New Zealand, to cover patient cost-sharing obligations. The survey finds that in the countries with universal coverage adults with higher incomes are much more likely to report private supplemental coverage than are adults with below-average income (Table 2). Overall, adults in UK were the least likely to have additional private health insurance—rates were much lower across income groups than the other three countries with universal coverage. Canada, on the other hand, had the highest rate (61%) of private coverage. In Canada, although not universal across Provinces, supplemental coverage generally includes benefits such as dental and prescription drug benefits that are not covered through province public health insurance. In US, low income adults were much more likely to be uninsured and less likely to have private coverage (rely on public only) than were adults with above-average incomes. Due to variations in public insurance policies, including the scope of benefits covered by core public programs, the five countries differ substantially in the extent to which residents are exposed to out of pocket costs for medical bills. US stands out for high out of pocket medical expenses—42% of US respondents estimated they paid US$ 500 or more out of pocket in the past year compared to only 5% of adults in UK (Table 2). UK stands out for extensive protections against out of pocket costs—43% of adults reported no out of pocket expenses during the year. Canadians were also relatively well protected with one-third of adults reporting no out of pocket costs in the past year. Reflecting national coverage and benefit poli-
313
cies, adults in Australia and New Zealand were more likely than those in Canada or UK to encounter costs when seeking care. Yet the level of medical bill exposure was well below that experienced in the United States. 3.2. Access problems and waiting times to get care when needed Despite having quite different health care systems, the survey found that adults in all five countries were similarly likely to report that they found it extremely or very difficult to see a specialist when needed, ranging from a low of 11% in New Zealand to a high of 17% in US (Table 3). In Australia, Canada, and UK, adults were about as likely (approximately 12%) to say it was difficult to get care at night or on weekends other than an emergency room. In contrast, only 6% of adults in New Zealand had difficulty getting care on nights or weekends. Less than 10% of adults in all five countries reported they were often unable to get care where they live. Income disparities emerged on difficulty seeing a specialist when needed (in all countries except Australia), getting care at night or on weekends (in Australia, Canada, and US), and the ability to get care where one lives (in New Zealand and US) (Table 3). The income gap was consistently widest in US—lower income adults were two to three times more likely than higher income adults to report access problems. The amount of time adults spent waiting to see a physician when sick, or scheduling necessary elective surgery among respondents reporting surgery in the past 2 years varied widely between countries (Table 3). Adults in Australia and New Zealand cited more rapid access to their physicians with two-thirds saying they were able to see their doctor the same day and few reporting waits of 5 days or more. In contrast, only one-third of US and Canadian adults and two of five UK adults said they could get in the same day; waiting times of 5 days or more were more frequent for adults in these countries. Among those needing elective surgery in the past 2 years, approximately 20% of adults in all countries (except in US) experienced long wait times of 6 months or more. Among the five nations, UK adults were the most likely to report long waits for elective surgery—27% of adults in UK
314
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
Table 2 Select demographic characteristics by country Demographic characteristics
Australia
Canada
New Zealand
United Kingdom
United States
36 19 39 7
34 22 39 5
27 21 48 4
40 23 33 4
32 21 43 3
20 26
16 22
12 24
20 20
22 27
35
30
28
30
36
47 25
37 24
40 21
41 20
49 27
– 48 52
– 39 61
– 55 45
– 77 23
16 8 76
33 68
36 79
24 60
10 44
53 90
– –
– –
– –
– –
29 7
Race/ethnicity or minority status Member of minority or racial/ethnic group White Black Hispanic Other
15 – – – –
20 – – – –
24 – – – –
14 – – – –
– 73 12 11 5
Immigration status Foreign-born
21
24
21
11
10
Out of pocket medical costs in past year None US$ 1–499 US$ 500 or more
4 51 19
35 46 12
6 65 16
43 46 5
7 39 42
(%)a
Income levels Below-average Average Above-average Missing
Health status Self-reported fair or poor health Illness/disability requiring serious medical care in past 2 years Sicker adults (either fair/poor health or disabled) Percent “sicker,” by income Below-average income Above-average income Health insurance coverage Uninsured Public coverage only Private coverageb Percent with private coverage, by income Below-average income Above-average income Percent uninsured in US, by income Below-average income Above-average income
Source: Commonwealth Fund 2001 International Health Policy Survey. a Survey respondents were provided the national median income and asked to categorize their own income in relation to the country’s average. b In UK, New Zealand, Australia, and Canada private coverage is in addition to national public coverage.
waited 6 months or more for elective surgery. In contrast, long waiting times in US were rare—only 2% of adults who needed elective surgery had to wait 6 or more months. Waiting times to see a physician when sick varied by income in US for both short and long waits. In Canada, low income adults were also more likely than
higher income adults to report waiting 5 days or more. Waiting times for elective surgery did not vary significantly by income. Among adults reporting surgery within the past 2 years, lower income adults in all five countries were no more likely than those with higher incomes to experience long waiting times for elective surgery.
Table 3 General and cost-related access and medical bill problems, and overall quality and physician ratings, by country and income levels, 2001 unadjusted percentages Australia
General access problem measures Extremely or very difficult to see a specialist when needed Very difficult to get care on weekends or evenings Often not able to get appropriate care where you live
Canada
New Zealand
Total
Belowaverage
Aboveaverage
Total
Belowaverage
Aboveaverage
Total
Belowaverage
12
14
11
16
20∗
14
11
21∗
12
15∗
10
15
19∗
13
5
3 ∗
United Kingdom Aboveaverage
Total
Belowaverage
6
13
16∗
5
12
11
United States Aboveaverage
Total
Belowaverage
Aboveaverage
9
17
30∗
8
13
18
26∗
16
∗
4
7
5
8
9
6
6
8
5
4
4
4
7
12
Time waited for doctor’s appointment when sick Same day 1–2 days 3–4 days Five days or more
62 27 6 5
58∗ 28 7 6∗
65 27 4 3
35 28 11 23
35 22 11 27∗
36 31 11 20
69 25 3 1
70 24 2 2
68 25 4 1
42 28 12 16
41 29 12 15
39 27 13 17
36 33 10 18
30∗ 33 10 21∗
40 33 9 16
Of those needing elective surgery in the past 2 years, time waited for it Less than 2 weeks 2 weeks, less than 1 month 1 month, less than 6 months 6 months or more
34 17 31 19
28 15 34 23
39 13 32 16
20 17 44 18
27 14 37 19
19 18 44 18
27 16 37 17
22 11 43 24
30 18 36 14
26 12 32 27
21 6∗ 42∗ 29
27 17 28 25
37 26 30 2
38 24 34 3
38 28 27 1
19 15
21 17
18 14
13 6
22∗ 9∗
7 4
15 14
20∗ 18∗
11 11
7 2
7 4∗
7 1
26 22
39∗ 36∗
18 14
11 33 11
14 38∗ 17∗
10 31 8
5 26 7
9∗ 42∗ 14∗
3 15 3
20 37 12
24∗ 40 20∗
18 36 7
3 19 3
4 20 4∗
2 19 2
24 35 21
36∗ 51∗ 35∗
15 24 11
63
64
59
54
51∗
60
67
66
70
53
56∗
45
57
45∗
65
80 73 69 63 72 59
82 76 71 66 75 60
80 74 68 63 72 57
79 74 62 59 67 55
80 76 61 60 69 54
81 77 65 61 69 59
84 75 71 67 73 64
84 74 70 65 72 63
85 77 72 68 74 64
73 67 54 51 58 48
73 67 55∗ 55∗ 58 51∗
70 63 49 43 55 41
72 65 58 57 63 52
61∗ 56∗ 46∗ 46∗ 53∗ 43∗
81 73 67 63 71 58
Cost-related access problems Did not do the following due to cost in the past year Did not fill a prescription Did not get recommended test, treatment, or follow-up Had medical problem but did not see doctor Needed dental care but did not see dentist Had problem paying medical bills Overall quality and physician ratings Excellent/very good rating on overall medical care received past year Physician ratings (excellent or very good on) Treating with dignity and respect Listening carefully health concerns Spending enough time with you Knowing you or your family situation Providing you with all information you want Being accessible by phone or in person
315
Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ Indicates significant difference from above-average income at P < 0.05 or better.
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
6
316
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
3.3. Access problems due to cost and medical bill burdens Overall, UK adults were the least likely and US adults were the most likely to report access problems due to cost or medical bill problems (Table 3). In Canada (5%) and UK (3%), few adults reported cost barriers to physician care. Similarly small proportions of adults in Canada (7%) and UK (3%) faced problems paying their medical bills. Dental care emerged as the only measure for which delays or foregone care due to costs in US rivaled levels reported in some other countries (e.g. Australia and New Zealand). In UK, gaps between income groups were generally small, and few adults—either high or low income—reported access problems due to cost. In contrast, the gap by income was significant and particularly wide in US on all five cost-related access problems and financial burdens measures, in New Zealand on four of five measures, and in Australia on two of five measures. In Canada, access inequities by income were widest on not filling prescriptions and forgoing needed dental care—the two benefits not included in the basic public program. Income gaps also existed in Canada for physician care and follow-up tests but, in general, few Canadians reported cost-related problems for these services. 3.4. Quality of care and physician ratings Ratings of overall quality of care and physician– patient dimensions of care were generally highest in Australia and New Zealand and lowest in UK and US; Canadian ratings were in between (Table 3). Although physician ratings for the total population were among the lowest in UK, in comparisons of quality ratings by income, UK emerged as the one country in which ratings of care were, if anything, more positive among lower income adults than among adults with above-average incomes. On four out of seven quality measures (overall quality of care, time spent with the doctor, having access to doctor by phone or in person, and knowing the patient well), a significantly larger share of UK adults with lower incomes rated their care as excellent or very good than those with higher incomes. There were no instances where quality ratings for low income UK adults were more negative.
In contrast, in the US, significant income disparities emerged on all quality and physician ratings measures with lower income adults significantly less likely to give positive ratings (Table 3). The United States was unique on these ratings disparities. In Australia, Canada, and New Zealand ratings did not vary significantly by income—lower income adults were generally as likely to rate care positively as were higher income adults. 3.5. Adjusted estimates of income and insurance-related disparities: multivariate results The country pattern of disparities by income generally persisted after controlling for the independent effects of insurance, health status, race/ethnicity or minority status, immigration status, education, residential location, and age (Tables 4–6). In logistic regressions, income gaps tended to narrow compared to unadjusted odds ratios.3 In US, regardless of insurance status, race/ethnicity, and immigration status income-related disparities remained wide and significant suggesting that income plays a dominating role in either facilitating or hampering access to the US health care system. Notably unique to US, there was an income effect for quality of care experiences, even after controlling for insurance, race/ethnicity, immigration status, and other demographic factors. Other studies in US have documented that minorities, immigrants, and those with lower education often report more negative health care experiences [5]. With respect to physician or quality ratings, however, there were no statistically significant differences between whites, Hispanics or blacks after controlling for income, insurance, and other demographic factors. In fact, in the multivariate analysis race/ethnicity was significant on only 3 out of 17 measures relating to access (seeing a specialist, care on weekends or nights, and dental access). Similarly, education was not statistically significant in any multivariate model once income, race/ethnicity, immigration status and other demographic factors were taken into account.4 In US, minorities are likely to be low income, have lower education attainment, and lack insurance which contribute to inequities in US. 3
Full results are available from authors. Full results for all parameter estimates are available from authors. 4
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
317
Table 4 The effects of income, insurance, and health status on general access problems by country, based on logit regressions Adjusted odds ratio Australia
New Zealand
United Kingdom
United States
Extremely or very difficult to see a specialist when needed Average income 1.0 1.1 Below-average income 1.2 1.5∗ Additional private insurance 0.5∗ 1.1 US no insurance – – US public insurance – – Sicker 1.4∗ 1.1
1.5 3.8∗ 0.5∗ – – 1.9∗
1.5 1.7∗ 0.6 – – 2.2∗
1.6 3.0∗ – 3.9∗ 1.0 1.8∗
Very difficult to get care on weekends Average income Below-average income Additional private insurance US no insurance US public insurance Sicker
1.1 1.5∗ 1.0 – – 1.1
0.7 0.6 1.0 – – 2.1∗
0.8 0.9 1.1 – – 1.2
0.6∗ 1.3 – 2.5∗ 0.7 2.4∗
Often not able to get appropriate care where you live Average income 0.9 Below-average income 1.5 Additional private Insurance 0.8 US no insurance – US public insurance – Sicker 1.1
1.2 1.3 0.8 – – 1.0
0.8 1.7 1.0 – – 1.9∗
1.0 0.7 1.1 – – 2.5∗
1.7 2.0 – 1.5 1.7 1.7
Waited 5 days or more for a doctor’s appointment Average income 0.7 Below-average income 1.9 Additional private insurance 1.5 US no insurance – US public insurance – Sicker 2.4∗
1.0 1.4∗ 1.0 – – 0.9
1.7 1.7 0.7 – – 1.2
0.8 0.9 1.0 – – 1.5∗
0.7 1.1 – 1.3
Waited 6 months or more for elective Average income Below-average income Additional private insurance US no insurance US public insurance Sicker
0.8 1.0 1.2 – – 1.8∗
1.6 1.1 0.2∗ – – 1.3
1.1 1.0 0.5∗ – – 1.3
1.0 1.6 – 4.0
or evenings 0.8 1.2 0.7 – – 1.9∗
surgery 0.5 1.0 0.2∗ – – 1.0
Canada
1.7∗
1.4
Logistic regressions also control for age, education, minority and immigration status, and residential location (estimates not shown). Note: Sicker adults include respondents in fair or poor health, or who have a disability or chronic disease. Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ P < .05.
The race/ethnicity and immigration effects, however, are being picked up by income and insurance variables in the multivariate models. Regression analyses indicate that having supplemental insurance protected adults from cost-related access problems in Australia, Canada, and New Zealand (Table 5). Australians and New Zealanders with supplemental coverage also encountered less diffi-
culty seeing a specialist when needed and waited shorter times for elective surgery (Table 4). Only in Canada were individuals with supplemental coverage more satisfied than those with only public coverage with their interactions with their doctors (Table 6). On several dimensions of physician–patient interactions—being treated with dignity and respect, feeling listened to, receiving all the necessary
318
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
Table 5 The effects of income, insurance, and health status on cost-related access problems by country, based on logit regressions Adjusted odds ratio Australia
Canada
New Zealand
United Kingdom
United States
1.9∗ 3.2∗ 0.6∗ – – 2.2∗
1.5 2.2∗ 0.6∗ – – 2.0∗
1.1 0.9 1.0 – – 2.2∗
1.4 2.2∗ – 2.6∗ 1.1 3.1∗
Had a medical problem but did not see doctor due to cost Average income 1.1 1.0 Below-average income 1.8∗ 3.0∗ Additional private insurance 0.5∗ 0.5∗ US no insurance – – US public insurance – – Sicker 1.8∗ 2.8∗
1.1 1.6∗ 0.5∗ – – 2.4∗
1.2 1.3 0.9 – – 1.5
1.5 2.0∗ – 5.2∗ 1.0 3.3∗
Did not get recommended medical test or treatment due to cost Average income 1.0 0.9 Below-average income 1.3 1.6 Additional private insurance 0.5∗ 0.6∗ US no insurance – – US public insurance – – Sicker 2.1∗ 3.3∗
1.3 1.8∗ 0.6∗ – – 2.3∗
1.1 2.5 0.4 – – 2.3∗
1.3 2.6∗ – 3.5∗ 0.7 3.0∗
Had dental problem but did not see dentist due to cost Average income 0.9 Below-average income 1.4 Additional private insurance 0.5∗ US no insurance – US public insurance – Sicker 1.9∗
1.6∗ 4.1∗ 0.4∗ – – 1.3
1.2 1.6∗ 0.6∗ – – 1.3∗
0.9 1.2 0.9 – – 1.3
1.5∗ 2.3∗ – 2.3∗ 0.9 1.9∗
Problem paying medical bills Average income Below-average income Additional private insurance US no insurance US public insurance Sicker
1.9 4.4∗ 0.4∗ – – 3.2∗
2.0∗ 3.9∗ 0.8 – – 3.5∗
3.3∗ 2.4 1.6 – – 3.7 ∗
1.6 3.0∗ – 2.3∗ 0.9 3.3∗
Did not fill a prescription due to cost Average income 1.2 Below-average income 1.2 Additional private insurance 0.6∗ US no insurance – US public insurance – Sicker 1.9∗
1.1 2.5∗ 0.7 – – 3.2∗
Logistic regressions also control for age, education, minority and immigration status, and residential location (not shown). Note: Sicker adults include respondents in fair or poor health, or who have a disability or chronic disease. Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ P < 0.05.
information, and spending sufficient amount of time, Canadians with supplemental coverage reported more positive ratings than those with only public coverage (Table 6). In contrast, in UK where the rates of supplemental coverage are much lower across income groups than the other three countries with universal coverage, hav-
ing supplemental private coverage made little difference to general access, experiencing cost-related barriers, or quality of care ratings (Tables 4–6). In fact, having supplemental coverage made a difference on only one measure—shorter wait times for elective surgery. In US, having no insurance significantly increased the risk of going without needed care, even after
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
319
Table 6 The effects of income, insurance, and health status on overall quality and physician ratings by country, based on logit regressions Adjusted odds ratio Australia
Canada
New Zealand
United Kingdom
United States
Excellent or very good rating on overall quality of medical care received past 12 months Average income 1.3 0.8 0.8 Below-average income 1.2 0.8 0.8 Additional private insurance 1.2 1.2 1.4∗ US no insurance – – – US public insurance – – – Sicker 0.6∗ 0.7∗ 0.7∗
1.5∗ 1.3 1.3 – – 0.6∗
0.9 0.6∗ – 0.5∗ 0.8 0.5∗
Excellent or very good rating treating with dignity and respect Average income 0.9 0.9 Below-average income 1.0 1.4 Additional private insurance 1.2 1.6∗ US no insurance – – US public insurance – – Sicker 1.0 0.9
0.9 0.9 1.3 – – 0.9
1.5∗ 1.0 1.1 – – 0.8
0.6∗ 0.5∗ – 0.5∗ 1.0 0.7∗
Excellent or very good rating listening carefully health concerns Average income 0.8 0.8 Below-average income 1.0 1.3 Additional private insurance 1.1 1.5∗ US public insurance – – US no insurance – – Sicker 0.9 0.9
0.9 0.8 1.1 – – 0.8
1.4 1.0 1.1 – – 0.8
0.7∗ 0.7∗ – 0.9 0.5∗ 0.7∗
Excellent or very good rating spending enough time with you Average income 1.1 0.9 Below-average income 1.1 1.1 Additional private insurance 1.4∗ 1.5∗ US no insurance – – US public insurance – – Sicker 0.8 0.8
1.0 0.8 1.1 – – 0.8
1.4∗ 1.0 1.0 – – 1.0
0.6∗ 0.5∗ – 0.7 1.2 0.8
Excellent or very good rating knowing you or your family situation as it affects your health Average income 0.9 0.9 0.8 Below-average income 1.0 1.0 0.7 Additional private insurance 1.0 1.0 0.9 US no insurance – – – US public insurance – – – Sicker 0.9 0.9 0.9
1.4∗ 1.3 1.2 – – 1.1
1.0 0.6∗ – 0.6∗ 1.0 0.6∗
Excellent or very good rating providing all the information you want Average income 0.9 0.9 Below-average income 1.0 1.3 Additional private insurance 1.1 1.5∗ US no insurance – – US public insurance – – Sicker 0.7 0.8
0.9 0.8 1.2 – – 1.0
1.3 0.9 0.9 – – 0.8
0.7 0.7 – 0.5∗ 0.7 0.7∗
Excellent or very good rating being accessible by phone or in-person Average income 1.0 0.8 Below-average income 0.9 0.9 Additional private insurance 0.9 1.2 US no insurance – – US public insurance – – Sicker 0.8 0.8
1.1 0.8 1.2 – – 0.8
1.4∗ 1.2 1.0 – – 0.9
0.9 0.7∗ – 0.5∗ 0.8 0.6∗
Logistic regressions also control for age, education, minority and immigration status, and residential location (not shown). Note: Sicker adults include respondents in fair or poor health, or who have a disability or chronic disease. Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ P < 0.05.
320
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
controlling for income, race/ethnicity, immigration status, residential location, and other confounders. The uninsured were two to five times more likely than those with private insurance to have difficulties seeing a specialist, getting care on weekends or evenings, and experiencing all cost-related access problems (Tables 4 and 5). The uninsured were particularly vulnerable to forgoing care because of cost; the uninsured were five times more likely to forgo medical care and three times more likely to not get recommended tests due to costs than adults with insurance (Table 5). Respondents without insurance also reported less positive doctor–patient interactions compared to those with insurance (Table 6). On six quality and physician ratings, the uninsured were about half as likely as the insured to give their doctors high ratings. On all access measures and quality ratings US respondents with public insurance were not significantly different than those with private insurance even after controlling for income, race/ethnicity, immigration status, and other covariates. In all five countries, adults with health problems were significantly more likely to have difficulty seeing specialists when needed and, in countries with out of pocket patient costs, forgo needed care due to costs, even after adjusting for the effects of income, health insurance, and other factors (Tables 4 and 5). Indeed, respondents with health problems were two to three times more likely than those without health problems to report difficulties accessing care. Notably, however, the United States was the only country among the five nations surveyed in which sicker adults were about half as likely as adults without health problems to rate their physicians and quality of care positively. In the other four countries with respect to physician ratings, the health status variable tended to be insignificant in regression analyses (Table 6). 3.6. Health system ratings: overall and by income When asked to rate their countries’ health care systems, a minority—one of five to one of four—thought their system of care basically worked well with only minor changes needed (Table 7). The majority in each country felt that their health care system needed either major changes or needed complete rebuilding. In UK and Australia—the countries with the lowest rates of access differences by income—lower and
higher income adults shared similar views of the system (Table 7). Country averages, however, hid divided opinions by income groups in US, Canada, and New Zealand. In these countries, low income adults were significantly more likely to say that their health care system needed complete rebuilding.
4. Summary of country results and discussion Overall, with respect to providing equity in access to care and health care experiences across income classes UK emerges as the most equitable and US as the least equitable. Among the five countries, UK generally stands out for the absence of income-related disparities in access and quality care. In UK, the gap between low and high income adults is typically small or non-existent. On only 2 out of 17 measures are responses by UK lower income adults more negative than those with above-average incomes, after controlling for health, insurance, minority status and other demographic factors (Table 8). Although in UK overall patient quality and physician ratings are among the lowest among the five nations (with US often sharing bottom rankings), notably low income UK adults are more likely to rate their care more positively than do adults with above-average incomes. In fact, in UK there are no instances where quality ratings for low income adults are more negative than those given by adults with higher incomes. UK is also distinct in the minimal effect of having supplemental coverage to reduce access barriers or improve quality of care. Findings suggest that the National Health Service’s emphasis on removal of financial barriers to care, and the relatively restricted role for private insurance and markets has resulted in more equitable access across income classes. At the other extreme lies US, where disparities by income are much more pronounced than in any other country. In 12 out of 17 measures, US lower income adults are significantly more likely than those with higher incomes to have access problems and give lower physician ratings (Table 7). This income effect persists after taking into account insurance, race/ethnicity, immigration status, health, education, residential location, and age. Furthermore, lack of insurance in US undermines access and quality of care on almost all measures included in the study. These
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
321
Table 7 Overall view of the health care system Percent responding
Australia
Canada
New Zealand
United Kingdom
United States
System works well—only minor changes needed (%) Some good things but fundamental change is needed (%) System has so much wrong needs complete rebuilding (%)
25 53 19
21 59 18
18 60 20
21 60 18
18 51 28
By income (%) System works well—only minor changes needed Below-average income Above-average income
21 24
19 24
18 18
23 18
15 20
System has so much wrong needs complete rebuilding (%) Below-average income Above-average income
22 18
23∗ 13
25∗ 18
19 17
35∗ 22
Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ Indicates significant difference from above-average income at P < 0.05.
findings suggest that the absence of a national insurance system that provides basic coverage to all and a reliance on the voluntary purchase of private insurance with marked variations in the scope of benefits results in widespread access inequities by income. It is well known and documented that the uninsured in US face access barriers. Studies indicate they also receive lower quality of care [5]. The fact that low in-
come adults continue to report access difficulties and negative care experiences even when controlling for insurance suggests that their insurance is less comprehensive, less likely to provide affordable access to care, and may indicate a more limited network of providers. Australia, Canada, and New Zealand are in the middle of the two extremes in terms of income-related
Table 8 Indicators of inequity: summary measures of general and cost-related access problems, and overall quality and physician ratings, by country, based on logit regressions Australia Lower income individuals have more negative Five general access measures Five access related to cost measures Seven quality and physician rating measures Total (17)
Canada
New Zealand
United Kingdom
0 2 0
3 4 0
1 5 0
1 1 0
1 5 6
2
7
6
2
12
1 0 0
2 5 6
Adults without private health insurance, and the uninsured in US, have more negative Five general access measures 2 0 2 Five access related to cost measures 4 5 4 Seven quality and physician rating measures 1 4 1 Total (17)
7
Adults in poorer health have more negative experiences Five general access measures 3 Five access related to cost measures 5 Seven quality and physician rating measures 1 Total (17)
United States
experiencesa
9
experiencesb
9
7
1
13
1 4 1
3 5 1
3 3 1
3 5 6
6
9
7
14
Based on Tables 4–6 (see tables for details). Source: The Commonwealth Fund 2001 International Health Policy Survey. a Adjusted for insurance, health status, age, education, minority and immigration status, and residential location. b Adjusted for income, health status, age, education, minority and immigration status, and residential location.
322
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
access differences and the extent to which having private supplemental insurance is associated with more positive access experiences. Access results for New Zealand, and in some instances Australia, indicate that even where services are covered, fees that would be modest by US standards can erect financial barriers to care [2]. Unlike the United States, however, access differences by income in these countries do not extend to basic care ratings: lower and higher income adults are about equally likely to rate their physicians highly. Looking across all five countries, more equal experiences within health care systems are associated with more united views of country health care systems overall in UK and Australia. In contrast, in US, Canada, and New Zealand more widespread disparities in access experiences are linked to more divided public opinion. The findings for Canada are particularly notable for the shift since an earlier 1998 survey that found few access differences by income in Canada except for services omitted from Canada’s Medicare program and no significant differences in system views by income [2]. Comparing the 1998 and 2001 Canadian findings suggest, however, that emerging differences in access experiences may translate into divisions in public opinions over a relatively short period of time. In summary, the 2001 survey findings indicate that low income residents are particularly vulnerable when a system is under stress. For physician care and basic services, non-financial barriers to care as well as costs can contribute to worse access experiences for lower income adults. As policies change, surveys offer a way to track the effect of these changes. For policy leaders, the five-nation survey indicates that experiences of low income adults are likely to be particularly sensitive to even modest cost-sharing levels. The survey also finds that reliance on private coverage to supplement public coverage can result in
access inequities. To the extent differential access by income exists, these policies can, in turn, result in more divided opinions about the need or direction of health care reform. In sum, more unequal access care experiences by income can undermine social solidarity around a national system.
Acknowledgements The Commonwealth Fund funded the survey and analysis. The views are those of the authors and not necessarily those of the directors or officers of the Fund. This paper builds on an earlier article published in the US journal Health Affairs in May/June 2002. Authors of this paper were Robert Blendon, Cathy Schoen, Catherine DesRoches, Robin Osborn, Kimberly Scoles and Kinga Zapert. The authors thank Deirdre Downey at the Commonwealth Fund for preparation of tables, charts, and the manuscript. We also thank our anonymous reviewers for their helpful comments.
References [1] Van Doorslaer E, et al. Equity in the delivery of health care in Europe and US. Journal of Health Economics 2000;19(5):553– 83. [2] Schoen C, et al. Health insurance markets and income inequality: findings from an international health policy survey. Health Policy 2000;51:67–85. [3] Blendon RJ, et al. Inequities in health care: a five country survey. Health Affairs May/June 2002;21(3):182–91 [The full text of the article is available at http://www.healthaffairs.org]. [4] Anderson GF, Petrosyan V, Hussey PS. Multinational comparisons of health systems data. The Commonwealth Fund Chartbook, October 2002. [5] Institute of Medicine. Unequal treatment: confronting racial and ethnic disparities in health care. March 2002.
Inequities in access to medical care in five countries: findings from the 2001 Commonwealth Fund International Health Policy Survey Cathy Schoen∗ , Michelle M. Doty1 Health Policy, Research and Evaluation, The Commonwealth Fund, One East 75th Street, New York, NY 10021, USA Accepted 16 September 2003
Abstract Objective: To examine across five countries inequities in access to health care and quality of care experiences associated with income, and to determine whether these inequities persist after controlling for the effect of insurance coverage, minority and immigration status, health and other important co-factors. Design: Multivariate analysis of a cross-sectional 2001 random survey of 1400 adults in five countries: Australia, Canada, New Zealand, United Kingdom, and United States. Main outcome measures: Access difficulties and waiting times, cost-related access problems, and ratings of physicians and quality of care. Results: The study finds wide and significant disparities in access and care experience between US adults with above and below-average incomes that persist after controlling for insurance coverage, race/ethnicity, immigration status, and other important factors. In contrast, differences in UK by income were rare. There were also few significant access differences by income in Australia; yet, compared to UK, Australians were more likely to report out of pocket costs. New Zealand and Canada results fell in the mid-range of the five nations, with income gaps most pronounced on services less well covered by national systems. In the four countries with universal coverage, adults with above-average income were more likely to have private supplemental insurance. Having private insurance in Australia, Canada, and New Zealand protects adults from cost-related access problems. In contrast, in UK having supplemental coverage makes little significant difference for access measures. Being uninsured in US has significant negative consequences for access and quality ratings. Conclusions: For policy leaders, the five-nation survey demonstrates that some health systems are better able to minimize among low income adults financial barriers to access and quality care. However, the reliance on private coverage to supplement public coverage in Australia, Canada, and New Zealand can result in access inequities even within health systems that provide basic health coverage for all. If private insurance can circumvent queues or waiting times, low income adults may also be at higher risks for non-financial barriers since they are less likely to have supplemental coverage. Furthermore, greater inequality in care experiences by income is associated with more divided public views of the need for system reform. This finding was particularly striking in Canada where an increased incidence of disparities by income in 2001 compared to a 1998 survey was associated with diverging views in 2001. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Health; Policy; International; Inequality; Access; Insurance
∗
Corresponding author. Tel.: +1-212-606-3864; fax: +1-212-606-3508. E-mail addresses: [email protected] (C. Schoen), [email protected] (M.M. Doty). 1 Tel.: +1-212-606-3860; fax: +1-212-606-3508.
0168-8510/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.healthpol.2003.09.006
310
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
1. Introduction Inequities in access to medical care by income can contribute to and exacerbate disparities in health and quality of life. Conversely, policies that seek to promote equity in access for lower income individuals offer the potential of moderating underlying health differences and, over the longer term, providing more equal opportunities for health and productivity. Countries vary widely in the extent to which public financing of health care and rules governing insurance markets seek to promote equity in access and health care experiences across income classes. Although the United States stands out among developed nations at one end of the spectrum for its lack of universal coverage, countries with universal coverage also differ in benefit design, patient-cost-sharing, and the role of private insurance. Non-financial barriers (such as waiting lists or queues, community shortages, complexity) as well as financial barriers may contribute to inequities in access. If private insurance enables more affluent families to have more ready access to care or to purchase a different standard of care, access inequities by income may emerge even where few financial barriers to health care exist. In an era of budgetary constraints and often rapid changes in public health care policies, cross-national comparisons of access and care experiences by income offer the opportunity for countries to learn from each other and to assess relative performance. Yet, such comparative studies have been relatively rare due to lack of common access measures beyond physician visits or other measures of service use, and the lack of international studies using similar data-bases and time periods [1]. To assess access experiences and variations by income, the 2001 Commonwealth Fund International Health Policy Survey interviewed adults in five countries—Australia, Canada, New Zealand, United Kingdom and United States—using an array of questions on difficulties accessing medical care, financial barriers to obtaining needed health care, physician quality ratings, and overall health care system views. Previous descriptive analysis of this survey explored the extent to which experiences varied by income, without taking into account the effects of important confounding factors such as health insurance coverage, health status, education, race/ethnicity, immigra-
tion status, and residential location [2]. Each of these confounding factors can vary across income groups within each country as well as influence access and quality of care. Thus, failure to control for these effects could potentially bias any observed income-related inequities. In this paper, we re-examine the relative importance of low income and its association with access and care experiences disparities in Australia, Canada, New Zealand, United Kingdom, and the United States controlling for the effects of health insurance, health status, education, race/ethnicity, immigration status, and residential location. In addition, we examine the role of health insurance in minimizing income-related access and quality of care disparities. The analysis focuses on four central questions: How do access, cost, and adults’ perceptions of quality vary between low and higher income adults within each country? To what extent are access inequities by income attenuated after adjusting for insurance, health, and other characteristics likely to influence care experiences? In countries with universal coverage, how do access and care experiences vary depending on whether or not adults have private insurance in addition to public coverage? Do system views vary by income, and if so, are divergent system views more likely in countries with inequities in care experiences?
2. Methods and data Data come from the 2001 Commonwealth Fund International Health Policy Survey, a five-nation survey consisting of interviews with a random sample of approximately 1400 adults age 18 and older in each of five countries: Australia (1412), Canada (1400), New Zealand (1400), United Kingdom (1400), and United States (1401). Conducted by telephone during April and May of 2001 by Harris Interactive, the survey explored problems accessing medical care, quality ratings, and adults’ views of their health care system. Details of the survey methodology are reported elsewhere [3]. The following describes key measures used in the analysis and the general analytical approach. 2.1. Relative income status Income status is the main independent variable of interest. To gauge relative income levels in each coun-
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
try, the survey quoted the national median household income in the specific country (Australia $ 38,000 Australian dollars, Canada $ 40,200 Canadian dollars; New Zealand $ 33,000 NZ dollars, UK £ 23,000, and US$ 37,000) and asked respondents whether their own income was much or somewhat below this amount, about average, or much or somewhat above. In general, this technique for collecting income information resulted in few missing observations; only between 3 and 7% of the sample in each country did not respond to the income question. Respondents who did not provide responses were dropped from the analysis.2 The analysis groups adults into three income categories: “below-average”, “average”, or “above-average” median household income (Table 2). 2.2. Health status and insurance The survey asked two questions about health status, self-reported health status and the presence of a chronic disease, injury, or disability. In the analysis, we used both questions to classify a respondent as “sicker” if he or she rated their health as fair or poor health status or said yes to a serious or chronic illness, injury, or disability that required a lot of medical care. In each of the four countries with universal coverage, the survey asked whether respondents had private health insurance in addition to public or national health coverage. In US, respondents were asked whether they had private or public coverage, or were uninsured. 2.3. Race/ethnicity and foreign-born In US, sample sizes were large enough to categorize respondents by race/ethnicity according to self-reported group membership—Latino/Hispanic, black, white, Asian and other. In the remaining four countries, due to small sample sizes, a dichotomous minority status variable was created using self-reported group membership and an additional question, “do you consider yourself to be a member of minority group?” For each country, respondents 2 This translates into 92 dropped cases in Australia, 66 in Canada, 47 in New Zealand, 57 in UK, and 48 in US. Excluded cases were not significantly different from those reporting their relative income with respect to education, minority and immigration status.
311
who reported that they were a member of a minority group or who indicated they were of Aboriginal, Asian, Central/South American, or of Middle Eastern descent in Australia; Aboriginal, black, Asian, South/Central American or Middle Eastern descent in Canada; Maori or Asian/Pacific Islander in New Zealand; and black, Asian/Pacific Islander in UK were considered minorities. In the absence of having a variable which measures language proficiency, we use foreign-born as a proxy measure. In all countries, respondents who indicated that they were born in another country were categorized as foreign-born. 2.4. Education A four category educational variable was created for each country. In regressions, the equivalent of high school graduate in US was used as the referent category in all country models. 2.5. Access to health care The survey included five measures of access not explicitly related to cost. Three items asked about difficulties getting care when needed: seeing a specialist, getting care in the evenings or weekends, and getting care where they live. Two items asked about waiting times to get care, including days waited to see a doctor when sick and waiting times for elective surgery among respondents with surgery in the past 2 years. 2.6. Cost-related access problems and financial burdens The survey includes four questions about going without needed care within the past 12 months because of costs: (1) not filling a prescription; (2) having a medical problem but not seeking doctor or medical care; (3) having a dental problem but not seeing a dentist; and (4) skipping recommended medical tests, treatment, or follow-up. The survey also asked respondents to estimate actual out of pocket costs during the year. In the analysis, all currency values were converted to US dollars using currency exchange rates current at the time of the survey. We use these out of pocket cost estimates as background information on the extent to which patients are paying directly for care. Respondents also indicated whether
312
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
or not in the past 12 months they had any problems paying their medical bills. 2.7. Quality of care and physician ratings Seven ratings of quality of care (rated as excellent or very good) are included—overall quality of care received in the past 12 months and six aspects of physician care: treating you with dignity and respect; listening to your questions and health concerns; providing you with the information you want; spending enough time with you; being accessible by phone or in person, and being aware of relevant family history as it pertained to your health. 2.8. Health care system views The survey also gauged adults’ views of their health care system by asking about the perceived need for reform: whether they believed their country’s health care system worked well with only minor changes, needed major changes, or required complete rebuilding. 2.9. Analytic approach To provide a country context, we describe and compare for each country insurance coverage and insurance policies, including benefits included in national public systems. Next we present survey findings overall and by income without adjusting for confounders. Then, we present a series of within country logistic regressions which assess the independent and main effects of income, insurance coverage, and health status on access and care experiences, controlling for
race/ethnicity or minority status, immigration status, education, age and residential location (rural, suburban or urban). In the analysis, data are weighted to reflect the demographic composition of each country and are analyzed using STATA statistical software. Tables report Chi-square tests of significance and log-likelihood at P < 0.05.
3. Results 3.1. Country context: benefit coverage and role of private insurance The health insurance systems in the five countries included in the study differ substantially with respect to the role of private insurance, the extent of patient cost-sharing, exposure to out of pocket costs for medical bills, and the range of benefits covered by public insurance systems. US stands out as the one country that does not provide universal coverage for at least a core set of benefits and for its reliance on a mixed system of voluntary private insurance and public coverage for the elderly and some of the poor. Based on national census data, as of 2001 nearly one out of five US adults (18%) under age 65 were uninsured. The four other nations in the survey provide universal public coverage yet differ in the scope of benefits, cost-sharing, and prevalence and the role played by supplemental private insurance (Table 1). Australia and New Zealand both include patient cost-sharing within their basic public systems, including fees for
Table 1 Insurance and cost-sharing policies in four countries with universal coverage Insurance and cost-sharing policies
Australia
Canada
New Zealand
United Kingdom
Private insurance for services covered by public Population with private coverage (%)
Permitted
Prohibited
Permitted
Permitted
33
Majority excluded benefits only None for basic
33
12
Co-payments for many services Covered School children
None for basic
Public plan patient cost-sharing Prescription drugs Dental
On most services balance billing allowed by doctors Covered Covered
Not covered Not covered
Covered Covered
Source: G. Anderson, Multinational Comparisons of Health Care: Expenditures, Coverage, and Outcomes, The Commonwealth Fund (New York, NY): Oct. 1998.
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
physician care and other services. Although Canadian health care benefits vary somewhat by Province, Canadian benefits for physicians and hospitals are the same across provinces with no patient fees at the point of services for doctors or hospital costs. Similarly, UK has few or no patient fees at the point of services for most benefits included in public coverage. However, Canada’s core benefits exclude prescription and dental benefits while UK includes these benefits. Canadian insurance rules restrict use of private insurance to benefits not otherwise covered by the public program or for out of country care [4]. The other three countries with universal coverage allow use of private insurance for access to private physicians or hospitals, more ready access to care, and in the case of Australia and New Zealand, to cover patient cost-sharing obligations. The survey finds that in the countries with universal coverage adults with higher incomes are much more likely to report private supplemental coverage than are adults with below-average income (Table 2). Overall, adults in UK were the least likely to have additional private health insurance—rates were much lower across income groups than the other three countries with universal coverage. Canada, on the other hand, had the highest rate (61%) of private coverage. In Canada, although not universal across Provinces, supplemental coverage generally includes benefits such as dental and prescription drug benefits that are not covered through province public health insurance. In US, low income adults were much more likely to be uninsured and less likely to have private coverage (rely on public only) than were adults with above-average incomes. Due to variations in public insurance policies, including the scope of benefits covered by core public programs, the five countries differ substantially in the extent to which residents are exposed to out of pocket costs for medical bills. US stands out for high out of pocket medical expenses—42% of US respondents estimated they paid US$ 500 or more out of pocket in the past year compared to only 5% of adults in UK (Table 2). UK stands out for extensive protections against out of pocket costs—43% of adults reported no out of pocket expenses during the year. Canadians were also relatively well protected with one-third of adults reporting no out of pocket costs in the past year. Reflecting national coverage and benefit poli-
313
cies, adults in Australia and New Zealand were more likely than those in Canada or UK to encounter costs when seeking care. Yet the level of medical bill exposure was well below that experienced in the United States. 3.2. Access problems and waiting times to get care when needed Despite having quite different health care systems, the survey found that adults in all five countries were similarly likely to report that they found it extremely or very difficult to see a specialist when needed, ranging from a low of 11% in New Zealand to a high of 17% in US (Table 3). In Australia, Canada, and UK, adults were about as likely (approximately 12%) to say it was difficult to get care at night or on weekends other than an emergency room. In contrast, only 6% of adults in New Zealand had difficulty getting care on nights or weekends. Less than 10% of adults in all five countries reported they were often unable to get care where they live. Income disparities emerged on difficulty seeing a specialist when needed (in all countries except Australia), getting care at night or on weekends (in Australia, Canada, and US), and the ability to get care where one lives (in New Zealand and US) (Table 3). The income gap was consistently widest in US—lower income adults were two to three times more likely than higher income adults to report access problems. The amount of time adults spent waiting to see a physician when sick, or scheduling necessary elective surgery among respondents reporting surgery in the past 2 years varied widely between countries (Table 3). Adults in Australia and New Zealand cited more rapid access to their physicians with two-thirds saying they were able to see their doctor the same day and few reporting waits of 5 days or more. In contrast, only one-third of US and Canadian adults and two of five UK adults said they could get in the same day; waiting times of 5 days or more were more frequent for adults in these countries. Among those needing elective surgery in the past 2 years, approximately 20% of adults in all countries (except in US) experienced long wait times of 6 months or more. Among the five nations, UK adults were the most likely to report long waits for elective surgery—27% of adults in UK
314
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
Table 2 Select demographic characteristics by country Demographic characteristics
Australia
Canada
New Zealand
United Kingdom
United States
36 19 39 7
34 22 39 5
27 21 48 4
40 23 33 4
32 21 43 3
20 26
16 22
12 24
20 20
22 27
35
30
28
30
36
47 25
37 24
40 21
41 20
49 27
– 48 52
– 39 61
– 55 45
– 77 23
16 8 76
33 68
36 79
24 60
10 44
53 90
– –
– –
– –
– –
29 7
Race/ethnicity or minority status Member of minority or racial/ethnic group White Black Hispanic Other
15 – – – –
20 – – – –
24 – – – –
14 – – – –
– 73 12 11 5
Immigration status Foreign-born
21
24
21
11
10
Out of pocket medical costs in past year None US$ 1–499 US$ 500 or more
4 51 19
35 46 12
6 65 16
43 46 5
7 39 42
(%)a
Income levels Below-average Average Above-average Missing
Health status Self-reported fair or poor health Illness/disability requiring serious medical care in past 2 years Sicker adults (either fair/poor health or disabled) Percent “sicker,” by income Below-average income Above-average income Health insurance coverage Uninsured Public coverage only Private coverageb Percent with private coverage, by income Below-average income Above-average income Percent uninsured in US, by income Below-average income Above-average income
Source: Commonwealth Fund 2001 International Health Policy Survey. a Survey respondents were provided the national median income and asked to categorize their own income in relation to the country’s average. b In UK, New Zealand, Australia, and Canada private coverage is in addition to national public coverage.
waited 6 months or more for elective surgery. In contrast, long waiting times in US were rare—only 2% of adults who needed elective surgery had to wait 6 or more months. Waiting times to see a physician when sick varied by income in US for both short and long waits. In Canada, low income adults were also more likely than
higher income adults to report waiting 5 days or more. Waiting times for elective surgery did not vary significantly by income. Among adults reporting surgery within the past 2 years, lower income adults in all five countries were no more likely than those with higher incomes to experience long waiting times for elective surgery.
Table 3 General and cost-related access and medical bill problems, and overall quality and physician ratings, by country and income levels, 2001 unadjusted percentages Australia
General access problem measures Extremely or very difficult to see a specialist when needed Very difficult to get care on weekends or evenings Often not able to get appropriate care where you live
Canada
New Zealand
Total
Belowaverage
Aboveaverage
Total
Belowaverage
Aboveaverage
Total
Belowaverage
12
14
11
16
20∗
14
11
21∗
12
15∗
10
15
19∗
13
5
3 ∗
United Kingdom Aboveaverage
Total
Belowaverage
6
13
16∗
5
12
11
United States Aboveaverage
Total
Belowaverage
Aboveaverage
9
17
30∗
8
13
18
26∗
16
∗
4
7
5
8
9
6
6
8
5
4
4
4
7
12
Time waited for doctor’s appointment when sick Same day 1–2 days 3–4 days Five days or more
62 27 6 5
58∗ 28 7 6∗
65 27 4 3
35 28 11 23
35 22 11 27∗
36 31 11 20
69 25 3 1
70 24 2 2
68 25 4 1
42 28 12 16
41 29 12 15
39 27 13 17
36 33 10 18
30∗ 33 10 21∗
40 33 9 16
Of those needing elective surgery in the past 2 years, time waited for it Less than 2 weeks 2 weeks, less than 1 month 1 month, less than 6 months 6 months or more
34 17 31 19
28 15 34 23
39 13 32 16
20 17 44 18
27 14 37 19
19 18 44 18
27 16 37 17
22 11 43 24
30 18 36 14
26 12 32 27
21 6∗ 42∗ 29
27 17 28 25
37 26 30 2
38 24 34 3
38 28 27 1
19 15
21 17
18 14
13 6
22∗ 9∗
7 4
15 14
20∗ 18∗
11 11
7 2
7 4∗
7 1
26 22
39∗ 36∗
18 14
11 33 11
14 38∗ 17∗
10 31 8
5 26 7
9∗ 42∗ 14∗
3 15 3
20 37 12
24∗ 40 20∗
18 36 7
3 19 3
4 20 4∗
2 19 2
24 35 21
36∗ 51∗ 35∗
15 24 11
63
64
59
54
51∗
60
67
66
70
53
56∗
45
57
45∗
65
80 73 69 63 72 59
82 76 71 66 75 60
80 74 68 63 72 57
79 74 62 59 67 55
80 76 61 60 69 54
81 77 65 61 69 59
84 75 71 67 73 64
84 74 70 65 72 63
85 77 72 68 74 64
73 67 54 51 58 48
73 67 55∗ 55∗ 58 51∗
70 63 49 43 55 41
72 65 58 57 63 52
61∗ 56∗ 46∗ 46∗ 53∗ 43∗
81 73 67 63 71 58
Cost-related access problems Did not do the following due to cost in the past year Did not fill a prescription Did not get recommended test, treatment, or follow-up Had medical problem but did not see doctor Needed dental care but did not see dentist Had problem paying medical bills Overall quality and physician ratings Excellent/very good rating on overall medical care received past year Physician ratings (excellent or very good on) Treating with dignity and respect Listening carefully health concerns Spending enough time with you Knowing you or your family situation Providing you with all information you want Being accessible by phone or in person
315
Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ Indicates significant difference from above-average income at P < 0.05 or better.
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
6
316
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
3.3. Access problems due to cost and medical bill burdens Overall, UK adults were the least likely and US adults were the most likely to report access problems due to cost or medical bill problems (Table 3). In Canada (5%) and UK (3%), few adults reported cost barriers to physician care. Similarly small proportions of adults in Canada (7%) and UK (3%) faced problems paying their medical bills. Dental care emerged as the only measure for which delays or foregone care due to costs in US rivaled levels reported in some other countries (e.g. Australia and New Zealand). In UK, gaps between income groups were generally small, and few adults—either high or low income—reported access problems due to cost. In contrast, the gap by income was significant and particularly wide in US on all five cost-related access problems and financial burdens measures, in New Zealand on four of five measures, and in Australia on two of five measures. In Canada, access inequities by income were widest on not filling prescriptions and forgoing needed dental care—the two benefits not included in the basic public program. Income gaps also existed in Canada for physician care and follow-up tests but, in general, few Canadians reported cost-related problems for these services. 3.4. Quality of care and physician ratings Ratings of overall quality of care and physician– patient dimensions of care were generally highest in Australia and New Zealand and lowest in UK and US; Canadian ratings were in between (Table 3). Although physician ratings for the total population were among the lowest in UK, in comparisons of quality ratings by income, UK emerged as the one country in which ratings of care were, if anything, more positive among lower income adults than among adults with above-average incomes. On four out of seven quality measures (overall quality of care, time spent with the doctor, having access to doctor by phone or in person, and knowing the patient well), a significantly larger share of UK adults with lower incomes rated their care as excellent or very good than those with higher incomes. There were no instances where quality ratings for low income UK adults were more negative.
In contrast, in the US, significant income disparities emerged on all quality and physician ratings measures with lower income adults significantly less likely to give positive ratings (Table 3). The United States was unique on these ratings disparities. In Australia, Canada, and New Zealand ratings did not vary significantly by income—lower income adults were generally as likely to rate care positively as were higher income adults. 3.5. Adjusted estimates of income and insurance-related disparities: multivariate results The country pattern of disparities by income generally persisted after controlling for the independent effects of insurance, health status, race/ethnicity or minority status, immigration status, education, residential location, and age (Tables 4–6). In logistic regressions, income gaps tended to narrow compared to unadjusted odds ratios.3 In US, regardless of insurance status, race/ethnicity, and immigration status income-related disparities remained wide and significant suggesting that income plays a dominating role in either facilitating or hampering access to the US health care system. Notably unique to US, there was an income effect for quality of care experiences, even after controlling for insurance, race/ethnicity, immigration status, and other demographic factors. Other studies in US have documented that minorities, immigrants, and those with lower education often report more negative health care experiences [5]. With respect to physician or quality ratings, however, there were no statistically significant differences between whites, Hispanics or blacks after controlling for income, insurance, and other demographic factors. In fact, in the multivariate analysis race/ethnicity was significant on only 3 out of 17 measures relating to access (seeing a specialist, care on weekends or nights, and dental access). Similarly, education was not statistically significant in any multivariate model once income, race/ethnicity, immigration status and other demographic factors were taken into account.4 In US, minorities are likely to be low income, have lower education attainment, and lack insurance which contribute to inequities in US. 3
Full results are available from authors. Full results for all parameter estimates are available from authors. 4
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
317
Table 4 The effects of income, insurance, and health status on general access problems by country, based on logit regressions Adjusted odds ratio Australia
New Zealand
United Kingdom
United States
Extremely or very difficult to see a specialist when needed Average income 1.0 1.1 Below-average income 1.2 1.5∗ Additional private insurance 0.5∗ 1.1 US no insurance – – US public insurance – – Sicker 1.4∗ 1.1
1.5 3.8∗ 0.5∗ – – 1.9∗
1.5 1.7∗ 0.6 – – 2.2∗
1.6 3.0∗ – 3.9∗ 1.0 1.8∗
Very difficult to get care on weekends Average income Below-average income Additional private insurance US no insurance US public insurance Sicker
1.1 1.5∗ 1.0 – – 1.1
0.7 0.6 1.0 – – 2.1∗
0.8 0.9 1.1 – – 1.2
0.6∗ 1.3 – 2.5∗ 0.7 2.4∗
Often not able to get appropriate care where you live Average income 0.9 Below-average income 1.5 Additional private Insurance 0.8 US no insurance – US public insurance – Sicker 1.1
1.2 1.3 0.8 – – 1.0
0.8 1.7 1.0 – – 1.9∗
1.0 0.7 1.1 – – 2.5∗
1.7 2.0 – 1.5 1.7 1.7
Waited 5 days or more for a doctor’s appointment Average income 0.7 Below-average income 1.9 Additional private insurance 1.5 US no insurance – US public insurance – Sicker 2.4∗
1.0 1.4∗ 1.0 – – 0.9
1.7 1.7 0.7 – – 1.2
0.8 0.9 1.0 – – 1.5∗
0.7 1.1 – 1.3
Waited 6 months or more for elective Average income Below-average income Additional private insurance US no insurance US public insurance Sicker
0.8 1.0 1.2 – – 1.8∗
1.6 1.1 0.2∗ – – 1.3
1.1 1.0 0.5∗ – – 1.3
1.0 1.6 – 4.0
or evenings 0.8 1.2 0.7 – – 1.9∗
surgery 0.5 1.0 0.2∗ – – 1.0
Canada
1.7∗
1.4
Logistic regressions also control for age, education, minority and immigration status, and residential location (estimates not shown). Note: Sicker adults include respondents in fair or poor health, or who have a disability or chronic disease. Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ P < .05.
The race/ethnicity and immigration effects, however, are being picked up by income and insurance variables in the multivariate models. Regression analyses indicate that having supplemental insurance protected adults from cost-related access problems in Australia, Canada, and New Zealand (Table 5). Australians and New Zealanders with supplemental coverage also encountered less diffi-
culty seeing a specialist when needed and waited shorter times for elective surgery (Table 4). Only in Canada were individuals with supplemental coverage more satisfied than those with only public coverage with their interactions with their doctors (Table 6). On several dimensions of physician–patient interactions—being treated with dignity and respect, feeling listened to, receiving all the necessary
318
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
Table 5 The effects of income, insurance, and health status on cost-related access problems by country, based on logit regressions Adjusted odds ratio Australia
Canada
New Zealand
United Kingdom
United States
1.9∗ 3.2∗ 0.6∗ – – 2.2∗
1.5 2.2∗ 0.6∗ – – 2.0∗
1.1 0.9 1.0 – – 2.2∗
1.4 2.2∗ – 2.6∗ 1.1 3.1∗
Had a medical problem but did not see doctor due to cost Average income 1.1 1.0 Below-average income 1.8∗ 3.0∗ Additional private insurance 0.5∗ 0.5∗ US no insurance – – US public insurance – – Sicker 1.8∗ 2.8∗
1.1 1.6∗ 0.5∗ – – 2.4∗
1.2 1.3 0.9 – – 1.5
1.5 2.0∗ – 5.2∗ 1.0 3.3∗
Did not get recommended medical test or treatment due to cost Average income 1.0 0.9 Below-average income 1.3 1.6 Additional private insurance 0.5∗ 0.6∗ US no insurance – – US public insurance – – Sicker 2.1∗ 3.3∗
1.3 1.8∗ 0.6∗ – – 2.3∗
1.1 2.5 0.4 – – 2.3∗
1.3 2.6∗ – 3.5∗ 0.7 3.0∗
Had dental problem but did not see dentist due to cost Average income 0.9 Below-average income 1.4 Additional private insurance 0.5∗ US no insurance – US public insurance – Sicker 1.9∗
1.6∗ 4.1∗ 0.4∗ – – 1.3
1.2 1.6∗ 0.6∗ – – 1.3∗
0.9 1.2 0.9 – – 1.3
1.5∗ 2.3∗ – 2.3∗ 0.9 1.9∗
Problem paying medical bills Average income Below-average income Additional private insurance US no insurance US public insurance Sicker
1.9 4.4∗ 0.4∗ – – 3.2∗
2.0∗ 3.9∗ 0.8 – – 3.5∗
3.3∗ 2.4 1.6 – – 3.7 ∗
1.6 3.0∗ – 2.3∗ 0.9 3.3∗
Did not fill a prescription due to cost Average income 1.2 Below-average income 1.2 Additional private insurance 0.6∗ US no insurance – US public insurance – Sicker 1.9∗
1.1 2.5∗ 0.7 – – 3.2∗
Logistic regressions also control for age, education, minority and immigration status, and residential location (not shown). Note: Sicker adults include respondents in fair or poor health, or who have a disability or chronic disease. Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ P < 0.05.
information, and spending sufficient amount of time, Canadians with supplemental coverage reported more positive ratings than those with only public coverage (Table 6). In contrast, in UK where the rates of supplemental coverage are much lower across income groups than the other three countries with universal coverage, hav-
ing supplemental private coverage made little difference to general access, experiencing cost-related barriers, or quality of care ratings (Tables 4–6). In fact, having supplemental coverage made a difference on only one measure—shorter wait times for elective surgery. In US, having no insurance significantly increased the risk of going without needed care, even after
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
319
Table 6 The effects of income, insurance, and health status on overall quality and physician ratings by country, based on logit regressions Adjusted odds ratio Australia
Canada
New Zealand
United Kingdom
United States
Excellent or very good rating on overall quality of medical care received past 12 months Average income 1.3 0.8 0.8 Below-average income 1.2 0.8 0.8 Additional private insurance 1.2 1.2 1.4∗ US no insurance – – – US public insurance – – – Sicker 0.6∗ 0.7∗ 0.7∗
1.5∗ 1.3 1.3 – – 0.6∗
0.9 0.6∗ – 0.5∗ 0.8 0.5∗
Excellent or very good rating treating with dignity and respect Average income 0.9 0.9 Below-average income 1.0 1.4 Additional private insurance 1.2 1.6∗ US no insurance – – US public insurance – – Sicker 1.0 0.9
0.9 0.9 1.3 – – 0.9
1.5∗ 1.0 1.1 – – 0.8
0.6∗ 0.5∗ – 0.5∗ 1.0 0.7∗
Excellent or very good rating listening carefully health concerns Average income 0.8 0.8 Below-average income 1.0 1.3 Additional private insurance 1.1 1.5∗ US public insurance – – US no insurance – – Sicker 0.9 0.9
0.9 0.8 1.1 – – 0.8
1.4 1.0 1.1 – – 0.8
0.7∗ 0.7∗ – 0.9 0.5∗ 0.7∗
Excellent or very good rating spending enough time with you Average income 1.1 0.9 Below-average income 1.1 1.1 Additional private insurance 1.4∗ 1.5∗ US no insurance – – US public insurance – – Sicker 0.8 0.8
1.0 0.8 1.1 – – 0.8
1.4∗ 1.0 1.0 – – 1.0
0.6∗ 0.5∗ – 0.7 1.2 0.8
Excellent or very good rating knowing you or your family situation as it affects your health Average income 0.9 0.9 0.8 Below-average income 1.0 1.0 0.7 Additional private insurance 1.0 1.0 0.9 US no insurance – – – US public insurance – – – Sicker 0.9 0.9 0.9
1.4∗ 1.3 1.2 – – 1.1
1.0 0.6∗ – 0.6∗ 1.0 0.6∗
Excellent or very good rating providing all the information you want Average income 0.9 0.9 Below-average income 1.0 1.3 Additional private insurance 1.1 1.5∗ US no insurance – – US public insurance – – Sicker 0.7 0.8
0.9 0.8 1.2 – – 1.0
1.3 0.9 0.9 – – 0.8
0.7 0.7 – 0.5∗ 0.7 0.7∗
Excellent or very good rating being accessible by phone or in-person Average income 1.0 0.8 Below-average income 0.9 0.9 Additional private insurance 0.9 1.2 US no insurance – – US public insurance – – Sicker 0.8 0.8
1.1 0.8 1.2 – – 0.8
1.4∗ 1.2 1.0 – – 0.9
0.9 0.7∗ – 0.5∗ 0.8 0.6∗
Logistic regressions also control for age, education, minority and immigration status, and residential location (not shown). Note: Sicker adults include respondents in fair or poor health, or who have a disability or chronic disease. Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ P < 0.05.
320
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
controlling for income, race/ethnicity, immigration status, residential location, and other confounders. The uninsured were two to five times more likely than those with private insurance to have difficulties seeing a specialist, getting care on weekends or evenings, and experiencing all cost-related access problems (Tables 4 and 5). The uninsured were particularly vulnerable to forgoing care because of cost; the uninsured were five times more likely to forgo medical care and three times more likely to not get recommended tests due to costs than adults with insurance (Table 5). Respondents without insurance also reported less positive doctor–patient interactions compared to those with insurance (Table 6). On six quality and physician ratings, the uninsured were about half as likely as the insured to give their doctors high ratings. On all access measures and quality ratings US respondents with public insurance were not significantly different than those with private insurance even after controlling for income, race/ethnicity, immigration status, and other covariates. In all five countries, adults with health problems were significantly more likely to have difficulty seeing specialists when needed and, in countries with out of pocket patient costs, forgo needed care due to costs, even after adjusting for the effects of income, health insurance, and other factors (Tables 4 and 5). Indeed, respondents with health problems were two to three times more likely than those without health problems to report difficulties accessing care. Notably, however, the United States was the only country among the five nations surveyed in which sicker adults were about half as likely as adults without health problems to rate their physicians and quality of care positively. In the other four countries with respect to physician ratings, the health status variable tended to be insignificant in regression analyses (Table 6). 3.6. Health system ratings: overall and by income When asked to rate their countries’ health care systems, a minority—one of five to one of four—thought their system of care basically worked well with only minor changes needed (Table 7). The majority in each country felt that their health care system needed either major changes or needed complete rebuilding. In UK and Australia—the countries with the lowest rates of access differences by income—lower and
higher income adults shared similar views of the system (Table 7). Country averages, however, hid divided opinions by income groups in US, Canada, and New Zealand. In these countries, low income adults were significantly more likely to say that their health care system needed complete rebuilding.
4. Summary of country results and discussion Overall, with respect to providing equity in access to care and health care experiences across income classes UK emerges as the most equitable and US as the least equitable. Among the five countries, UK generally stands out for the absence of income-related disparities in access and quality care. In UK, the gap between low and high income adults is typically small or non-existent. On only 2 out of 17 measures are responses by UK lower income adults more negative than those with above-average incomes, after controlling for health, insurance, minority status and other demographic factors (Table 8). Although in UK overall patient quality and physician ratings are among the lowest among the five nations (with US often sharing bottom rankings), notably low income UK adults are more likely to rate their care more positively than do adults with above-average incomes. In fact, in UK there are no instances where quality ratings for low income adults are more negative than those given by adults with higher incomes. UK is also distinct in the minimal effect of having supplemental coverage to reduce access barriers or improve quality of care. Findings suggest that the National Health Service’s emphasis on removal of financial barriers to care, and the relatively restricted role for private insurance and markets has resulted in more equitable access across income classes. At the other extreme lies US, where disparities by income are much more pronounced than in any other country. In 12 out of 17 measures, US lower income adults are significantly more likely than those with higher incomes to have access problems and give lower physician ratings (Table 7). This income effect persists after taking into account insurance, race/ethnicity, immigration status, health, education, residential location, and age. Furthermore, lack of insurance in US undermines access and quality of care on almost all measures included in the study. These
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
321
Table 7 Overall view of the health care system Percent responding
Australia
Canada
New Zealand
United Kingdom
United States
System works well—only minor changes needed (%) Some good things but fundamental change is needed (%) System has so much wrong needs complete rebuilding (%)
25 53 19
21 59 18
18 60 20
21 60 18
18 51 28
By income (%) System works well—only minor changes needed Below-average income Above-average income
21 24
19 24
18 18
23 18
15 20
System has so much wrong needs complete rebuilding (%) Below-average income Above-average income
22 18
23∗ 13
25∗ 18
19 17
35∗ 22
Source: The Commonwealth Fund 2001 International Health Policy Survey. ∗ Indicates significant difference from above-average income at P < 0.05.
findings suggest that the absence of a national insurance system that provides basic coverage to all and a reliance on the voluntary purchase of private insurance with marked variations in the scope of benefits results in widespread access inequities by income. It is well known and documented that the uninsured in US face access barriers. Studies indicate they also receive lower quality of care [5]. The fact that low in-
come adults continue to report access difficulties and negative care experiences even when controlling for insurance suggests that their insurance is less comprehensive, less likely to provide affordable access to care, and may indicate a more limited network of providers. Australia, Canada, and New Zealand are in the middle of the two extremes in terms of income-related
Table 8 Indicators of inequity: summary measures of general and cost-related access problems, and overall quality and physician ratings, by country, based on logit regressions Australia Lower income individuals have more negative Five general access measures Five access related to cost measures Seven quality and physician rating measures Total (17)
Canada
New Zealand
United Kingdom
0 2 0
3 4 0
1 5 0
1 1 0
1 5 6
2
7
6
2
12
1 0 0
2 5 6
Adults without private health insurance, and the uninsured in US, have more negative Five general access measures 2 0 2 Five access related to cost measures 4 5 4 Seven quality and physician rating measures 1 4 1 Total (17)
7
Adults in poorer health have more negative experiences Five general access measures 3 Five access related to cost measures 5 Seven quality and physician rating measures 1 Total (17)
United States
experiencesa
9
experiencesb
9
7
1
13
1 4 1
3 5 1
3 3 1
3 5 6
6
9
7
14
Based on Tables 4–6 (see tables for details). Source: The Commonwealth Fund 2001 International Health Policy Survey. a Adjusted for insurance, health status, age, education, minority and immigration status, and residential location. b Adjusted for income, health status, age, education, minority and immigration status, and residential location.
322
C. Schoen, M.M. Doty / Health Policy 67 (2004) 309–322
access differences and the extent to which having private supplemental insurance is associated with more positive access experiences. Access results for New Zealand, and in some instances Australia, indicate that even where services are covered, fees that would be modest by US standards can erect financial barriers to care [2]. Unlike the United States, however, access differences by income in these countries do not extend to basic care ratings: lower and higher income adults are about equally likely to rate their physicians highly. Looking across all five countries, more equal experiences within health care systems are associated with more united views of country health care systems overall in UK and Australia. In contrast, in US, Canada, and New Zealand more widespread disparities in access experiences are linked to more divided public opinion. The findings for Canada are particularly notable for the shift since an earlier 1998 survey that found few access differences by income in Canada except for services omitted from Canada’s Medicare program and no significant differences in system views by income [2]. Comparing the 1998 and 2001 Canadian findings suggest, however, that emerging differences in access experiences may translate into divisions in public opinions over a relatively short period of time. In summary, the 2001 survey findings indicate that low income residents are particularly vulnerable when a system is under stress. For physician care and basic services, non-financial barriers to care as well as costs can contribute to worse access experiences for lower income adults. As policies change, surveys offer a way to track the effect of these changes. For policy leaders, the five-nation survey indicates that experiences of low income adults are likely to be particularly sensitive to even modest cost-sharing levels. The survey also finds that reliance on private coverage to supplement public coverage can result in
access inequities. To the extent differential access by income exists, these policies can, in turn, result in more divided opinions about the need or direction of health care reform. In sum, more unequal access care experiences by income can undermine social solidarity around a national system.
Acknowledgements The Commonwealth Fund funded the survey and analysis. The views are those of the authors and not necessarily those of the directors or officers of the Fund. This paper builds on an earlier article published in the US journal Health Affairs in May/June 2002. Authors of this paper were Robert Blendon, Cathy Schoen, Catherine DesRoches, Robin Osborn, Kimberly Scoles and Kinga Zapert. The authors thank Deirdre Downey at the Commonwealth Fund for preparation of tables, charts, and the manuscript. We also thank our anonymous reviewers for their helpful comments.
References [1] Van Doorslaer E, et al. Equity in the delivery of health care in Europe and US. Journal of Health Economics 2000;19(5):553– 83. [2] Schoen C, et al. Health insurance markets and income inequality: findings from an international health policy survey. Health Policy 2000;51:67–85. [3] Blendon RJ, et al. Inequities in health care: a five country survey. Health Affairs May/June 2002;21(3):182–91 [The full text of the article is available at http://www.healthaffairs.org]. [4] Anderson GF, Petrosyan V, Hussey PS. Multinational comparisons of health systems data. The Commonwealth Fund Chartbook, October 2002. [5] Institute of Medicine. Unequal treatment: confronting racial and ethnic disparities in health care. March 2002.