Socioeconomic disadvantage and the practice location of recently Fellowed Australian GPs: a cross-sectional analysis

Keywords: family practice, health accessibility, health equity.

In Australia, as in many countries, areas of greatest socioeconomic disadvantage have higher rates of illness and poorer health outcomes (Starfield et al. 2005; Gordon et al. 2016). Research findings consistently reflect the ‘inverse care law’, whereby those who have the greatest need for health care receive the lowest levels of care (Hart 1971; Furler et al. 2002; Mercer et al. 2021). There are many factors that influence this relationship, but access to quality primary health care in an affordable and timely manner has been shown to attenuate this inequity (Furler et al. 2002; Starfield et al. 2005; Norbury et al. 2011).

Access to best-practice primary care is the most important health systems determinant of the health of communities, and of health care efficiency and equity (Starfield et al. 2005). Equitable access is predicated on sufficient numbers and the distribution of adequately educated and trained primary care practitioners, including GPs. A recent comparison of the healthcare systems of 11 developed countries ranked Australia as having one of the best, including placing first on measures of equity (encompassing timeliness, affordability and patient engagement; Schneider et al. 2021). However, Australian data consistently shows a disparity of GP care between those most in need and those with lesser need. There is evidence that, even in Australia, this disparity is widening (The Royal Australian College of General Practitioners 2019).

One of the means by which the Australian Government attempts to redress this access disparity is via workforce distribution considerations within the general practice vocational training program (Campbell et al. 2011; Hays and Morgan 2011; Department of Health 2020a). Although overarching vocational GP training policy indirectly addresses this via rural training obligations (Department of Health 2020a), regional training organisations (RTOs) may implement further training location policies that include consideration of the socioeconomic status (SES) of practice locations (GP Synergy 2016). This is based on the premise that experiencing training in areas of socioeconomic disadvantage will encourage trainees to build an ongoing career in these areas, and can help prepare GPs to respond to the medical, psychological and social needs within Australia’s evolving primary healthcare system. Although there is evidence for this contention in the context of rural practice location (Campbell et al. 2011; McGrail et al. 2016), it remains unclear whether, and/or how, vocational training practice location and experience relate to GPs remaining in areas of high socioeconomic disadvantage in their early post-Fellowship years. However, there is evidence to suggest that even within GP training, registrars are opting to move to practice in areas of social advantage, with later stage (Term 3) registrars less likely than Term 1 registrars to practice in areas of greater socioeconomic disadvantage (D. Moad, A. Tapley, A. Fielding, M. van Driel, E. Holliday, J. Ball, A. Davey, K. FitzGerald, N. Spike, P. Magin, unpubl. obs).

Establishing the SES of the clinical practice location of recently Fellowed GPs and its association with practice location during training will inform workforce planning and structural aspects of training programs.

The aim of this study was to establish the relationship between GP training location and the subsequent practice of early career GPs (recent alumni of the Australian GP vocational training program) in areas of social disadvantage. In addition, the study explored other associations of SES of alumni practice location.

This analysis took place within the New alumni EXperiences of Training, and independent Unsupervised Practice (NEXT-UP) study (Magin et al. 2019).

NEXT-UP was a cross-sectional questionnaire-based study of recently Fellowed GPs. Participants were alumni of three Australian GP RTOs, namely GP Synergy, Eastern Victoria GP Training and General Practice Training Tasmania, who achieved Fellowship between January 2016 and July 2018 inclusive and were between 6-months and 2 years post-Fellowship.

The NEXT-UP methodology has been fully described elsewhere (Magin et al. 2019). Briefly, participants were invited to complete a questionnaire regarding their current practice, training experiences and perceptions of the utility of their training in preparing them for unsupervised practice. Although this was essentially a cross-sectional study design, where consent was provided questionnaire data were linked to data collected by registrars’ RTOs as part of the participants’ vocational training program.

The outcome factor was whether the participant’s current main GP location was in an area of low SES. This was defined by the Socio-Economic Indexes for Areas – Index of Relative Disadvantage 2016 (SEIFA-IRSD) (Australian Bureau of Statistics 2018a) using alumni’s current main practice postcode. Where postcodes straddled more than one SEIFA-IRSD score, the highest score was used.

National SEIFA-IRSD deciles (Australian Bureau of Statistics 2018a, 2018b) were used, with lower deciles indicating lower SES and therefore higher relative socioeconomic disadvantage. For analysis purposes in this study, the outcome factor of an area of lower SES was defined as the practice’s postcode being within the four lowest SEFIA-IRSD deciles.

Independent variables were those related to the participants, their current general practice and their training. Participant-related variables included sex, age, country of medical qualification (Australia or international), relationship status (along with spousal employment status, dependent children), regional, rural or urban schooling before university and number of years in an Australian hospital before commencing their first general practice training term (GPT1).

The current practice variable included was whether the participant previously worked at their current practice.

Training-related variables included training in a practice located rurally and/or in an area of lower SES (using practice postcodes), any part-time training, examination performance (failure in any Fellowship examination), leave taken during training and year of Fellowship.

Rural training experience (yes/no) was defined as having undertaken ≥13 full-time equivalent (FTE) weeks in a practice location classified as regional, rural, remote or very remote, using the Modified Monash Model (MMM) system (i.e. MMM2–7 locations, henceforth referred to as a ‘rurally located practice/s’; Department of Health 2020b).

Low SES training location experience (yes/no) was defined as having undertaken ≥13 FTE weeks in areas of relative disadvantage based on SEIFA-IRSD deciles within the study sample (Australian Bureau of Statistics 2018a), with Deciles 1–4 considered to be relatively disadvantaged.

This was a cross-sectional analysis. Descriptive statistics included frequencies for categorical variables and the mean ± s.d. for continuous variables. The frequencies of categorical variables were compared between outcome categories using Chi-squared tests for all variables, except when Fisher’s exact test was used (due to an expected count <5 in ≥25% of cells). For continuous variables, means were compared using t-tests. Missing data were handled by using complete case analysis.

Logistic regression was used to compare the log-odds of a registrar working in a practice in a low SES area (Deciles 1–4) compared with working in a practice in higher SES areas (Deciles 5–10). Effect sizes were presented as odds ratios (ORs) with 95% confidence intervals (CIs).

First, univariate analyses were conducted on each covariate, with the outcome. Of the 14 covariates of interest, five were univariately associated with the outcome (P < 0.20) and were considered for inclusion in the multiple regression model. After the multivariable model was fitted, all covariates were significant (P < 0.20), so no testing for removal from the final multivariable model was performed.

Associations were considered significant at the conventional two-sided P < 0.05 level. Analyses were completed using Stata 14.2 (StataCorp, College Station, TX, USA) and SAS version 9.4 (SAS Institute, Cary, NC, USA).

Ethics approval was provided by the University of Newcastle Human Research Ethics Committee (Reference H-2018-0333).

This analysis included 354 participants (response rate 28%), of whom 337 (95%) were currently working in clinical general practice. Of these 337 participants, 317 had current practice postcode data available for the outcome, with 83 (26%) currently practicing in areas within the four deciles of greatest socioeconomic disadvantage (i.e. SEIFA-IRSD Deciles 1–4; Fig. 1).

Fig. 1.  Current practice location of recently Fellowed GPs according to SEIFA-IRSD decile.

Characteristics of alumni, their current practices and their training variables are presented in Table 1.

Table 1.  Characteristics of recently Fellowed GPs participating in this study (n = 354)
Data are presented as n (%) or mean ± s.d. ACRRM, Australian College of Rural and Remote Medicine; EVGPT, Eastern Victoria General Practice Training; GPTT, General Practice Training Tasmania; RACGP, Royal Australian College of General Practitioners

Fourteen independent variables were included in the univariate analyses, and characteristics associated with current practice location within the lowest four SEIFA-IRSD deciles versus the six highest deciles are presented in Table 2.

Table 2.  Characteristics associated with recently Fellowed GPs currently practicing in an area of greater socioeconomic disadvantage (n = 317^A)
Unless indicated otherwise, data are given as the mean ± s.d. or n (%). SES, socioeconomic status

In the adjusted model, statistically significant multivariable associations of recently Fellowed GPs currently practicing in lowest SEIFA-IRSD deciles practices included having worked at the current practice during vocational training (OR 2.99; 95% CI 1.60–5.61), having worked in a practice in a low SES location during training (OR 2.21; 95% CI 1.24–3.92) and having worked in a rural practice location during training (OR 3.14; 95% CI 1.69–5.81). Full univariate and multivariable regression results are presented in Table 3.

Table 3.  Univariate and multivariable logistic regression analysis of associations of recently Fellowed GPs currently practicing in an area of greater socioeconomic disadvantage
SES, socioeconomic status

We found that a modest proportion of recently Fellowed GPs (26%) currently practice in areas of greater socioeconomic disadvantage (the four lowest national SEIFA-IRSD deciles). Only 10% of alumni practised in the lowest two SEIFA-IRSD deciles. This figure is difficult to compare to the distribution of longer-established GPs because, to the best of the authors’ knowledge, this is the first Australian study to investigate the SES of GPs’ practice location.

Our primary study finding, namely that training location has a statistically significant association with post-Fellowship practice location, is consistent with previous research that found significant associations for practice location based on rurality (Wilkinson et al. 2003; McGrail et al. 2016; Fielding et al. 2019).

We also found a significant association of alumni currently working in areas of greater socioeconomic disadvantage post-Fellowship with having worked at their current practice during training. This is consistent with earlier work in a registrar population, where we found that currently working in a lower SES area practice was associated with having worked at that practice previously (D. Moad, A. Tapley, A. Fielding, M. van Driel, E. Holliday, J. Ball, A. Davey, K. FitzGerald, N. Spike, P. Magin, unpubl. obs).

This study of early career GPs has demonstrated a strong association of alumni having practiced in areas of greater socioeconomic disadvantage during training with their current practice location being socioeconomically disadvantaged. This will be important in considerations of the training–workforce nexus. From a training perspective, these results provide evidence to support the role of training programs in helping to address GP workforce issues (by providing evidence for prioritising placements in lower SEIFA-IRSD areas, identified areas of need for service provision).

Because patients from lower SES environments have greater burdens of disease (including greater multimorbidity) and particular, often complex, biopsychosocial needs (Australian Institute of Health and Welfare 2020), enhanced support of registrars undertaking training in areas of socioeconomic disadvantage could be considered. Notably, there are educational benefits to the registrar, as well as societal benefit, in working in lower SES area practices (D. Moad, A. Tapley, A. Fielding, M. van Driel, E. Holliday, J. Ball, A. Davey, K. FitzGerald, N. Spike, P. Magin, unpubl. obs). As with training in rural practice (Tapley et al. 2020), there is evidence from our earlier work of practices in lower SES areas being particularly rich education and training environments (D. Moad, A. Tapley, A. Fielding, M. van Driel, E. Holliday, J. Ball, A. Davey, K. FitzGerald, N. Spike, P. Magin, unpubl. obs).

Considering the above, the significant association of current practice and training raises the possibility that training in areas of socioeconomic disadvantage may ‘open the eyes’ of registrars to working in an area of higher need. Training in such practices may provide registrars with an opportunity to experience the singular demands and rewards of practising in socioeconomically disadvantaged areas. These positive experiences may, in turn, encourage trainees to practice post-Fellowship in a disadvantaged area, including in the same practice. These associations may also be a reflection of trainees’ pre-existing commitment to social equity in health care, and a commitment to seek out training opportunities and post-Fellowship employment in high-need areas. Either way, training providers fulfil an essential role in supporting and promoting training in areas of greater socioeconomic disadvantage.

However, this study highlights that there is a lesser proportion of recently Fellowed GPs taking up practice in the areas of greatest socioeconomic disadvantage. Almost three-quarters of recently Fellowed GPs (74%) opted to practice in areas of less disadvantage. The highest proportion of alumni currently practice in the top two deciles (17% and 19%, respectively), servicing areas that reflect the highest levels of socioeconomic advantage. This reflects continuation of the shift away from more socioeconomically disadvantaged areas during GP vocational training demonstrated in a previous study (D. Moad, A. Tapley, A. Fielding, M. van Driel, E. Holliday, J. Ball, A. Davey, K. FitzGerald, N. Spike, P. Magin, unpubl. obs). It is also consistent with the broader literature on socioeconomic disparity and workforce maldistribution/supply and will only contribute to Australia’s concerns regarding equity in health care (Kattererl 2011; Schneider et al. 2021). Although Australian workforce policies have focused on maldistribution in areas such as rurality and vulnerable groups (e.g. Aboriginal and Torres Strait Islander people, aged care and disability), low SES, which is interwoven across these factors, needs further attention if Australia is to improve its equity in health care.

Future research is required to understand in greater detail the experiences of recently Fellowed GPs in choosing their post-Fellowship practice location, and what influence their training had on their practice location choices. Further research could look beyond the 2-year post-Fellowship period, to further understand longer-term patterns for GPs’ practice location.

Further research in this area would assist in understanding the experiences of those working in areas of greater socioeconomic disadvantage, and the role that recently Fellowed GPs play in meeting the ongoing needs of these communities.

A strength of this study is that it was conducted across three Australian RTOs. The three RTOs account for 43.4% of all registrars in general practice terms in Australia (Taylor et al. 2021) and have a demographic and geographic presence across the range of Australian GP vocational training. The study included many variables, and the use of RTO-held contemporaneously collected training data provides reliability beyond that of the retrospective questionnaire-elicited demographic and training data. The modest response rate of 28% is also consistent with contemporary cross-sectional research in GP populations (Bonevski et al. 2011), and the majority of demographic data reflects that of the wider GP population.

A limitation of this study is that, because of the cross-sectional design, we can demonstrate association, but not causation.

Another possible limitation is the use of practice location to define levels of disadvantage, as opposed to possible measures of individual patients’ SES. However, although practices themselves may see a mix of patients from different socioeconomic backgrounds, the focus of this study was to investigate the current practice location of recently Fellowed GPs, and the potential impact that training variables may have on post-Fellowship practice location.

Access to quality primary health care within Australia and other countries is inequitable, with areas of higher socioeconomic disadvantage experiencing a disproportionate burden of disease. Low numbers and poor retention rates of recently Fellowed GPs suggests the continuation of the inverse care law in Australia. Government and training provider policies aimed at increasing GP registrar training in areas of socioeconomic disadvantage have the potential to contribute to a more equitable workforce distribution.

The data that support this study cannot be shared due to ethical or privacy reasons.

The authors declare no conflicts of interest.

This study was supported by a competitive Educational Research Grant of the Royal Australian College of General Practitioners (Grant no. ERG020). In-kind support was provided by GP Synergy, Eastern Victoria General Practice Training and General Practice Training Tasmania.