At the heart of health workforce shortages and surpluses is the inflexibility of health professional training and roles between different types of doctors and between doctors and other health professionals. Arguments that inflexibility of roles and increasing specialisation are necessary to maintain quality are valid up to a point, but there is a clear trade-off between the additional quality gained from specialisation and the quality foregone because of inflexibility. Nevertheless, there is an increasing consensus in Australia that we are past that point, that the gains in health outcomes from specialisation are now less than the health outcomes foregone created by inflexibility.
Re-balancing the distribution of doctors between generalism and specialism is difficult and there is no single solution. It is true that the preferences of doctors do not seem to match community needs, with high demand and competition for specialties already well supplied and demand for specialties in most need.2 The attractiveness of specialties does not seem to change in response to changes in community needs. There is no ‘market’ mechanism to ensure that changes in demand for specialties by the community leads to changes in the relative supply of generalists or specialists.
Doctors’ preferences are influenced by a range of factors including their own characteristics, skills and aptitudes, experiences during undergraduate and postgraduate training, and the expected rewards and other characteristics of working in a specialty. Unfortunately, there is little evidence on which of these factors are most important in driving specialty choice, and little evidence that employers or governments can change doctors’ choices.3 The perceived rewards for specialisation are not only financial but also status, prestige and reputation, and are influenced by culture and competition during medical training. Nevertheless, remuneration has been identified as a key barrier to increased generalism.4
The aim of this short paper is to discuss issues related to the relative remuneration of GPs and specialists while recognising that 1) the promotion of generalism outside of general practice is also a major issue and that other factors to promote generalism, such as reform of education and training, also need to be considered, and 2) the distinction between GPs and specialists is too simple and doesn’t recognise the blurring of generalist and specialist roles within many medical specialties.5
Evidence on earnings differentials
In Australia, there are large earnings differentials between qualified specialists and general practitioners. In 2012, data from the Medicine in Australia: Balancing Employment and Life (MABEL) panel survey6 shows that GPs’ average earnings (before tax but after practice expenses) was $193 958, whereas the average earnings of specialists was $360 000, a gap of $166 041. Specialists’ earnings are 1.86 times higher than GPs. Earnings per hour were $114 for GPs and $184 for specialists. These gaps have remained steady over time. Figure 1 shows trends in real earnings (adjusted for inflation) between 2008 and 2012. Hours worked have also fallen slightly for both groups over time, such that the gap in hourly earnings is also relatively stable over time. Though generally steady, there is some evidence that specialists’ earnings have fallen slightly more than GPs’ earnings, reducing the gap from $185 076 ($78 per hour) in 2008 to $166 042 ($72 per hour) in 2012.
Figure 1. Real average annual earnings of GPs and specialists (2008–2012)
Source: MABEL data Waves 1 (2008) to Wave 5 (2012). Data are inflated to 2013 prices using the ABS Wage Index for healthcare professionals. Average annual earnings are before tax but after practice expenses from the question: ‘What are your (approximate) TOTAL PERSONAL earnings from ALL of the work you do as a doctor? (If possible, base this on your last personal income tax return or payslip.) This should be your personal earnings rather than total practice earnings.’ Data are from just under 3000 GPs and around 3500 specialists each year. Cross-sectional survey response weights are used to help ensure national representativeness each year (details in Yan et al6).
These averages mask large differences and variations in earnings. For example, in 2012 the median earnings of specialists in the bottom 25% of their earnings distribution were $217 524 (working a median 36 hours per week); whereas the median earnings of GPs in the top 25% of the earnings distribution was $248 291 (working a median 47 hours per week). So the highest earning GPs were earning more than the lowest earning specialists, partly because they were working more hours per week. The earnings gap depends on a range of factors, including rurality. Figure 2 compares average specialist earnings with average GP earnings in metropolitan and non-metropolitan areas. Although GPs in the most rural areas (Australian Standard Geographical Classification (ASGC) 3-5) earn about $55 000 more than GPs in metropolitan areas, the gap between GPs and specialists is still high at about $122 000. Controlling for a range of other factors, such as hours worked, gender and practice size, Cheng et al7 found that GPs in ASGC 3-5 earned, on average, 11% more than GPs in metropolitan areas.
Figure 2. Real average annual earnings of specialists and GPs, by geographical area (2008–2012)
Source: see Figure 1 for data sources. Data for GPs in this figure are from around 1750 GPs per year in major cities, from around 600 GPs per year in inner regional areas and from around 450 GPs in ASGC 3-5.
In exploring the reasons for the earnings gap between GPs and specialists using MABEL data from 2008, Cheng et al7 showed that the annual earnings of GPs are lower than those of specialists because GPs work fewer hours, are more likely to be women, are less likely to undertake after-hours or on-call work and have lower returns to experience.7 The latter finding indicates that specialists are paid more as they accumulate more years of experience, whereas GPs do not, partly because specialists in hospitals are salaried and progress up a pay scale with some certainty.
Gender is an important part of this, as female doctors are paid less than males, even after controlling for differences in hours worked, productivity, experience and a range of other factors. This has been shown in many countries and industries, including for doctors in Australia.8 This research shows that having children has a major role in influencing the gender-gap in earnings. However, even after accounting for the fact that there are more female GPs than specialists, and differences in a whole range of other characteristics between GPs and specialists, Cheng et al7 showed that up to half of the difference in earnings remains unexplained. This may be due to historical differences in fee schedules that provide higher rewards for procedural work.
Do earnings differentials influence career choices?
Do these large differences in earnings influence doctors when choosing a specialty? Using data from the US, a handful of studies have shown that expected future earnings, together with other factors such as flexibility of hours and the level of educational debt, influence specialty choice.9–12
In Australia, a survey of junior doctors already enrolled in vocational specialist and GP training were asked retrospectively about their reasons for specialty choice; only 16% rated ‘financial prospects’ of the specialty as important, compared with other factors.13 A study in 2008 using MABEL data used a discrete choice experiment administered to junior doctors before they chose their specialty training program.14 This may more accurately reflect preferences and does not suffer from post-justification bias of doctors who couldn’t get into their most preferred specialty program.
Sivey et al14 found that expected future earnings was an important factor, along with opportunities for procedural work, hours worked, control over hours worked, on-call, opportunities for academic work and continuity of care. Future earnings were more important for those with higher levels of educational debt. In a policy simulation, the authors found that if GP earnings were to increase by $50 000 per year to $230 000, then the probability of choosing general practice would increase by 10.5 percentage points from 39.9% to 50.4%. An increase in procedural work and opportunities for academic work had similar-sized effects, suggesting that other ways to make general practice more attractive as a career should also be considered. This study used hypothetical choices, and MABEL data is currently being used to examine actual choices as junior doctors can be tracked moving into specialty training programs using the longitudinal data. Remuneration seems to matter along with other factors.
Policy responses
It seems clear that to make primary care a more attractive career, additional and targeted investment is needed. Changing the relative funding between primary and hospital care is therefore an important policy lever – the issue is how this funding should be used to improve population health.
Option 1: changing fee relativities
It is clear that in many countries, procedural work is more highly rewarded than ‘cognitive’ work, and in part this is due to the historical evolution of fee schedules. Changing remuneration by altering the relativities in the Medicare Benefits Schedule (MBS) is controversial and has been tried before in Australia but was not adopted.15 In a system where doctors can charge patients what the market will bear, changing the relative levels of subsidy may not have much of an effect on relative earnings as providers can make up for lost earnings by charging higher prices or increasing the volume of care provided. We know that increasing Medicare subsidies can lead to increases in fees and co-payments.16 Even if relative prices could be changed, there is evidence from the US when Medicare changed fee relativities that physicians and surgeons can increase the volume of services provided to maintain their incomes.17
A better solution in Australia would be a more vigilant and evidence-based Medical Services Advisory Committee that is more cautious about funding new devices and technologies, rewarding procedural work by suitably trained GPs and reviewing the existence of, and subsidies for, the MBS items that have never been reviewed.18
Option 2: new funding streams
A second option is to provide more funding to primary care in other ways to enhance the attractiveness of general practice as a career. As many young and increasingly female GPs do not want to be self-employed and run a business, the growth of corporate practice continues to dominate the sector especially in areas of low socioeconomic status. There are now more salaried GPs than principals and it is important for governments to consider new contractual options for GPs that offer a better career structure, better career progression and new payment systems that reward high-quality care for chronic disease. Investing in academically focused health centres and in primary care-led procedural services is also important.
An alternative that has not been considered is for Medicare and/or Medicare locals to enter into different types of contracts with primary care practices. Should practices be given a choice to opt out of fee-for-service and receive their revenue from Medicare in different ways, such as a mix of fixed base payments, some fee-for-service and some payments for quality? Different types of Medicare contracts could be offered, including fee-for-service. These options could build on the Practice Incentive program, where payments are already being made to practices and would contribute greatly to reducing the red tape of fee-for-service, provide more stable jobs for the future generation of mainly female GPs who do not all want to be principals or business owners, and provide alternatives to for-profit corporate ownership.
Conclusion
Although expected levels of remuneration do play a part in doctors’ choice of specialty, it is difficult to see how the earnings gap can be reduced in the short or medium term by altering fees or changing payment systems. The feasibility of these options would require strong political will and support from the medical profession. Providing a higher proportion of health funding to primary care is more achievable and realistic, such that the growth in primary care funding is higher than the growth in hospital funding. This could filter down into GPs’ and other primary care professionals’ earnings and make primary care a more attractive sector in which to work. The danger is that this extra funding would be taken as profit by an increasingly corporate sector and not invested in services that improve health outcomes or increase the attractiveness of the sector. Governments need to think of alternative contractual models for primary care practices and ensure that any additional funding in primary care is used to its best effect.
Competing Interests: None.
Provenance and peer review: Commissioned; externally peer reviewed.
Acknowledgements
Anthony Scott was funded by an ARC Future Fellowship Grant. This paper used data from the MABEL longitudinal survey of doctors conducted by the University of Melbourne and Monash University. Funding for MABEL comes from the National Health and Medical Research Council (Health Services Research Grant: 2008–2011, and Centre for Research Excellence in Medical Workforce Dynamics: 2012–2016) with additional support from the Department of Health (in 2008) and Health Workforce Australia (in 2013).