Sex of the GP — 20 years on

Over the past 20 years in Australia, the proportion of active general practitioners who are female has almost doubled, from 19.6%1 in 1990–1991 to 37.1% in 2009–2010.2 The increased participation of women in the GP workforce, in Australia and abroad, has led to examination of differences between male and female GPs.

Australian and overseas studies show that compared with males, female GPs are more likely to be younger,3,4 have longer consultations,5-9 work fewer hours,10 manage more female genital problems4,5,11 and social problems,5,12 see female patients6 (and female patients are more likely to prefer to see a female GP5,13,14), and have a patient-oriented consultative style.7,12,15 Female GPs are more likely to hold Fellowship of the Royal Australian College of General Practitioners (FRACGP),3 work in a group practice,3,4 provide preventive care,12,16 use psychosocial counselling,7 make referrals17 and prescribe fewer medications.15

However, most of these variables correlate. For example, if female patients prefer seeing female GPs, then female GPs will see more female patients and so will manage more female genital problems and social problems. If female GPs manage more of these problems, they would use more tests and psychosocial counselling. To test the extent to which GP sex independently affects clinical practice, we must control for all other variables.

One of the first studies18 controlling for these other factors was carried out in Australia using data from the 1990–1991 Australian Morbidity and Treatment Survey.1 It examined the effect of GP sex on patients’ “reasons for encounter” (RFEs) (adjusting for other GP and patient characteristics), and the effect of GP sex on the morbidity managed (adjusting for patients’ RFEs and other GP and patient characteristics).

After adjustment, patients seeing female GPs were more likely to have RFEs related to endocrine, psychological, pregnancy, female genital and social issues, and less likely to have RFEs related to the eyes or ears, or the musculoskeletal, respiratory, skin or male genital systems. After further statistical adjustment, female GPs were found to manage more general, psychological, pregnancy, female genital and social problems. The authors concluded that, while some differences were due to the mix of patients and their RFEs, other differences were “inherent” to the sex of the GP. In our study we investigated whether these inherent differences remain with today’s GPs, or if they have levelled out with the increased numbers of female GPs.

We analysed data from the Bettering the Evaluation and Care of Health (BEACH) program, April 2009 – March 2010. BEACH is a continuous, paper-based, national study of general practice activity in Australia. Sampling and methods have been described in detail elsewhere.2 In summary, each year about 1000 GPs from a national, rolling random sample (drawn by the Department of Health and Ageing) participate. Each participating GP provides patient and encounter information for 100 consecutive encounters with consenting, unidentified patients.

We weighted the April 2009 – March 2010 dataset (of about 100 000 encounters) to match the age–sex distribution of all GPs in the sample frame and for each GP’s activity level (measured by number of Medicare GP items claimed). The annual weighted BEACH encounter samples have repeatedly been demonstrated to be representative of all encounters claimed through Medicare.2

GP-related data elements include age (years), sex, years in practice, country of graduation, FRACGP status (yes or no), size and postcode of practice. Practice postcodes were used to define major cities using the Australian Standard Geographical Classification.19

The information collected at BEACH encounters includes patient age and sex, Indigenous status (self-reported by patient), non-English speaking background status (self-reported by patient), if a patient is new to the practice, place of consultation, up to three patient RFEs and up to four problems managed. GPs are instructed to describe the problem under management at the highest diagnostic level possible. For each problem, management data include medications, pathology and imaging orders and referrals. Problems and RFEs were classified according to the International Classification of Primary Care (version 2) (ICPC-2).20

Ethics approval for the BEACH program was obtained through the University of Sydney Human Research Ethics Committee and the Australian Institute of Health and Welfare Ethics Committee.

Between April 2009 and March 2010, 988 GPs participated in the BEACH program; 557 (56.4%) males and 431 (43.6%) females. After weighting, male GPs accounted for 63.0% and females 37.0% of encounters. There were 98 800 encounters recorded, and 101 349 after weighting.

Two decades after the previous study, our results show that clear differences in the problems managed by male and female GPs remain. The effect of other GP characteristics, the patient mix and patients’ demand for care account for some of these differences but, even after controlling for these factors, female GPs remain more likely to manage problems of a psychosocial, female-specific or general nature. These are similar findings to those of 20 years ago. Moving beyond the previous study, we also found differences in management practices that were inherent to GP sex, with female GPs more often using clinical treatments, pathology tests, referrals and imaging.

As found in previous studies, female GPs were younger,3,4,18 less likely to work as a sole practitioner3,4,18 and more likely to hold FRACGP status.3 These results are all consistent with the influx of new female registrars,21 as younger GPs are more likely to work in larger practices22 and hold FRACGP.22

As suggested by previous research, the patients at female-GP encounters were more likely to be female6,18 and younger.18 In the Australian Morbidity and Treatment Survey, seven out of 10 female-GP encounters were with female patients,18 and two decades later, this has not changed. This is surprising, since we had hypothesised that as female GPs increased their share of the GP workforce (and workload) they would see proportionally more male patients.

We also found that female GPs were less likely to see patients who were Commonwealth health care card holders, which probably relates to their seeing fewer older patients. Unlike the earlier study, we did not find male GPs doing more home visits, probably due to reduced frequency of home visits by both sexes.

Differences in patient RFEs given to male and female GPs remained after adjustment. This shows that patients, no matter what age or sex they are, still select GPs based on their sex for presentation of specific types of problems (especially sex-specific problems).

In our study, there was no difference in patient presentation rates of psychological or social RFEs to male or female GPs, either before or after adjustment, whereas in the previous study, patients presented more of both RFEs to female GPs, before and after adjustment. This suggests that patients now present with such problems equally often to both sexes, and that earlier preferences to take them to a female GP have disappeared. However, psychological and social problems were managed more by female GPs, before and after adjustment. This may suggest that female GPs are more attuned to the possible presence of such problems in their patients, which in turn may be related to their patient-focused consultative style.7,12,15

While female genital problems remained more likely to be managed by female GPs after adjustment, pregnancy and family planning problems did not, suggesting that their higher rate of management by female GPs is determined by patients rather than GPs.

Higher rates of clinical treatments and test orders add to the body of evidence that female GPs provide more preventive care.11,12,16 Individual encounters provided by female GPs take up more resources than those provided by male GPs, as female GPs have longer consultations,8,9 and our study shows that they refer and test more often.

A limitation of our study is that the data are cross-sectional, the denominator being individual problems or encounters. We know from the Australian medical labour force survey10 that male GPs work longer hours. This, combined with their shorter consultations,8,9 means that male GPs will perform more consultations in a year. If patients of male GPs visit them more often, it is possible that over 1 year, patients of male GPs will receive a similar amount of counselling and number of tests and referrals as those seeing female GPs. One must also consider whether statistically significant results presented are of genuine clinical significance.

This study shows that, other things being equal, for female and male GPs, patients bring different problems to be managed, different problems are managed and different management methods are used. It seems that the Australian general practice workforce is split by GP sex into groups practising different styles of medicine. The types of RFEs and problems managed more often by female GPs are similar to those found in data 20 years ago,18 suggesting that these differences may continue. Further research into how often male and female GPs see their patients in a year is needed to examine whether the higher use of resources per female-GP encounter (through additional tests and referrals) translates to higher or lower resource use overall.