Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Social Inequality: Utilisation of general practitioner services by socio‐economic disadvantage and geographic remoteness

Social Inequality: Utilisation of general practitioner services by socio‐economic disadvantage... Abstract Objective: To examine the association between socio-economic status (SES) and GP utilisation across Statistical Local Areas (SLAs) that differed in their geographic remoteness, and to assess whether Indigenous status and GP availability modified the association. Design: Retrospective analysis of Medicare data for all unreferred GP consultations (1996/97) for 952 SLAs comprising the six Australian States. Geographic remoteness was ascertained using the Area Remoteness Index of Australia (ARIA), and SES was measured by grouping SLAs into tertiles based on their Index of Relative Socioeconomic Disadvantage score. Main outcome measure: Age/sex standardised rates of GP utilisation for each SLA. Main results: In SLAs classified as ‘highly accessible’, rates of GP use were 10.8% higher (95% CI 5.7-16.0) in the most socioeconomically disadvantaged tertile after adjustment for Indigenous status and GP availability. A very different pattern of GP utilsation was found in ‘remote/very remote’ SLAs. After adjustment, rates of GP use in the most socio-economically disadvantaged tertile were 25.3% lower (95% CI 5.9-40.7) than in the most advantaged tertile. Conclusions: People in socio-economically disadvantaged metropolitan SLAs have higher rates of GP utilisation, as would be expected due to their poorer health. This is not true for people living in disadvantaged remote/very remote SLAs: in these areas, those most in need of GP services are least likely to receive them. Australia may lay claim to having a primary health care system that provides universal coverage, but we are still some way from having a system that is economically and geographically accessible to all. (Aust N Z J Public Health 2004; 28: 152-8) Gavin Turrell, Brian F. Oldenburg School of Public Health, Queensland University of Technology, Queensland Elizabeth Harris Centre for Health Equity Research, Training and Evaluation, South West Sydney Area Health Service, New South Wales Damien Jolley School of Health Sciences, Deakin University, Victoria large and growing Australian literature documents an association between socio-economic status (SES) and morbidity, with disadvantaged groups experiencing a greater burden of physiological disease, psychosocial illhealth, and accidents and injury.1 The poorer health status (and hence greater medical need) of disadvantaged groups is the main reason for their higher consultation rates for general practitioner (GP) services.2 In this study, we build on and extend earlier work3 by examining the relationship between SES, geographic remoteness, and GP use, and whether and to what extent this relationship is modif ied by Indigenous status and GP availability. In metropolitan areas, distributional and financial factors associated with the provision of GP services appear to facilitate equity of access; that is, higher rates of GP use among disadvantaged groups reflecting greater levels of need. GPs are disproportionately concentrated in metropolitan areas,4,5 and in one capital city at least, GP surgeries are more likely to be located in socio-economically disadvantaged areas.6 GPs in metropolitan areas charge lower fees than their counterparts in other areas7 and are more likely to bulk bill patients, 8 thus providing services at no direct (out-ofpocket) cost. 9 Further, travel distances between GPs and patients are likely to be shorter in metropolitan areas and transport more readily available. Based on this type of evidence, we hypothesise that socioeconomically disadvantaged groups in metropolitan areas will use GP services at a significantly higher rate than more advantaged groups. Evidence pertaining to the provision of GP services in non-metropolitan areas leads us to propose a very different hypothesis. Nonmetropolitan areas are underserved by GPs,4,5,10,11 doctors in these areas tend to charge more7,12 and are less likely to bulk bill.13-15 Moreover, GPs and patients in nonmetropolitan areas are often separated by large distances, and ready accessibility to transport is likely to be limited to a private vehicle. Given the greater barriers to GP use in non-metropolitan areas, we hypothesise that despite their greater need for medical services, socio-economically disadvantaged groups will exhibit lower rates of GP utilisation. Put differently, in non-metropolitan areas, we expect to find that socio-economic equity is a long way from being realised. Submitted: September 2003 Revision requested: December 2003 Accepted: January 2004 Correspondence to: Dr Gavin Turrell, School of Public Health, Queensland University of Technology, Victoria Park Road, Kelvin Grove, Queensland 4059. Fax: (07) 3864 3369; e-mail: g.turrell@qut.edu.au AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2004 VOL. 28 NO. 2 Social Inequality Use of GPs by disadvantage and location Our examination of the relationship between SES, geographic remoteness, and GP use considers two important additional factors: potential confounding by Indigenous status and the impact of GP availability on rates of utilisation. We know from previous research that SES and Indigenous status are correlated;16 that Indigenous peoples are proportionately more concentrated in rural and remote areas;17,18 that they experience much poorer health than the non-Indigenous; 19 and that they make less use of Medicare services, especially in remote areas.20 For these reasons, we examine the relationship between SES and GP use in metropolitan and non-metropolitan areas before and after accounting for the proportion of Indigenous peoples in these areas. We hypothesise that the impact of Indigenous peoples on the relation between SES and GP use will be minimal in metropolitan areas (where the proportion of Indigenous is smallest) and greatest in remote areas (where the Indigenous comprise a large proportion of the population). Finally, our assessment of the impact of GP availability on utilisation serves two purposes. First, adjusting the relationship between SES and GP use for GP availability represents a partial indirect test of the fundamental Medicare principle of equity, as higher rates of GP utilisation for disadvantaged groups independent of GP availability would suggest equity of access. Second, if socio-economic groups show a different utilisation pattern after equalising for GP availability, and if the nature and magnitude of this socio-economic patterning varies across geographic areas, it will suggest that the relationship between SES and GP use in metropolitan and non-metropolitan areas is influenced by qualitatively different factors. on information collected during the 1996 Census. Data relating to an SLA’s geographic remoteness were extracted from files compiled by DHAC and the National Key Centre for Social Applications of Geographic Information Systems,22 and pertain to each area’s boundaries as at 1996. Measures The socio-economic characteristics of each SLA were measured using the Index of Relative Socioeconomic Disadvantage (IRSD).23 The IRSD is derived from area attributes such as low income, low educational attainment, high levels of public sector housing, and high unemployment. We categorised IRSD scores into approximate tertiles: low SES (n=318, 33.4%), medium SES (n=312, 32.7%) and high SES (n=322, 33.8%). The geographic remoteness of each SLA was ascertained using the Accessibility/Remoteness Index of Australia (ARIA). 22 ARIA supersedes the Rural, Remote and Metropolitan Area (RRMA) classification, which was limited as an indicator of people’s access to services by its use of straight-line distance measurements between places, which did not capture the reality of road travel.24 By contrast, ARIA was developed using Geographic Information Systems and calculates the minimum distance that people have to travel to reach an urban centre containing basic services. We used the categorical ARIA measure, which groups together SLA with similar degrees of access and remoteness to services: 1. Highly Accessible SLA – areas with relatively unrestricted access to a wide range of goods and services and opportunities for social interaction. 2. Accessible SLA – areas with some restrictions to accessibility of some goods, services and opportunities for social interaction. 3. Moderately accessible SLA – areas with significantly restricted accessibility of goods, services and opportunities for social interaction. 4. Remote SLA – areas with very restricted accessibility of goods, services and opportunities for social interaction. 5. Very remote SLA – areas with very little accessibility of goods, services and opportunities for social interaction. Due to the small number of SLAs classif ied as ‘remote’ or ‘very remote’ when grouped into SES tertiles, ARIA categories four and five were combined. The proportion of Aboriginal and Torres Strait Islanders (ATSIs) in each SLA was calculated as the number of self-identified persons of ATSI origin (numerator) relative to the total population in each area (denominator). Across the 952 SLA, Indigenous persons comprised an average of 3.4% (SD 8.4) of the population of each area (median 1.2%, range 0-88.6%). GP workforce data relate only to services provided by general practitioners (including vocationally registered GP) for items defined as ‘unreferred attendances’, ‘enhanced primary care’ and ‘other unreferred attendances’ as per the Medicare Benefit Schedule.25 Services provided by specialist medical practitioners are not included. Availability of GP services in each SLA pertains to Methods Unit of analysis, data sources and reference periods This paper focuses on the socio-economic and remoteness characteristics of Statistical Local Areas (SLAs). SLAs correspond to council boundaries defined by Local Government Areas and cover the whole of Australia without gaps or overlaps.21 Due to data limitations, it was not possible to include all SLAs for each State and Territory (n=1,134). SLAs comprising the ACT (n=106) and Northern Territory (n=63) could not be used due to boundary discrepancies across data collections. Of the 965 SLAs comprising the six States, eight had incomplete data on one or more variables and six were classified as ‘off-shore/migratory’ areas: each of these SLAs were excluded. The f inal data set comprised 952 SLAs, or 98.65% of the total for the States. Data pertaining to GP availability and utilisation were extracted from data files released as part of the Social Health Atlas of Australia (SHAA) project.18 The availability and utilisation data relate to the years 1996 and 1996/97 respectively, and were originally sourced from the Medicare Statistics Section of the Commonwealth Department of Health and Aged Care (DHAC). Data pertaining to each SLA’s Indigenous and socio-economic characteristics were also obtained from SHAA files, and are based 2004 VOL. 28 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Turrell et al. Article the number of full-time equivalent (FTE) GPs per practice site. For the period 1995/96, GPs were classified by Medicare as fulltime if their Medical Benefits Schedule fee value was at or above a threshold of $71,725. For all GP earning in excess of this figure, the average income was calculated (estimated at $165,000 for the reference period), and those with a fee income at or above this average were assigned a value ‘1’. All other GPs were given a proportion of 1, having regard to the schedule fee income for the practitioner concerned relative to the average income of $165,000.18 We derived an estimate of FTE GP per 10,000 population for each SLA using 1996 Census counts for each area. GP utilisation data pertains to services provided at a surgery or clinic, a patient’s home, or an institution such as a hostel or nursing home. GP services not covered by Medicare were excluded.18 These comprised in-patient services and attendances at accident and emergency/casualty departments in public acute hospitals, community health services that did not bill their clients, and services provided by the Aboriginal Medical Service, private companies, armed forces, and the Royal Flying Doctor Service. Utilisation of GP services was measured in two parts: (i) the actual number of services provided in each SLA in 1996 and (ii) the number of services that would have been expected to occur in each SLA if the area had the same age and sex profile of the State in which it was located. The actual and expected number of GP services were then expressed as an age/sex standardised utilisation rate for each SLA. between ARIA, SES and GP utilisation was examined by negative binomial regression (with log-link) using the GENMOD procedure in SAS, with the actual number of GP visits per SLA being the response variable, and the log of the expected number of visits the offset term. The use of a negative binomial model to f it the data was necessary to allow for extra Poisson variation. The results are expressed as age/sex standardised utilisation ratios and 95% confidence intervals. Stage 2 used GLM to examine the relationship between SES, per cent ATSI and GP availability, stratified by each of the ARIA categories. Here the focus was on whether the association between SES, ATSI and GP availability was similar for SLA that differed in their geographic remoteness. Stage 3 used negative binomial regression to model the association between the SES tertiles and GP utilisation within categories of ARIA both before (Model 1) and after (Model 2) adjustment for per cent ATSI and FTE GP per 10,000 population. In addition, we formally tested whether the relation between SES and GP utilisation was differently patterned across the ARIA categories by modelling interactions between ARIA and SES. Results Table 1 shows that ARIA and SES were signif icantly associated with per cent ATSI, GP availability, and GP utilisation. The proportion of Indigenous peoples was, on average, smallest in ‘highly accessible’ and high SES areas, and largest in SLAs classified as ‘remote/very remote’ and socio-economically disadvantaged (i.e. bottom tertile). The mean number of FTE GPs was markedly higher in highly accessible areas compared with remote/ very remote SLAs, and SLAs in the most advantaged tertile had more GPs per head of population than those in the bottom tertile, although the difference was not large. Rates of GP utilisation were highest among highly accessible SLAs and lowest in remote/very remote areas. GP use was also significantly higher among socioeconomically advantaged SLAs. Analysis Analyses were undertaken in three stages. First, we investigated bivariate relationships between ARIA, SES and per cent ATSI and GP availability using the General Linear Models (GLM) procedure in SAS.26 Specifically, ARIA categories and SES tertiles were compared in terms of the per cent ATSI and the mean number of FTE GP per 10,000 population (with associated 95% confidence intervals). Also, as part of Stage 1, the bivariate relation Table 1: Per cent Aboriginal and Torres Strait Islander (ATSI), full-time equivalent GPs per 10,000 population, and GP utilisation, by Area Remoteness Index of Australia (ARIA) and Index of Relative Socioeconomic Disadvantage (IRSD). Statistical Local Areas n ARIA category Highly accessible Accessible Moderately accessible Remote/Very remote IRSD tertiled High SES Middle SES Low SES Per cent ATSIa Mean 95% CI Full-time equivalent GPs per 10,000 populationb Mean 95% CI GP utilisationc Standardised ratio 95% CI 0.5-1.8 1.5-3.4 3.3-5.7 12.6-15.4 0.3-2.0 1.3-3.1 6.1-7.9 7.3-7.9 5.2-6.1 4.9-6.0 3.0-4.3 6.4-7.2 5.8-6.6 5.6-6.4 – 75.3-82.6 75.9-84.6 55.1-62.2 – 89.5-99.4 89.8-99.6 Notes: (a) Based on data collected as part of the 1996 National Census. (b) Based on 1996 data from the Medicare Statistics Section of the (then) Commonwealth Department of Health and Aged Care. (c) Pertains to the period 1996/97. (d) Each tertile comprises approximately one-third of the Statistical Local Areas in the six Australian States. The high and low tertiles comprise the 33% least and most socio-economically disadvantaged SLA respectively. AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2004 VOL. 28 NO. 2 Social Inequality Use of GPs by disadvantage and location Table 2 examines the bivariate relation between SES and per cent ATSI, and SES and GP availability, within each ARIA category. Independent of ARIA, statistically signif icant associations were observed between SES and per cent ATSI, with socio-economically advantaged areas recording the lowest proportions and disadvantaged areas the highest. The smallest concentration of Indigenous peoples was found in highly accessible, high SES areas (mean 0.8%, 95% CI 0.6-1.0), and the largest concentrations were located in remote/very remote, socio-economically disadvantaged areas (mean 24.4%, 95% CI 19.8-28.8). There was no evidence of an association between SES and GP availability when the data were stratified by ARIA category. The greatest number of GPs per 10,000 population was found in highly accessible, high SES areas (mean 7.8, 95% CI 7.3-8.3), and the fewest in the most socio-economically disadvantaged remote/very remote areas (mean 3.1, 95% CI 2.4-3.8). Table 3 examines the relation between SES and GP utilisation within ARIA categories, prior to (Model 1) and after adjustment for per cent ATSI and GP availability (Model 2). Tests for interactions between ARIA and SES (not shown) were highly significant for both models (p<0.0001), indicating that the pattern of association between SES and GP utilisation was qualitatively different across the ARIA categories. Among highly accessible SLAs there was a significant negative association between SES and average rate of GP utilisation: low SES areas recorded higher utilisation rates than high SES areas independent of per cent ATSI and GP availability. No association was found between per cent ATSI and GP use; however, GP availability was significantly related with utilisation: a unit increase in the number of FTE GPs per 10,000 population was associated with a 1.5% increase in the use of GP services. Among accessible SLAs, areas classified as low SES exhibited a significantly higher rate of GP utilisation prior to adjustment for per cent ATSI and GP availability, but after adjustment, the socio-economic differences were attenuated to non-signif icance. Per cent ATSI and GP use were not significantly related; however, an association between GP availability and utilisation was again evident, with a unit increase in the number of FTE GPs resulting in a 4.5% increase in GP visits. Among moderately accessible SLAs, there were no significant differences between the SES tertiles in terms of GP utilisation rates (prior to or after adjustment), nor was per cent ATSI related to GP use, but utilisation rates were related with GP availability. Among remote/very remote SLAs, we witnessed a reversal of the relation between an area’s socio-economic characteristics and GP utilisation. For these SLAs, a strong positive association was found between the SES tertiles and the average rate of GP utilisation. Specifically, after adjustment, GP utilisation rates in areas classified as ‘low SES’ were 25.3% lower than in high SES areas. There was also a significant relation between per cent ATSI and GP use, with rates of utilisation being lower in areas with greater concentrations of Indigenous peoples. Further, there was a strong association between GP availability and GP utilisation, with a unit increase in the number of FTE GPs being associated with an estimated 14.4% increase in GP use. The patterning of results in Table 3 suggests that the strength of association between GP availability and utilisation increased as a function of an area’s remoteness. A correlation analysis showed support for this, with Table 2: Per cent Aboriginal and Torres Strait Islander (ATSI) and number of full-time equivalent GPs per 10,000 population, by tertiles of socio-economic disadvantage: within categories of the Area Remoteness Index of Australia (ARIA). ARIA categories and socio-economic tertilesa Highly accessible SLA High SES Middle SES Low SES Accessible SLA High SES Middle SES Low SES Moderately accessible SLA High SES Middle SES Low SES Remote/very remote SLA High SES Middle SES Low SES Statistical Local Areas Per cent ATSI Mean 95% CI 0.6-1.0 1.0-1.5 1.4-2.0 0.9-2.6 1.1-2.2 2.9-4.0 0-4.5 1.6-5.0 5.8-9.7 0-9.2 0-11.0 19.8-28.8 Full-time equivalent GPs per 10,000 population Mean 95% CI 7.8 7.2 7.6 4.7 5.7 5.9 4.5 5.9 5.5 3.8 4.5 3.1 7.3-8.3 6.5-7.8 6.9-8.3 4.0-5.5 5.2-6.2 5.4-6.4 3.3-5.7 5.1-6.7 4.5-6.4 2.8-4.9 3.5-5.4 2.4-3.8 Note: (a) Each tertile comprises approximately one-third of the Statistical Local Areas in the six Australian States. The high and low tertiles comprise the 33% least and most socio-economically disadvantaged SLAs respectively. 2004 VOL. 28 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Turrell et al. Article correlations between GP availability and GP use across the four ARIA categories being r=0.23 (p<0.0001), r=0.51 (p<0.0001), r=0.52 (p<0.0001) and r=0.76 (p<0.0001) respectively. Discussion Two important findings emerged from this study. First, the strength and direction of the relationship between SES and GP utilisation differed by ARIA category. Among ‘highly accessible’ and ‘accessible’ SLAs, a negative association between SES and GP use was found, with socio-economically disadvantaged areas having significantly higher rates of GP utilisation. Among ‘remote/very remote’ SLAs, however, a positive relationship was found: here, GP use was lowest in the most socio-economically disadvantaged SLAs and highest in the least disadvantaged areas. Moreover, these associations between SES and GP utilisation were evident even after accounting for the proportion of Indigenous peoples in each SLA. Indeed, across the four ARIA categories, adjustment for per cent ATSI attenuated the relationship between SES and GP use in a manner consistent with how Indigenous peoples are concentrated throughout Australia. 17,19 In highly accessible SLAs, where Indigenous peoples are least concentrated, adjustment for per cent ATSI made no difference to the association between SES and GP use, whereas in accessible, moderately accessible, and remote/very remote areas, adjustment for per cent ATSI accounted for 1.1%, 2.4%, and 10% of this association respectively. Second, within each ARIA category, greater numbers of FTE GPs per 10,000 population were associated with higher rates of utilisation, and the strength of this association tended to increase across the ARIA categories: it was weakest in highly accessible areas, intermediate in accessible and moderately accessible areas, and strongest in remote/very remote SLAs. These findings suggest that the relationship between GP availability and utilisation in areas with relatively unrestricted access to services is influenced by qualitatively different factors than the same relationship in areas with more restricted access. Similar findings have been reported and discussed elsewhere.27 Moreover, the patterning of the relationship between FTE GPs and utilisation across the ARIA categories seems to be indicating that utilisation is not strongly influenced by GP supply in highly accessible areas; however, supply clearly does influence utilisation in more remote areas, implying that rates of use would change with concomitant changes in supply. Study limitations A number of data limitations need to be considered when interpreting the results. The data pertained to services provided by Table 3: Modelling the association between the Index of Relative Socioeconomic Disadvantage and GP utilisation within categories of the Area Remoteness Index of Australia (ARIA). Model 1 Standardised ratio Highly accessible SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population Accessible SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population Moderately accessible SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population Remote/Very remote SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population 100.0 100.9 110.0 – – 100.0 102.9 110.3 – – 100.0 103.1 101.0 – – 100.0 95.1 69.2 – – 95% CI – 96.4-105.6 105.1-115.2 – – – 94.2-112.4 101.1-120.4 – – – 92.9-114.4 90.5-112.7 – – – 71.1-127.2 53.1-90.1 – – Model 2a Standardised ratio 95% CI – 97.4-106.4 105.7-116.0 98.2-100.9 100.9-102.0 – 91.4-107.1 95.9-112.6 99.6-101.7 103.3-105.7 – 89.6-107.8 86.5-105.7 99.9-100.9 102.4-104.6 – 65.3-103.2 59.3-94.1 98.8-99.8 110.7-118.2 Note: (a) Adjusted for the percentage of Aboriginal and Torres Strait Islander peoples in each SLA and the number of FTE GP per 10,000 population. AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2004 VOL. 28 NO. 2 Social Inequality Use of GPs by disadvantage and location GPs and not specialist medical practitioners. Rates of GP utilisation were calculated using Medicare data for unreferred services provided at a surgery or clinic, a patient’s residence, or an institution such as a nursing home. Services not covered by Medicare were excluded. Our results, therefore, represent a partial, but nevertheless important picture of how the availability and utilisation of GP services in Australia is patterned by an area’s socioeconomic and geographic characteristics. The data on which this study is based used the Medical Benefits fee schedule to estimate the number of FTE GPs per practice site, a method that is widely employed by health service researchers and others interested in medical workforce issues. 14 In recent years, however, measures of FTE have been criticised and alternative measures proposed.5 It is unlikely that the use of alternative measures would have fundamentally changed our results. All workforce databases are inaccurate to varying degrees, there is no universally accepted method for estimating the actual clinical contribution of GPs, and all methods of measuring GP contribution produce similar results, indicating that all may be a reasonable basis for estimating FTE.28 FTE measures are also considered to be problematic when data are compared on the basis of small areas such as SLAs, as many GPs practice in more than one area.5 Multiple counting, however, was unlikely to have changed the geographic and socio-economic patterning that was observed. SLAs were grouped into four broad geographic categories that shared the same remoteness and accessibility characteristics, and by definition were spatially distant. Thus a GP might work in two or more adjoining or closely spaced SLA within the same ARIA category, but they were unlikely to be practising in SLAs located in different categories. SES tertiles within an ARIA category comprised SLAs from all over Australia, and thus separated by large distances, minimising the likelihood of contamination sufficient to adversely affect the results. For SLAs that were contiguous or spatially close within the same ARIA category, it seems reasonable to assume that error associated with multiple counting would be similar for high and low SES areas, effectively equalising noise associated with this problem. Further, our examination of the association between SES and GP use tells us very little (if anything) about the appropriateness and quality of care received by different socio-economic groups. The results of previous research suggests that in addition to difficulties experienced by some disadvantaged groups in terms of accessing GP services, they are also less likely to receive care consistent with that provided to socio-economically advantaged groups. 29-31 services. Remote and very remote SLAs are under-served by GPs and residents of these areas probably travel long distances to attend the closest practice. GPs in rural and remote areas tend to charge more for their services7 and are less likely to bulk bill,8,1315 resulting in higher out-of-pocket costs. Among the general population in remote and very remote regions, these factors are likely to act as a disincentive to the use of GP services for all but the more serious conditions. Among socio-economically disadvantaged groups, however, who typically experience more straitened social and economic circumstances, these factors are likely to have a disproportionately negative influence on their propensity to seek the services of a GP. Low rates of GP use in socio-economically disadvantaged SLAs in remote/very remote areas might also reflect a different pattern of health service use compared with that which occurs in more advantaged areas. High SES groups probably have the necessary financial and material resources to visit GPs at their clinics or surgeries, whereas low SES groups might make greater use of community health services, nurse practitioners, and accident and emergency departments of local hospitals. None of these services were included in the calculation of the GP utilisation rates; thus the lower use of GP services in socio-economically disadvantaged SLA in remote/very remote areas might reflect these exclusions. More generally, lower rates of GP use in some remote/very remote regions might also be due to a greater reliance on alternative services (e.g. salaried medical practitioners, community services by nurses) provided by some State governments in sparsely populated areas where general practice is not viable. Further, geographic variability in GP use might partially result from urban-rural (and related occupational) differences in illness perceptions or responses to minor complaints or injuries.5 Small (non-signif icant) differences in the rate of clinical treatments that involve the administration of a medical certificate are also found between urban and rural areas (lower in the latter),32 although this is not likely to account for the geographic differences in utilisation observed in this present study. Finally, in areas with high concentrations of Indigenous people, lower rates of GP utilisation might also reflect the fact that in the absence of GPs, Indigenous groups may seek treatment instead from Aboriginal Community Controlled Health Services and other health care services not covered by Medicare (e.g. Aboriginal Health Worker or nurse practitioner).20 Conclusions In areas of Australia where GPs are in sufficient numbers, and are geographically and financially accessible, it seems that the Medicare principle of equity of access is being realised: given greater levels of morbidity, socio-economically disadvantaged groups in urban areas appear to make greater use of GP services. By contrast, in socio-economically disadvantaged areas in remote/ very remote regions, where GPs are least available and affordable and medical need is high, the principle of equity is apparently Explanation of findings Why was the relationship between SES and GP utilisation in remote/very remote SLAs opposite in direction to that observed in highly accessible and accessible ARIA categories? The different patterning possibly reflects concomitant geographic differences in social and economic factors, road distances to population centres, and the distribution, financing, and operation of GP 2004 VOL. 28 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Turrell et al. Article some way from being realised, and the Inverse Care Law seems to operate with considerable effect.33 Acknowledgements The first (and corresponding) author is supported by a National Health and Medical Research Council/National Heart Foundation Career Development Award (CR 013 0502). Thanks to Professor Mark Harris (University of New South Wales) for comments on an earlier draft, and to Mr John Glover (University of Adelaide) for assisting us with data-related issues. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australian and New Zealand Journal of Public Health Wiley

Social Inequality: Utilisation of general practitioner services by socio‐economic disadvantage and geographic remoteness

Loading next page...
 
/lp/wiley/social-inequality-utilisation-of-general-practitioner-services-by-UIYIHZs8dg

References (34)

Publisher
Wiley
Copyright
2004 The Public Health Association of Australia Inc
ISSN
1326-0200
eISSN
1753-6405
DOI
10.1111/j.1467-842X.2004.tb00929.x
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective: To examine the association between socio-economic status (SES) and GP utilisation across Statistical Local Areas (SLAs) that differed in their geographic remoteness, and to assess whether Indigenous status and GP availability modified the association. Design: Retrospective analysis of Medicare data for all unreferred GP consultations (1996/97) for 952 SLAs comprising the six Australian States. Geographic remoteness was ascertained using the Area Remoteness Index of Australia (ARIA), and SES was measured by grouping SLAs into tertiles based on their Index of Relative Socioeconomic Disadvantage score. Main outcome measure: Age/sex standardised rates of GP utilisation for each SLA. Main results: In SLAs classified as ‘highly accessible’, rates of GP use were 10.8% higher (95% CI 5.7-16.0) in the most socioeconomically disadvantaged tertile after adjustment for Indigenous status and GP availability. A very different pattern of GP utilsation was found in ‘remote/very remote’ SLAs. After adjustment, rates of GP use in the most socio-economically disadvantaged tertile were 25.3% lower (95% CI 5.9-40.7) than in the most advantaged tertile. Conclusions: People in socio-economically disadvantaged metropolitan SLAs have higher rates of GP utilisation, as would be expected due to their poorer health. This is not true for people living in disadvantaged remote/very remote SLAs: in these areas, those most in need of GP services are least likely to receive them. Australia may lay claim to having a primary health care system that provides universal coverage, but we are still some way from having a system that is economically and geographically accessible to all. (Aust N Z J Public Health 2004; 28: 152-8) Gavin Turrell, Brian F. Oldenburg School of Public Health, Queensland University of Technology, Queensland Elizabeth Harris Centre for Health Equity Research, Training and Evaluation, South West Sydney Area Health Service, New South Wales Damien Jolley School of Health Sciences, Deakin University, Victoria large and growing Australian literature documents an association between socio-economic status (SES) and morbidity, with disadvantaged groups experiencing a greater burden of physiological disease, psychosocial illhealth, and accidents and injury.1 The poorer health status (and hence greater medical need) of disadvantaged groups is the main reason for their higher consultation rates for general practitioner (GP) services.2 In this study, we build on and extend earlier work3 by examining the relationship between SES, geographic remoteness, and GP use, and whether and to what extent this relationship is modif ied by Indigenous status and GP availability. In metropolitan areas, distributional and financial factors associated with the provision of GP services appear to facilitate equity of access; that is, higher rates of GP use among disadvantaged groups reflecting greater levels of need. GPs are disproportionately concentrated in metropolitan areas,4,5 and in one capital city at least, GP surgeries are more likely to be located in socio-economically disadvantaged areas.6 GPs in metropolitan areas charge lower fees than their counterparts in other areas7 and are more likely to bulk bill patients, 8 thus providing services at no direct (out-ofpocket) cost. 9 Further, travel distances between GPs and patients are likely to be shorter in metropolitan areas and transport more readily available. Based on this type of evidence, we hypothesise that socioeconomically disadvantaged groups in metropolitan areas will use GP services at a significantly higher rate than more advantaged groups. Evidence pertaining to the provision of GP services in non-metropolitan areas leads us to propose a very different hypothesis. Nonmetropolitan areas are underserved by GPs,4,5,10,11 doctors in these areas tend to charge more7,12 and are less likely to bulk bill.13-15 Moreover, GPs and patients in nonmetropolitan areas are often separated by large distances, and ready accessibility to transport is likely to be limited to a private vehicle. Given the greater barriers to GP use in non-metropolitan areas, we hypothesise that despite their greater need for medical services, socio-economically disadvantaged groups will exhibit lower rates of GP utilisation. Put differently, in non-metropolitan areas, we expect to find that socio-economic equity is a long way from being realised. Submitted: September 2003 Revision requested: December 2003 Accepted: January 2004 Correspondence to: Dr Gavin Turrell, School of Public Health, Queensland University of Technology, Victoria Park Road, Kelvin Grove, Queensland 4059. Fax: (07) 3864 3369; e-mail: g.turrell@qut.edu.au AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2004 VOL. 28 NO. 2 Social Inequality Use of GPs by disadvantage and location Our examination of the relationship between SES, geographic remoteness, and GP use considers two important additional factors: potential confounding by Indigenous status and the impact of GP availability on rates of utilisation. We know from previous research that SES and Indigenous status are correlated;16 that Indigenous peoples are proportionately more concentrated in rural and remote areas;17,18 that they experience much poorer health than the non-Indigenous; 19 and that they make less use of Medicare services, especially in remote areas.20 For these reasons, we examine the relationship between SES and GP use in metropolitan and non-metropolitan areas before and after accounting for the proportion of Indigenous peoples in these areas. We hypothesise that the impact of Indigenous peoples on the relation between SES and GP use will be minimal in metropolitan areas (where the proportion of Indigenous is smallest) and greatest in remote areas (where the Indigenous comprise a large proportion of the population). Finally, our assessment of the impact of GP availability on utilisation serves two purposes. First, adjusting the relationship between SES and GP use for GP availability represents a partial indirect test of the fundamental Medicare principle of equity, as higher rates of GP utilisation for disadvantaged groups independent of GP availability would suggest equity of access. Second, if socio-economic groups show a different utilisation pattern after equalising for GP availability, and if the nature and magnitude of this socio-economic patterning varies across geographic areas, it will suggest that the relationship between SES and GP use in metropolitan and non-metropolitan areas is influenced by qualitatively different factors. on information collected during the 1996 Census. Data relating to an SLA’s geographic remoteness were extracted from files compiled by DHAC and the National Key Centre for Social Applications of Geographic Information Systems,22 and pertain to each area’s boundaries as at 1996. Measures The socio-economic characteristics of each SLA were measured using the Index of Relative Socioeconomic Disadvantage (IRSD).23 The IRSD is derived from area attributes such as low income, low educational attainment, high levels of public sector housing, and high unemployment. We categorised IRSD scores into approximate tertiles: low SES (n=318, 33.4%), medium SES (n=312, 32.7%) and high SES (n=322, 33.8%). The geographic remoteness of each SLA was ascertained using the Accessibility/Remoteness Index of Australia (ARIA). 22 ARIA supersedes the Rural, Remote and Metropolitan Area (RRMA) classification, which was limited as an indicator of people’s access to services by its use of straight-line distance measurements between places, which did not capture the reality of road travel.24 By contrast, ARIA was developed using Geographic Information Systems and calculates the minimum distance that people have to travel to reach an urban centre containing basic services. We used the categorical ARIA measure, which groups together SLA with similar degrees of access and remoteness to services: 1. Highly Accessible SLA – areas with relatively unrestricted access to a wide range of goods and services and opportunities for social interaction. 2. Accessible SLA – areas with some restrictions to accessibility of some goods, services and opportunities for social interaction. 3. Moderately accessible SLA – areas with significantly restricted accessibility of goods, services and opportunities for social interaction. 4. Remote SLA – areas with very restricted accessibility of goods, services and opportunities for social interaction. 5. Very remote SLA – areas with very little accessibility of goods, services and opportunities for social interaction. Due to the small number of SLAs classif ied as ‘remote’ or ‘very remote’ when grouped into SES tertiles, ARIA categories four and five were combined. The proportion of Aboriginal and Torres Strait Islanders (ATSIs) in each SLA was calculated as the number of self-identified persons of ATSI origin (numerator) relative to the total population in each area (denominator). Across the 952 SLA, Indigenous persons comprised an average of 3.4% (SD 8.4) of the population of each area (median 1.2%, range 0-88.6%). GP workforce data relate only to services provided by general practitioners (including vocationally registered GP) for items defined as ‘unreferred attendances’, ‘enhanced primary care’ and ‘other unreferred attendances’ as per the Medicare Benefit Schedule.25 Services provided by specialist medical practitioners are not included. Availability of GP services in each SLA pertains to Methods Unit of analysis, data sources and reference periods This paper focuses on the socio-economic and remoteness characteristics of Statistical Local Areas (SLAs). SLAs correspond to council boundaries defined by Local Government Areas and cover the whole of Australia without gaps or overlaps.21 Due to data limitations, it was not possible to include all SLAs for each State and Territory (n=1,134). SLAs comprising the ACT (n=106) and Northern Territory (n=63) could not be used due to boundary discrepancies across data collections. Of the 965 SLAs comprising the six States, eight had incomplete data on one or more variables and six were classified as ‘off-shore/migratory’ areas: each of these SLAs were excluded. The f inal data set comprised 952 SLAs, or 98.65% of the total for the States. Data pertaining to GP availability and utilisation were extracted from data files released as part of the Social Health Atlas of Australia (SHAA) project.18 The availability and utilisation data relate to the years 1996 and 1996/97 respectively, and were originally sourced from the Medicare Statistics Section of the Commonwealth Department of Health and Aged Care (DHAC). Data pertaining to each SLA’s Indigenous and socio-economic characteristics were also obtained from SHAA files, and are based 2004 VOL. 28 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Turrell et al. Article the number of full-time equivalent (FTE) GPs per practice site. For the period 1995/96, GPs were classified by Medicare as fulltime if their Medical Benefits Schedule fee value was at or above a threshold of $71,725. For all GP earning in excess of this figure, the average income was calculated (estimated at $165,000 for the reference period), and those with a fee income at or above this average were assigned a value ‘1’. All other GPs were given a proportion of 1, having regard to the schedule fee income for the practitioner concerned relative to the average income of $165,000.18 We derived an estimate of FTE GP per 10,000 population for each SLA using 1996 Census counts for each area. GP utilisation data pertains to services provided at a surgery or clinic, a patient’s home, or an institution such as a hostel or nursing home. GP services not covered by Medicare were excluded.18 These comprised in-patient services and attendances at accident and emergency/casualty departments in public acute hospitals, community health services that did not bill their clients, and services provided by the Aboriginal Medical Service, private companies, armed forces, and the Royal Flying Doctor Service. Utilisation of GP services was measured in two parts: (i) the actual number of services provided in each SLA in 1996 and (ii) the number of services that would have been expected to occur in each SLA if the area had the same age and sex profile of the State in which it was located. The actual and expected number of GP services were then expressed as an age/sex standardised utilisation rate for each SLA. between ARIA, SES and GP utilisation was examined by negative binomial regression (with log-link) using the GENMOD procedure in SAS, with the actual number of GP visits per SLA being the response variable, and the log of the expected number of visits the offset term. The use of a negative binomial model to f it the data was necessary to allow for extra Poisson variation. The results are expressed as age/sex standardised utilisation ratios and 95% confidence intervals. Stage 2 used GLM to examine the relationship between SES, per cent ATSI and GP availability, stratified by each of the ARIA categories. Here the focus was on whether the association between SES, ATSI and GP availability was similar for SLA that differed in their geographic remoteness. Stage 3 used negative binomial regression to model the association between the SES tertiles and GP utilisation within categories of ARIA both before (Model 1) and after (Model 2) adjustment for per cent ATSI and FTE GP per 10,000 population. In addition, we formally tested whether the relation between SES and GP utilisation was differently patterned across the ARIA categories by modelling interactions between ARIA and SES. Results Table 1 shows that ARIA and SES were signif icantly associated with per cent ATSI, GP availability, and GP utilisation. The proportion of Indigenous peoples was, on average, smallest in ‘highly accessible’ and high SES areas, and largest in SLAs classified as ‘remote/very remote’ and socio-economically disadvantaged (i.e. bottom tertile). The mean number of FTE GPs was markedly higher in highly accessible areas compared with remote/ very remote SLAs, and SLAs in the most advantaged tertile had more GPs per head of population than those in the bottom tertile, although the difference was not large. Rates of GP utilisation were highest among highly accessible SLAs and lowest in remote/very remote areas. GP use was also significantly higher among socioeconomically advantaged SLAs. Analysis Analyses were undertaken in three stages. First, we investigated bivariate relationships between ARIA, SES and per cent ATSI and GP availability using the General Linear Models (GLM) procedure in SAS.26 Specifically, ARIA categories and SES tertiles were compared in terms of the per cent ATSI and the mean number of FTE GP per 10,000 population (with associated 95% confidence intervals). Also, as part of Stage 1, the bivariate relation Table 1: Per cent Aboriginal and Torres Strait Islander (ATSI), full-time equivalent GPs per 10,000 population, and GP utilisation, by Area Remoteness Index of Australia (ARIA) and Index of Relative Socioeconomic Disadvantage (IRSD). Statistical Local Areas n ARIA category Highly accessible Accessible Moderately accessible Remote/Very remote IRSD tertiled High SES Middle SES Low SES Per cent ATSIa Mean 95% CI Full-time equivalent GPs per 10,000 populationb Mean 95% CI GP utilisationc Standardised ratio 95% CI 0.5-1.8 1.5-3.4 3.3-5.7 12.6-15.4 0.3-2.0 1.3-3.1 6.1-7.9 7.3-7.9 5.2-6.1 4.9-6.0 3.0-4.3 6.4-7.2 5.8-6.6 5.6-6.4 – 75.3-82.6 75.9-84.6 55.1-62.2 – 89.5-99.4 89.8-99.6 Notes: (a) Based on data collected as part of the 1996 National Census. (b) Based on 1996 data from the Medicare Statistics Section of the (then) Commonwealth Department of Health and Aged Care. (c) Pertains to the period 1996/97. (d) Each tertile comprises approximately one-third of the Statistical Local Areas in the six Australian States. The high and low tertiles comprise the 33% least and most socio-economically disadvantaged SLA respectively. AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2004 VOL. 28 NO. 2 Social Inequality Use of GPs by disadvantage and location Table 2 examines the bivariate relation between SES and per cent ATSI, and SES and GP availability, within each ARIA category. Independent of ARIA, statistically signif icant associations were observed between SES and per cent ATSI, with socio-economically advantaged areas recording the lowest proportions and disadvantaged areas the highest. The smallest concentration of Indigenous peoples was found in highly accessible, high SES areas (mean 0.8%, 95% CI 0.6-1.0), and the largest concentrations were located in remote/very remote, socio-economically disadvantaged areas (mean 24.4%, 95% CI 19.8-28.8). There was no evidence of an association between SES and GP availability when the data were stratified by ARIA category. The greatest number of GPs per 10,000 population was found in highly accessible, high SES areas (mean 7.8, 95% CI 7.3-8.3), and the fewest in the most socio-economically disadvantaged remote/very remote areas (mean 3.1, 95% CI 2.4-3.8). Table 3 examines the relation between SES and GP utilisation within ARIA categories, prior to (Model 1) and after adjustment for per cent ATSI and GP availability (Model 2). Tests for interactions between ARIA and SES (not shown) were highly significant for both models (p<0.0001), indicating that the pattern of association between SES and GP utilisation was qualitatively different across the ARIA categories. Among highly accessible SLAs there was a significant negative association between SES and average rate of GP utilisation: low SES areas recorded higher utilisation rates than high SES areas independent of per cent ATSI and GP availability. No association was found between per cent ATSI and GP use; however, GP availability was significantly related with utilisation: a unit increase in the number of FTE GPs per 10,000 population was associated with a 1.5% increase in the use of GP services. Among accessible SLAs, areas classified as low SES exhibited a significantly higher rate of GP utilisation prior to adjustment for per cent ATSI and GP availability, but after adjustment, the socio-economic differences were attenuated to non-signif icance. Per cent ATSI and GP use were not significantly related; however, an association between GP availability and utilisation was again evident, with a unit increase in the number of FTE GPs resulting in a 4.5% increase in GP visits. Among moderately accessible SLAs, there were no significant differences between the SES tertiles in terms of GP utilisation rates (prior to or after adjustment), nor was per cent ATSI related to GP use, but utilisation rates were related with GP availability. Among remote/very remote SLAs, we witnessed a reversal of the relation between an area’s socio-economic characteristics and GP utilisation. For these SLAs, a strong positive association was found between the SES tertiles and the average rate of GP utilisation. Specifically, after adjustment, GP utilisation rates in areas classified as ‘low SES’ were 25.3% lower than in high SES areas. There was also a significant relation between per cent ATSI and GP use, with rates of utilisation being lower in areas with greater concentrations of Indigenous peoples. Further, there was a strong association between GP availability and GP utilisation, with a unit increase in the number of FTE GPs being associated with an estimated 14.4% increase in GP use. The patterning of results in Table 3 suggests that the strength of association between GP availability and utilisation increased as a function of an area’s remoteness. A correlation analysis showed support for this, with Table 2: Per cent Aboriginal and Torres Strait Islander (ATSI) and number of full-time equivalent GPs per 10,000 population, by tertiles of socio-economic disadvantage: within categories of the Area Remoteness Index of Australia (ARIA). ARIA categories and socio-economic tertilesa Highly accessible SLA High SES Middle SES Low SES Accessible SLA High SES Middle SES Low SES Moderately accessible SLA High SES Middle SES Low SES Remote/very remote SLA High SES Middle SES Low SES Statistical Local Areas Per cent ATSI Mean 95% CI 0.6-1.0 1.0-1.5 1.4-2.0 0.9-2.6 1.1-2.2 2.9-4.0 0-4.5 1.6-5.0 5.8-9.7 0-9.2 0-11.0 19.8-28.8 Full-time equivalent GPs per 10,000 population Mean 95% CI 7.8 7.2 7.6 4.7 5.7 5.9 4.5 5.9 5.5 3.8 4.5 3.1 7.3-8.3 6.5-7.8 6.9-8.3 4.0-5.5 5.2-6.2 5.4-6.4 3.3-5.7 5.1-6.7 4.5-6.4 2.8-4.9 3.5-5.4 2.4-3.8 Note: (a) Each tertile comprises approximately one-third of the Statistical Local Areas in the six Australian States. The high and low tertiles comprise the 33% least and most socio-economically disadvantaged SLAs respectively. 2004 VOL. 28 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Turrell et al. Article correlations between GP availability and GP use across the four ARIA categories being r=0.23 (p<0.0001), r=0.51 (p<0.0001), r=0.52 (p<0.0001) and r=0.76 (p<0.0001) respectively. Discussion Two important findings emerged from this study. First, the strength and direction of the relationship between SES and GP utilisation differed by ARIA category. Among ‘highly accessible’ and ‘accessible’ SLAs, a negative association between SES and GP use was found, with socio-economically disadvantaged areas having significantly higher rates of GP utilisation. Among ‘remote/very remote’ SLAs, however, a positive relationship was found: here, GP use was lowest in the most socio-economically disadvantaged SLAs and highest in the least disadvantaged areas. Moreover, these associations between SES and GP utilisation were evident even after accounting for the proportion of Indigenous peoples in each SLA. Indeed, across the four ARIA categories, adjustment for per cent ATSI attenuated the relationship between SES and GP use in a manner consistent with how Indigenous peoples are concentrated throughout Australia. 17,19 In highly accessible SLAs, where Indigenous peoples are least concentrated, adjustment for per cent ATSI made no difference to the association between SES and GP use, whereas in accessible, moderately accessible, and remote/very remote areas, adjustment for per cent ATSI accounted for 1.1%, 2.4%, and 10% of this association respectively. Second, within each ARIA category, greater numbers of FTE GPs per 10,000 population were associated with higher rates of utilisation, and the strength of this association tended to increase across the ARIA categories: it was weakest in highly accessible areas, intermediate in accessible and moderately accessible areas, and strongest in remote/very remote SLAs. These findings suggest that the relationship between GP availability and utilisation in areas with relatively unrestricted access to services is influenced by qualitatively different factors than the same relationship in areas with more restricted access. Similar findings have been reported and discussed elsewhere.27 Moreover, the patterning of the relationship between FTE GPs and utilisation across the ARIA categories seems to be indicating that utilisation is not strongly influenced by GP supply in highly accessible areas; however, supply clearly does influence utilisation in more remote areas, implying that rates of use would change with concomitant changes in supply. Study limitations A number of data limitations need to be considered when interpreting the results. The data pertained to services provided by Table 3: Modelling the association between the Index of Relative Socioeconomic Disadvantage and GP utilisation within categories of the Area Remoteness Index of Australia (ARIA). Model 1 Standardised ratio Highly accessible SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population Accessible SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population Moderately accessible SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population Remote/Very remote SLA High SES Middle SES Low SES Per cent ATSI FTE GP per 10,000 population 100.0 100.9 110.0 – – 100.0 102.9 110.3 – – 100.0 103.1 101.0 – – 100.0 95.1 69.2 – – 95% CI – 96.4-105.6 105.1-115.2 – – – 94.2-112.4 101.1-120.4 – – – 92.9-114.4 90.5-112.7 – – – 71.1-127.2 53.1-90.1 – – Model 2a Standardised ratio 95% CI – 97.4-106.4 105.7-116.0 98.2-100.9 100.9-102.0 – 91.4-107.1 95.9-112.6 99.6-101.7 103.3-105.7 – 89.6-107.8 86.5-105.7 99.9-100.9 102.4-104.6 – 65.3-103.2 59.3-94.1 98.8-99.8 110.7-118.2 Note: (a) Adjusted for the percentage of Aboriginal and Torres Strait Islander peoples in each SLA and the number of FTE GP per 10,000 population. AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2004 VOL. 28 NO. 2 Social Inequality Use of GPs by disadvantage and location GPs and not specialist medical practitioners. Rates of GP utilisation were calculated using Medicare data for unreferred services provided at a surgery or clinic, a patient’s residence, or an institution such as a nursing home. Services not covered by Medicare were excluded. Our results, therefore, represent a partial, but nevertheless important picture of how the availability and utilisation of GP services in Australia is patterned by an area’s socioeconomic and geographic characteristics. The data on which this study is based used the Medical Benefits fee schedule to estimate the number of FTE GPs per practice site, a method that is widely employed by health service researchers and others interested in medical workforce issues. 14 In recent years, however, measures of FTE have been criticised and alternative measures proposed.5 It is unlikely that the use of alternative measures would have fundamentally changed our results. All workforce databases are inaccurate to varying degrees, there is no universally accepted method for estimating the actual clinical contribution of GPs, and all methods of measuring GP contribution produce similar results, indicating that all may be a reasonable basis for estimating FTE.28 FTE measures are also considered to be problematic when data are compared on the basis of small areas such as SLAs, as many GPs practice in more than one area.5 Multiple counting, however, was unlikely to have changed the geographic and socio-economic patterning that was observed. SLAs were grouped into four broad geographic categories that shared the same remoteness and accessibility characteristics, and by definition were spatially distant. Thus a GP might work in two or more adjoining or closely spaced SLA within the same ARIA category, but they were unlikely to be practising in SLAs located in different categories. SES tertiles within an ARIA category comprised SLAs from all over Australia, and thus separated by large distances, minimising the likelihood of contamination sufficient to adversely affect the results. For SLAs that were contiguous or spatially close within the same ARIA category, it seems reasonable to assume that error associated with multiple counting would be similar for high and low SES areas, effectively equalising noise associated with this problem. Further, our examination of the association between SES and GP use tells us very little (if anything) about the appropriateness and quality of care received by different socio-economic groups. The results of previous research suggests that in addition to difficulties experienced by some disadvantaged groups in terms of accessing GP services, they are also less likely to receive care consistent with that provided to socio-economically advantaged groups. 29-31 services. Remote and very remote SLAs are under-served by GPs and residents of these areas probably travel long distances to attend the closest practice. GPs in rural and remote areas tend to charge more for their services7 and are less likely to bulk bill,8,1315 resulting in higher out-of-pocket costs. Among the general population in remote and very remote regions, these factors are likely to act as a disincentive to the use of GP services for all but the more serious conditions. Among socio-economically disadvantaged groups, however, who typically experience more straitened social and economic circumstances, these factors are likely to have a disproportionately negative influence on their propensity to seek the services of a GP. Low rates of GP use in socio-economically disadvantaged SLAs in remote/very remote areas might also reflect a different pattern of health service use compared with that which occurs in more advantaged areas. High SES groups probably have the necessary financial and material resources to visit GPs at their clinics or surgeries, whereas low SES groups might make greater use of community health services, nurse practitioners, and accident and emergency departments of local hospitals. None of these services were included in the calculation of the GP utilisation rates; thus the lower use of GP services in socio-economically disadvantaged SLA in remote/very remote areas might reflect these exclusions. More generally, lower rates of GP use in some remote/very remote regions might also be due to a greater reliance on alternative services (e.g. salaried medical practitioners, community services by nurses) provided by some State governments in sparsely populated areas where general practice is not viable. Further, geographic variability in GP use might partially result from urban-rural (and related occupational) differences in illness perceptions or responses to minor complaints or injuries.5 Small (non-signif icant) differences in the rate of clinical treatments that involve the administration of a medical certificate are also found between urban and rural areas (lower in the latter),32 although this is not likely to account for the geographic differences in utilisation observed in this present study. Finally, in areas with high concentrations of Indigenous people, lower rates of GP utilisation might also reflect the fact that in the absence of GPs, Indigenous groups may seek treatment instead from Aboriginal Community Controlled Health Services and other health care services not covered by Medicare (e.g. Aboriginal Health Worker or nurse practitioner).20 Conclusions In areas of Australia where GPs are in sufficient numbers, and are geographically and financially accessible, it seems that the Medicare principle of equity of access is being realised: given greater levels of morbidity, socio-economically disadvantaged groups in urban areas appear to make greater use of GP services. By contrast, in socio-economically disadvantaged areas in remote/ very remote regions, where GPs are least available and affordable and medical need is high, the principle of equity is apparently Explanation of findings Why was the relationship between SES and GP utilisation in remote/very remote SLAs opposite in direction to that observed in highly accessible and accessible ARIA categories? The different patterning possibly reflects concomitant geographic differences in social and economic factors, road distances to population centres, and the distribution, financing, and operation of GP 2004 VOL. 28 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Turrell et al. Article some way from being realised, and the Inverse Care Law seems to operate with considerable effect.33 Acknowledgements The first (and corresponding) author is supported by a National Health and Medical Research Council/National Heart Foundation Career Development Award (CR 013 0502). Thanks to Professor Mark Harris (University of New South Wales) for comments on an earlier draft, and to Mr John Glover (University of Adelaide) for assisting us with data-related issues.

Journal

Australian and New Zealand Journal of Public HealthWiley

Published: Apr 1, 2004

There are no references for this article.