Inequality and inequity in health care use among older people in the United Kingdom. Sara Allin and Elias Mossialos

Transcription

1 Inequality and inequity in health care use among older people in the United Kingdom Sara Allin and Elias Mossialos

2 Abstract An area that remains relatively unexamined in the literature on older people is horizontal equity in health care service use. The purpose of this study is to investigate the extent of incomerelated inequity in the use of GP, inpatient, outpatient and dentist services among individuals aged 65 and older in the UK between 1997 and 2003 using data from the BHPS. The probability of GP, outpatient, dentist or inpatient service use between 1997 and 2003 was predicted using multiple random effects logistic regression panel models and income-related horizontal inequity was calculated. An inequity index that is equal to zero implies that after controlling for differences in need across income groups, individuals have equal probability of service use, regardless of income. The results indicate that individuals on lower income are significantly more likely to visit a GP, or use outpatient and inpatient care than the better-off, while the reverse is seen for dental care. After adjusting for differences in need with income, horizontal inequity is found favouring those on higher income for all service areas, particularly outpatient and dental care. 2

3 Introduction Horizontal equity, commonly defined as equal access to health care services for equal need, is one of the fundamental goals of most universally funded health care systems. While equal access presupposes that individuals are given equal opportunities to access services, for example by not charging fees, the goal of equal utilisation for equal need implies a different set of conditions. Although inequity in utilisation may not solely reflect inappropriate or unfair differentials in service use (e.g. different preferences, culture, etc.), it is the measure of equity most commonly studied to date. There has been considerable research in the area of equity in utilisation of health services in the past decade. However, this research has largely focussed on the general population, despite older people being by far the highest consumers of health services, and facing potentially greater difficulties in accessing health care services. For instance, limitations in mobility, insufficient social support, reduced access to health information sources such as the internet are likely to become more pronounced with age. These barriers are unlikely to be equally distributed across socio-economic groups, with well-educated, more financially secure older people experiencing less barriers to access than the less educated, lower income groups. These differences are likely to be reflected in differential, or inequitable patterns of service use across income groups. Studies of health care utilisation, often measured by self-reported use of services in surveys, in the general population tend to reveal similar findings. In Canada and parts of Europe, studies of physician and hospital service use have found that after controlling for health status or need, service use does not vary, or varies only slightly, across income or socio-economic groups (Keskimaki et al. 1995; Roos and Mustard 1997; Finkelstein 2001; Van Der Heyden et al. 2003). However, there appear to be differences in the probability of specialist consultation and preventive care, favouring higher income and educated groups (Roos and Mustard 1997; Whitehead 1997; Dunlop 2000; Finkelstein 2001; Van Der Heyden et al. 2003). Similarly in the UK, utilization of primary care and hospital services appears equitable or propoor, whereas preventive and specialist care tends to favour the better off. GP and inpatient service use has been shown to be pro-poor (O Donnell and Propper 1991; Nolan 1994; Propper 3

4 1998; Goddard and Smith 2001; Dixon et al. 2003), while others contend that the distribution of health care services favours the wealthy (Le Grand 1978; Le Grand 1991). Pooling data from the Health Survey for England ( ), Sutton et al. found that, consistent with previous studies, GP use is not affected by socio-economic indicators, inpatient service use favours individuals on higher income and higher education, and outpatient services favour the employed and higher educated (Sutton et al. 2002). More recently Morris et al found that in England while low income individuals are more likely to visit a GP, they have lower use of secondary care than would be predicted based on need (Morris et al. 2005). In the case of arthritis care, while the probability of receiving both NHS and private care is driven by illness severity, the higher educated are more likely to use private care, and once NHS care is taken, the higher educated individuals receive more care (Propper et al. 2005). Using 1996 data from the European Community Household Panel, van Doorslaer et al found that after controlling for differences in need across income groups, there is a small degree of income-related inequity in the probability of contacting a GP favouring the rich and more significant inequity in specialist care in the UK and most European countries (van Doorslaer et al. 2004). On the contrary, using 2001 data from the British Household Panel Survey, there was no significant pro-rich inequity found in GP, specialist or inpatient services, although dentist services were significantly inequitable favouring the higher income groups (van Doorslaer and Masseria 2004). There is some evidence revealing inequity in service use among older people in Europe. One can argue that equity is achieved in cases where need predicts health service use, alternatively inequity is revealed when socio-economic factors are significant. Need, as measured by selfreported health status, has been shown to significantly predict health care use in Canada and the UK (Roos 1989; Dening et al. 1998). However in Sweden, high socio-economic position, rather than need, predicts having at least one health care contact among those over 60, but not under (Merlo et al. 2003). In Finland, Hakkinen and Luoma report higher income is associated with physician service utilization in the general population, including adults over 64 (Hakkinen and Luoma 1995). This pro-rich inequity has also been found in Switzerland (Schellhorn et al. 2000). Likewise, higher levels of education tend to increase the likelihood of contacting one s GP or going to the hospital among older people in London, UK (Nelson et al. 2002). Also in London, older people in the top four income quintiles were more likely to consult a GP and 4

5 have an outpatient (but not inpatient) visit than those from the poorest quintile, after taking health status into account (Evandrou 2003). The purpose of the present study is to measure income-related inequity after controlling for differences in need in the use of GP, outpatient, inpatient and dental services among older people in the UK during the period 1997 to Methods This study was conducted using data from the British Household Panel Survey (BHPS). 1 The BHPS is a longitudinal cohort survey of adult members of a nationally representative sample of British households, including Scotland, Northern Ireland and Wales. The survey commenced in 1991, and the latest wave of the BHPS with available data was collected in The survey collects data from all adult members of the household. Those in the initial sample are followed until they refuse to participate, die, or are lost to follow-up. The present study included all individuals aged 65 or over in the period 1997 to Only those with complete responses were included in the analysis, therefore those with proxy respondents (due to inability to respond themselves) were excluded. The percentage of proxy respondents is around 2%. Further information on the methodology of the BHPS is available from the online documentation. 2 The demographic variables included in the analysis were age, sex, and marital status. Age is measured at the time of the interview, and is grouped into 5-year age bands: 65-69; 70-74; 75-79; 80-84, and 85+. Marital status is categorized as: married; divorced, separated or never married; and widowed. Information on whether or not the individual smokes was also included. Educational qualifications are separated into three groups: no qualifications; non-advanced qualifications (including GSE A and O levels, CSE grades 2-5); and advanced qualifications (apprenticeships, higher degree, first degree, teaching and other qualifications). Income is measured as gross household income in the last month, which is derived from disaggregated income sources including labour and non-labour income, transfer income, investment income, benefit income and pension income. Income is equivalised for household composition using the BHPS equivalisation scale (before housing costs). 5

6 Measures of need are examined separately and are approximated from several health indicators, including self-reported health status, which is the most preferred proxy for need (Newbold et al.1995; Sutton et al. 1999). Information on self-reported health status came from the following question: Please think back over the last 12 months about how your health has been. Compared to people of your own age, would you say that your health has on the whole been: excellent, good, fair, poor, or very poor? Additional health indicators include whether or not the respondent has any of the 12 listed health problems in the survey, which is used to create a variable indicating the presence of three or more health problems. Also, there is a question of whether or not health limits daily activities. The score on the General Health Questionnaire (a measure of mental health; GHQ) below or above 3 (out of 12) is also included in the analysis. Health service use is measured by the following questions (all referring to 1 September of the past year): Approximately how many times have you talked to, or visited a GP or family doctor about your own health? Have you yourself made use of hospital consultant/outpatient services? Have you been in hospital or clinic as an in-patient overnight or longer? Have you had a dental check-up? The probability of a GP, outpatient, dentist or inpatient service use between 1997 and 2003 was predicted using random effects logistic regression panel models. Most of the variables included within the models are categorical, thus bivariate or multivariate representations are created. In order to measure income-related inequality and inequity, concentration (inequality) indices were calculated according to methodology developed by van Doorslaer et al (van Doorslaer et al 2004). The inequality index would reach zero if all individuals had equal probability of seeking health care, regardless of income; the inequity index would near zero if after controlling for differences in need across income groups, individuals on different income would have equal probability of service use. In other words, the horizontal inequity index addresses the question: after controlling for differences in need (as measured by health status and demographics) across income groups, are individuals on higher income more likely to use health care services than lower income comparators? All analyses were conducted using STATA version 8. 6

7 Results The 1997 sample of the BHPS included 1855 older individuals (with complete data), which increased to 2644 individuals in This increase results both from individuals ageing thus entering the age 65+ age group, and additional individuals being included in the sample as they enter a household with an original sample member. In 1997, 82.8% of the sample visited the GP at least once, with 31.9% reporting 1-2 visits, 27.6% 3-5 visits, 11.4% 6-10 visits, and 11.9% more than 10 visits. Also in 1997, 28.5% had an outpatient visit, 39.2% visited a dentist, and 14.5% were admitted to hospital (for an average of 14.5 days). In 2003, 86% of the sample visited the GP at least once, with 30.5% reporting 1-2 visits, 27.6% 3-5 visits, 16.3% 6-10 visits, and 11.5% more than 10 visits. That same year, 43.3% of the sample had an outpatient visit, 46.2% visited a dentist, and 15.1% spent at least one night in hospital (for an average of 12.8 days). The mean age in both 1997 and 2003 was 74 years. Table 1 shows the results of the random effects logit model measuring the probability of health care service use during the period 1997 to The health indicators, in particular selfreported poor and very poor health, are the strongest predictors of health service use in all areas except dental care, where better health, not worse health, is associated with seeking dental care. Among the socio-economic factors, having private medical insurance (PMI) is significantly associated with all four health service areas, in particular with dental care where individuals with PMI are over twice as likely to have seen a dentist in the past year. Home ownership and higher educational qualifications are significantly associated with outpatient and, more strongly, dental services (individuals with advanced qualifications are more than 15 times more likely to have had a dental check-up). Regional effects are less significant. Compared to those living in London, individuals in Northern Ireland are less likely to have an outpatient visit. Those living in Scotland are more likely to have had an inpatient stay. Finally, individuals in England (not the south-east), Wales, Scotland, or Northern Ireland are less likely to have a dental check-up. One would predict based on the results from the logit models revealing significant income effects in dental, outpatient and inpatient care, that income-related inequity would appear most 7

8 strongly in dental care, and secondary care (both outpatient and inpatient services), and less so in primary (GP) care. Insert Table 1 here Table 2 displays the income-related inequality indices, revealing a pro-rich inequity in all four health service areas. Cm denotes actual/total income-related inequality - the inequality in the actual distribution of service use (unadjusted). Cn denotes the distribution of need according to income, with a negative index implying greater need concentrated among the worse off, and vice versa. HI denotes horizontal inequity - the inequality remaining in the distribution of service use after standardisation for need differences across the income distribution. Cm is positive and significant in GP, outpatient and dental care, suggesting overall inequality in probability of use favouring the higher income groups. The distribution of need, as measured by health status and age, is concentrated among lower income groups, as indicated by the negative Cn. However in the case of dental care, only age and sex was considered to be need-related (therefore standardised in the analysis) since it is generally the case that the healthy, younger individuals accessing dental care, therefore there is little difference between the unstandardised (actual; Cm) and standardised inequality (HI) in dental care. Once need differences are standardised for, the remaining income-related inequality, or horizontal inequity (HI), is positive, and significant, in all four health care service areas. Insert Table 2 here 8

9 Discussion The purpose of the present study was to investigate the degree of income-related inequality in use of health care services among older people in the UK in the period Results support the existence of significant inequality to varying degrees in GP, outpatient, inpatient and dental care. Therefore, there is evidence of violation of the principle of equal use for equal need by income GP visits Primary care appears the least inequitable of the four service areas. However, contrary to previous findings investigating income-related inequality in the general population in the UK (e.g. van Doorslaer et al. 2004; van Doorslaer and Masseria 2004; Morris et al. 2005), evidence suggests that individuals on higher income are more likely to see their GP than those on lower income. Also, despite indicators of health (or need) being the most strongly associated with GP service use, non-need factors are also significant, namely being a non-smoker, being married, and having PMI. Since take-up of PMI is associated with several other socio-economic factors such as education, holding more pro-conservative views and voting preferences, and income (King and Mossialos 2005), it is likely that these factors, and not access to private sector per se, are driving this relationship. Other studies have demonstrated that education and income may play a role in GP service use. Studies in the UK have found that less educated people see their GP more frequently (Goddard and Smith 2001), however studies of older people find higher education and higher income are associated with greater primary care use (Nelson et al. 2002; Evandrou 2003). Similar effects of income on physician service use among older people has been shown in Finland (Hakkinen and Luoma 1995). A recent analysis of determinants of primary care use in England also found that income significantly affects the probability of visiting a GP after controlling for health status (Bago d'uva 2004). Using similar methods with the general population, an earlier study using European Community Household Panel data found that the probability of a GP visit was prorich, despite the total number of visits being pro-poor (van Doorslaer et al. 2004). Inpatient care 9

10 In regards to inpatient care, the results suggest that the probability of an inpatient stay is greater among the higher income groups once need is controlled for. This pro-rich income-related inequity in inpatient care was not demonstrated in the general population (van Doorslaer et al. 2004; van Doorslaer and Masseria 2004), nor with a more recent analysis that pooled five years of data (Masseria et al. 2004), nor in a study of older people in London (Evandrou 2003). Unlike in the UK, in Portugal, Greece, Austria, Italy, Ireland, Germany and France, the better off are more likely to be admitted to hospital than the poor (Masseria et al. 2004). The strongest predictor of inpatient service use is health status, as shown in Canada (e.g. Roos 1997). However, higher income, having PMI and living in Scotland are also significantly positively associated with spending at least one night in hospital. Education was not found to be significantly associated with inpatient care, contrary to a study of older people in London (Nelson et al. 2002). It is interesting to note that when the individuals who used private inpatient care were excluded from the analysis (5% of the those who reported at least one night in hospital), the positive effect of income and PMI on the probability of having an inpatient stay (in NHS hospitals only) was no longer significant. Also the needs-adjusted income-related inequity (HI), although positive, was no longer significant (See Table 3). This finding suggests that while inpatient care in the NHS is equitable, the existence of a private sector creates income-related inequities in utilisation. Insert Table 3 here Outpatient care As found in the general UK population in 1996 (van Doorslaer et al. 2004) and a more recent analysis in England (Morris et al. 2005), outpatient service use was found to favour the rich older people over the poorer older people. Contrary to this finding, however, using similar methodology with the general UK population, outpatient service use was neither significantly pro-rich nor pro-poor (van Doorslaer and Masseria 2004). Among the factors significantly associated with outpatient visit are indicators of wealth and socio-economic status such as home ownership, higher education, and PMI. Also individuals in Northern Ireland are less likely than those in London to have an outpatient visit. Consistent with the present findings, socio-economic determinants of specialist use have been reported in other 10

11 countries, for example in Belgium (Van Der Heyden 2003), Spain (Fernandez 1996), the Netherlands (Mackenbach 1992) and Canada (Roos 1997). The regional variation observed, including the significantly lower likelihood of using outpatient care in Northern Ireland than London, and the lower, but not significantly, likelihood in regions outside of London could be attributed to the significantly greater proportion of private activity among consultants (as measured by the proportion of consultants on part-time contracts) in the south-east of England compared to the rest of the UK (King and Mossialos 2005). As with inpatient care, it was possible to separate the NHS and private sector outpatient activity. When NHS outpatient care was examined on its own, the significant pro-rich inequity remained, but decreased by about a third (see Table 3). Also, while the impact of income on service use remained (as revealed through the panel logit model), the association between PMI and outpatient service use disappeared, as seen with inpatient care. Therefore, unlike with inpatient care, within the NHS income-related inequity is still significant in use of outpatient care. Dental care The most significant degree of pro-rich inequity was found in dental care. Despite many older individuals having special dental needs, such as tooth decay and gum disease, 82% of the over- 60 age group receive no financial assistance with the significant user charges in the public sector (NHS patients have to pay 80% of the treatment costs) (Robinson et al. 2004). It is not surprising, therefore, that income-related inequity in use of dental care favouring the wealthy is substantial. The regional variation in dental care utilisation is worth noting, where individuals living in London are significantly more likely to have a dental check-up than those living in the rest of England (other than the south-east), Wales, Scotland and Northern Ireland. It is likely this observation results from a higher concentration of dentists practicing in London than the rest of the UK. The impact of supply on utilisation needs to be considered in further depth. In England, supply of health care services was found to have a positive impact on utilisation, and there was strong evidence of supply-based horizontal inequity (Morris et al. 2005). Unlike inpatient and outpatient care, when dental care in the public sector (NHS) is examined separately, income-related inequality favouring the higher income groups decreases only 11

12 slightly (see Table 3). Also, income and PMI remain significantly associated with dentist visit. This finding further supports the assertion that hefty user fees in the NHS are most likely deterring individuals on lower income from seeking dental care. Why do we observe significant pro-rich inequities in all service areas? It has been suggested that high utilisation among more privileged individuals may be explained by greater willingness to seek care and more appropriate responses to symptoms. Among the over-60 age group, those in higher socio-economic groups are significantly more likely to express immediate health seeking behaviour, as revealed through vignettes describing medical symptoms (Adamson et al. 2003). However it is likely a multitude of factors, both individual and societal, and both at the demand and supply-side, interact to create these existing inequalities. Limitations There are several potential biases in self-reported health measures that should be addressed. First, errors in self-reporting have been found to vary systematically across socio-economic groups (O Donnell and Propper 1991), which is consistent with the finding that lower socioeconomic groups tend to underreport longstanding illness (Adamson et al. 2003). This may then lead to underestimation of inequalities across socio-economic or income groups. Second, research from Canada reveals that although some older people report being in good health, they actually consumed more services over time than those reporting poor health (Black et al. 1995). This trend may be explained in part by the finding that despite the presence of many physical symptoms, older people often rate their overall health as good, suggesting a bias towards optimism (Dening et al. 1998). Therefore, one must interpret the results of the inequity analyses with caution, since all the need-related variables were based on self-report. In its defence, several studies have supported the validity of self-reported health status, demonstrating significant relationships with other measures of health status including physician assessments and utilisation data (Mossey and Shapiro 1982; Blaxter 1985). Also, self-reported health has been shown to predict future mortality better than other measures (Mossey and Shapiro 1982; Idler et al. 1990; Sutton et al. 1999), thus it is likely the best available proxy for need for health care. 12

13 Self-reported utilization may also be biased due to effects of social desirability or recall bias. Some researchers believe self-reporting of physicians visits may be unreliable (Roberts et al. 1996). Recall for hospital visits is generally better than that for physician contacts (Barer et al. 1982) however, using a one-year recall period is a common limitation of time series survey data. Also, as the data from the BHPS come from private households, the institutionalised individuals are not included in the analysis. Approximately 20% of those aged 85 and over live in institutions, and since entry to an institution is strongly affected by health, marital status, and socio-economic variables (Grundy and Sloggett 2003), the present analysis may be biased. Finally, as mentioned at the outset, this area of research is limited to the investigation of equity in utilisation of health care services, and not access to care. Therefore these studies may neglect important barriers to access and resulting inequalities that are important from a policy perspective. Conclusions The present study offers some support for the claim that health care service use is inequitable favouring those on higher income among the older population in the UK. Despite being in better health (in terms of number of health problems, self-reported health status, and activity limitations), wealthier older people are significantly more likely to see a doctor, have an outpatient visit, see a dentist and be admitted to hospital. Income-related inequities are most significant for specialist and dental care. While a recent analysis of equity in service use among the general British population only found significant inequity in dental care and not in other areas of medical care (van Doorslaer and Masseria 2004), it appears that income matters more among the older age groups. This is the first study to investigate fairness in service use specifically among older people and measure the presence and extent of inequity, thus it serves as a useful tool in evaluating the health system and highlighting weaknesses. As the highest consumers of health care, and with potentially more barriers to access, more attention should be paid to patterns of service use among older people and to addressing existing inequalities 13

14 Table 1. Health and socio-economic factors associated with health service use GP Outpatient Hospital Dentist Log of equivalised income Health/demographic indicators Good health Fair health Poor health Very poor health Female >3 health problems Health limits behaviour Disabled Ghq> Smoker Socio-economic indicators Owns home PMI Non-advanced qualifications Advanced qualifications SE England (not incl. London) Rest of England Wales Scotland N. Ireland Not married Widowed

15 Note: Adjusted Odds Ratios in bold are significant at p<0.05 Table 2. Income-related inequality in probability of GP, outpatient, inpatient and dentist use GP Outpatient Inpatient Dentist Cm ns Cn HI Note. NS is not significant. All others significant at p>0.01 Table 3. Income-related inequality in probability outpatient and inpatient care, NHS only Outpatient Inpatient Dentist Cm ns Cn HI ns Note. NS is not significant. All others significant at p>

16 References Adamson, J., Ben-Shlomo, Y., Chaturvedi, N., & Donovan, J. (2003). Ethnicity, socioeconomic position and gender-do they affect reported health-care seeking behaviour? Social Science and Medicine 57: Bago d'uva, T. (2004). Latent class models for use of primary care: evidence from 11 waves of the BHPS. European Conference on Health Economics, London School of Economics and Political Science. Barer, M. L., Manga, P., and Shillington, R. (1982). Income class and hospital use in Ontario. Toronto, Ontario Economic Council,. Black, C., Roos, N.P., Havens B., & MacWilliam L. (1995). Rising use of physician services by the elderly: the contribution of morbidity. Canadian Journal on Aging 24: Blaxter, M. (1985). Self-definition of health status and consulting rates in primary care. Quarterly Journal of Social Affairs 1: Dening, T. R., Chi, L.Y., Brayne, C., Huppert, F.A., Paykel, E.S., & O'Connor, D.W. (1998). Changes in self-rated health, disability and contact with services in a very elderly cohort: a 6-year follow-up study. Age and Ageing 27: Dixon, A., Le Grand, J., Henderson, J., Murray, R., Poteliakhoff, E. (2003). Is the NHS equitable? A review of the evidence. LSE Health and Social Care Discussion Paper No. 11. London, London School of Economics and Political Science. Dunlop, S., Coyte, P.C., & McIsaac, W. (2000). Socio-economic status and the utilization of physician's services: results from the Canadian National Population Health Survey. Social Science and Medicine 52: Evandrou, M. (2003). Growing old in London: socio-economic inequalities. SAGE Discussion paper no. 15. London, Stimulating Social Policy in an Ageing Society (SAGE). Fernandez, K., & Leon, D.A. (1996). Self-perceived health status and inequalities in use of health services in Spain. International Journal of Epidemiology 25: Finkelstein, M. M. (2001). Do factors other than need determine utilization of physician services in Ontario? Canadian Medical Association Journal 165: Goddard, M., & Smith, P. (2001). Equity of access to health care services: theory and evidence from UK. Social Science and Medicine 53: Grundy, E., & Sloggett, A. (2003). Health inequalities in the older population: the role of personal capital, social resources and socio-economic circumstances. Social Science and Medicine 56: Hakkinen, U., Luoma, K. (1995). Determinants of expenditure variation in health care and care of the elderly among Finnish municipalities. Health Economics 4: Idler, E., Kasl, S., & Lemke J. (1990). Self-evaluated health and mortality among the elderly in New Haven, Connecticut and Iowa and Washington counties, Iowa, American Journal of Epidemiology 131: 91. Keskimaki, I., Salinto, M., & Aro, S. (1995). Socioeconomic equity in Finnish hospital care in relation to need. Social Science and Medicine 41: King, D., Mossialos, E. (2005). The determinants of private medical insurance prevalence in England. Health Services Research. Le Grand, J. (1978). The distribution of public expenditure: the case of health care. Economica 45: Le Grand, J. (1991). The distribution of health care revisited: a commentary on Wagstaff, van Doorslaer and Paci, and O'Donnell and Propper. Journal of Health Economics 10:

17 Mackenbach, J. P. (1992). Socio-economic health differences in The Netherlands: a review of recent empirical findings. Social Science and Medicine 34(3): Masseria, C., Koolman, X., van Doorslaer, E. (2004). Income-related inequality in the probability of a hospital admission in Europe, Ecuity Project 3: Working Paper No. 13. Merlo, J., Gerdtham, U., Lynch, J., Beckman, A., Norlund, A., & Lithman, T. (2003). Social inequalities in health-do they diminish with age? International Journal of Equity in Health 2: 2. Morris, S., Sutton, M., Gravelle, H. (2005). Inequity and inequality in the use of health care in England: an empirical investigation. Social Science and Medicine 60: Mossey, J., & Shapiro, E. (1982). Self-rated health: a predictor of mortality among the elderly. American Journal of Public Health 72: 800. Nelson, T., Livingston, G., Knapp, M., Manela, M., Kitchen, G., & Katona, C. (2002). Slicing the health service cake: the Islington study. Age and Aging 31: Newbold, K. B., Eyles, J., & Birch, S. (1995). Equity in health care: methodological contributions to the analysis of hospital utilization within Canada. Social Science and Medicine 40: Nolan, B. (1994). General practitioner utilization in Ireland: the role of socio-economic factors. Social Science and Medicine 38: O'Donnell, O., & Propper, C. (1991). Equity and distribution of UK National Health Service resources. Journal of Health Economics 10: Propper, C. (1998). Who pays for and who gets health care: equity in the finance and delivery of health care in the UK. Nuffield Occasional Paper, Health Economics Series, paper 5. London, The Nuffield Trust. Propper, C., Eachus, J., Chan, P., Pearson, N., Davey Smith, G. (2005). Access to health care resources in the UK: the case of care for arthritis. Health Economics 14: Roberts, R. O., Bergstralh, E.J., Schmidt, I., & Jacobsen, S.J. (1996). Comparison of selfreported and medical record health care utilization measures. Journal of Clinical Epidemiology 49: Robinson, R., Patel, D., Pennycate, R. (2004). The Economics of Dental Care. London, Office of Health Economics. Roos, N. P. (1989). Predicting hospital utilization by the elderly. Medical Care Research 27: Roos, N. P., & Mustard, C.A. (1997). Variation in health and health care use by socioeconomic status in Winnipeg, Canada. Milbank Quarterly 75: Roos, N. P., & Shapiro, E. (1981). The Manitoba Longitudinal Study on Aging: preliminary findings on health care utilization by the elderly. Medical Care Research 19: Schellhorn, M., Stuck, A.E., Minder, C.E., & Beck, J.C. (2000). Health services utilization of elderly Swiss: evidence from panel data. Health Economics 9: Sutton, M., Carr-Hill, R., Gravelle, H., & Rice, N. (1999). Do measures of self-reported morbidity bias the estimation of the determinants of health care utilization? Social Science and Medicine 49: Sutton, M., Gravelle, H., Morris, S., Leyland, A., Windmeijer, F., Dibben, C. (2002). Allocation of resources to English areas. Edinburgh, Information and Service Division, Department of Health. Van Der Heyden, S., Demarest, J., Tafforeau, H., & Van Oyen. (2003). Socio-economic differences in the utilization of health services in Belgium. Health Policy 65: van Doorslaer, E., Koolman, X., Jones, A. (2004). Explaining income-related inequalities in doctor utilisation in Europe. Health Economics 13(7):

18 van Doorslaer, E., Masseria, C. (2004). Income-related inequality in the use of medical care in 21 OECD countries. Paris, OECD. van Doorslaer, E., Wagstaff, A., van der Burg, H., et al. (2000). Equity in the delivery of health care in Europe and the US. Journal of Health Economics 19(5): Whitehead, M., Evandrou, M., Haglund, B., & Diderichson, F. (1997). As the health divide widens in Sweden and Britain, what's happening to access to care? British Medical Journal 315: Notes 1 The data and tabulations used in this publication were made available through the ESRC Data Archive. The ESRC Research Centre originally collected the data on Microsocial Change at the University of Essex (now incorporated within the Institute for Social and Economic Research). Neither the original collectors of the data nor the Archive bear any responsibility for the analyses or interpretations presented here. 2 The Institute for Social and Economic Research website. [1 February 2005] 18