Female medical leadership: cross sectional study Editorial by Showalter
BMJ 1999; 318 doi: https://doi.org/10.1136/bmj.318.7176.91 (Published 09 January 1999) Cite this as: BMJ 1999;318:91- Kari J Kværner (K.j.kvarner{at}ioks.uio.no), associate professora,
- Olaf G Aasland, professorb,
- Grete S Botten, professorc
- aDepartment of Otorhinolaryngology, Ullevǻl University Hospital, N-0407 Oslo, Norway
- bResearch Institute, Norwegian Medical Association, N-0107 Oslo
- cCentre for Health Administration, University of Oslo, National Hospital, N-0027 Oslo
- Correspondence to: Dr Kværner
- Accepted 1 October 1998
Abstract
Objective: To assess the relation between male and female medical leadership.
Design: Cross sectional study on predictive factors for female medical leadership with data on sex, age, specialty, and occupational status of Norwegian physicians.
Setting: Oslo,Norway.
Subjects: 13844 non-retired Norwegian physicians.
Main outcome measure: Medical leaders, defined as physicians holding a leading position in hospital medicine, public health, aacademic medicine, or private health care.
Results: 14.6% (95% confidence interval 14.0% to 15.4%) of the men were leaders compared with 5.1% (4.4% to 5.9%) of the women. Adjusted for age men had a higher estimated probability of leadership in all categories of age and job, the highest being in academic medicine with 0.57(0.42to 0.72) for men aged over 54years compared with 0.39(0.21to 0.63) for women in the same category. Among female hospital physicians there was a positive relation between the proportion of women in their specialty and the probability of leadership.
Conclusion: Women do not reach senior positions as easily as men. Medical specialties with high proportions of women have more female leaders.
Editorial by Showalter
Key messages
Men have significantly higher probability for all leadership positions
Women are underrepresented in higher medical administrative positions
Leadership positions for women are especially prevalent in public health, where working hours are regulated
In hospitals the proportion of female physicians in the specialty is a determinant for female leadership
Introduction
The Norwegian labour market has adjusted to the increased proportion of career women with an equal opportunities policy of positive sex discrimination in favour of women and improved social benefits. Dual career families are now the norm rather than the exception; current employment rates for women and men in Norway are 56% and 70%, respectively. Maternity leave has improved, and working women are entitled to full cash benefits for 42weeks or 80% benefits for 52weeks, and men are entitled to 2weeks' paternity leave. The public day care coverage is 60%, and parents are allowed 10days' paid leave a year per child for illness in children. Despite this liberal equity policy and an increase in the number of women working, relatively few women hold positions of leadership in industry, business, or the public sector, including health care.1
Up to the 1990s, when the proportion of women among medical students reached 50%, women had the advantage of positive sex discrimination in medical schools. Now, though universities use this to increase the proportion of women in academic posts, there is no such policy in the appointment of hospital doctors and public health positions.
Data from Finland suggest that female physicians are largely excluded from decision making in medical research,2–4 but few studies have explored the relation between specialty and career choices and female leadership in medicine. As the purpose of the Norwegian equal opportunities policy has been to favour women in the workforce and strengthen their participation in all employment positions to obtain equity, it is of interest to study the extent to which female Norwegian doctors have succeeded in reaching senior positions.
With data on specialty and occupational status of Norwegian physicians we examined the distribution of female and male leaders with respect to age and type of job. We then estimated the probability of leadership for different categories of female and male physicians. Finally, we explored whether leadership among female hospital physicians is related to the proportion of women in the specialty.
Methods
Study population
Data on all Norwegian physicians were made available from the masterfile of the Norwegian Medical Association, which is a complete register of all Norwegian physicians, including the 6% who are not members of the association. All those active in October 1997with information on sex, age, specialty, and job category were included, comprising 13844 physicians, 3939women and 10131 men.
Variables
On the basis of the differences in career patterns between clinical and non-clinical doctors and differences between public and private health care, employment posts were grouped into four categories: hospital medicine (including teaching hospitals), public health (state, county, or municipality), academic medicine (university positions outside hospitals), and private health care (including occupational health care).
Leadership was the outcome of interest. For the purpose of this study leaders were defined as chief executive officers or senior consultants of the hospitals or clinical departments, full time and part time professors and principal research officers, and central and local government medical officers. Within private health care and industrial concerns chief medical physicians and medical directors in research departments were defined as leaders.
We assumed that the potential for support from colleagues —such as mentoring and mutual information on career opportunities —was likely to be associated with membership of a specialist category, particularly in hospitals. To explore the impact of the proportion of women in a specialty on female leadership we stratified the female physicians working in hospital into four groups: non-specialists, those belonging to the nine specialties with women comprising 2% to 14% of physicians (neurological surgery (2%), general surgery (5%), plastic surgery and dental, oral, and maxillofacial surgery (7%), otorhinolaryngology (8%), medical genetics (10%), clinical chemistry (11%), general medicine (12%), and community medicine (14%)), those belonging to the 10specialties with women comprising 15% to 24% (anaesthetics (16%), ophthalmology and clinical neurophysiology (19%), immunology (20%), rheumatology (21%), general practice (22%), paediatrics (23%), and radiotherapy, physiotherapy and microbiology-bacteriology (24%)), and finally those belonging to the nine specialties with more than 24% of women (neurology (25%), radiology (26%), dermatovenereology (28%), occupational medicine, pathological anatomy, and psychiatry (29%), obstetrics and gynaecology (32%), and child psychiatry (59%)). The female specialists working outside hospital were excluded from the analyses because they, to a large extent, are employed in working settings where leadership is irrelevant according to our criteria about mentoring support from colleagues.
Statistical analyses
Numbers of leaders with respect to sex, age, and job category are described as proportions with 95% confidence intervals. The probabilities of leadership for different sex and categories of age and job as well as the probability of leadership among female hospital physicians belonging to groups with different proportions of female specialists were estimated with logistic regression models.
Results
Of all physicians in the study, 28% were women. The mean age for women was 40.5years and for men 46.5years. The proportion of male physicians who were leaders was 14.6% (95% confidence interval 14.0% to 15.4%) compared with 5.1% (4.4% to 5.9%) for women. The mean age for female leaders was 49.7(48.6to 50.7) years and for male leaders 52.8(52.4to 53.3) years. Table 1 shows the numbers of physicians in the different age and job categories and the corresponding proportion of leaders.
The probability of being a leader was estimated by a logistic model with sex, age, and job categories as additive covariates. Table 2 gives the estimated probabilities of leadership for the different combinations of age and job situation. As some of these groups are quite small the confidence intervals are correspondingly wide —for instance, the 95% confidence interval for the estimated probability of 0.39for female academics over 54years was 0.21 to 0.63whereas the corresponding interval for the male probability of 0.57was 0.42to 0.72.The age gradient, assumed to be the same for men and women, was significant as was the sex effect, with an adjusted odds ratio of 2.0(1.6to 2.4) in favour of men.
A possible period effect (that it is easier for female physicians to reach leadership now than, for example, 10to 15years ago) was explored by including the interaction of women under 45years versus men and older women in the above model. This covariate had no significant effect, however, and the estimated probabilities in table 2 remained almost the same.
The 2033women currently employed in hospital medicine were included in a logistic regression analysis of the possible impact of proportion of women in different specialties on the probability of female leadership. Covariates were age, the proportion of women in the specialty (squared), and a dummy variable for specialist or non-specialist. All three covariates showed significant effects, with t values of 7.13,4.0,and 3.45,respectively. The figure shows the estimated probabilities of leadership for five major specialties plus the female dominated specialty of child psychiatry by using this model. There was a slight increase in probability of leadership with increased proportion of women, but the differences between the groups were not significant.
Discussion
Although the present study on careers of female physicians shows that few women reach senior positions, an encouraging finding, considering the steady increase in female medical graduates, is that the proportion of women in any medical specialty may be a determinant for female leadership. Nevertheless, there was no increased probability for leadership positions among the new generation of female physicians.
Because the medical career structure has been slow to adapt to the changing profile of medicine5 it has been questioned whether female physicians have the same opportunities as their male colleagues or if they are exposed to barriers which mean that they are likely to have a secondary role. 3–5 In addition to individual constraints such as family preferences and lack of leadership ambitions,6–8 attention has been drawn to structural constraints as the explanation of why female physicians who aspire to top positions meet invisible barriers. Deficiencies in career advice and counselling from patrons and colleagues may contribute to the “glass ceiling” effect on women's upward mobility.9–11 According to Kanter, who studied the impact of organisational characteristics on internal labour market patterns, the gradient of the existing sex ratio in a corporation is a strong predictor for career opportunities for women.12 She argues that women in a minority are exposed to exaggerated attention and lack of integration, reactions that will not appear in groups with less skewed sex ratios. Hence the size of the female minority group has to be sufficient to achieve the positions necessary to change the organisational structure.12 The fact that the relation between female leadership and the proportion of women in the specialty is positive supports this view if we are willing to equate the specialty union with Kanter's organisation.
Despite the recognition of dual careers, a policy of equal opportunity, and improvement of crèche facilities, the present study shows that women still do not reach senior positions as easily as men. The few female leaders in the study worked in public health and female supported hospital specialties. General practice and community health are deemed less prestigious than hospital career jobs, 3 13 and women report less interest in academic careers and leadership than men. 14 15 Even though the proportion of women in medicine is increasing, women tend to have employment characterised by routine work and low salary. 2 3 5 16 The progress towards parity remains slow and in certain sex stereotyped specialties, such as surgery and general medicine, women still comprise a tiny minority of the most senior grades. 2 3 16–18
Without disregard of discrimination issues, the underrepresentation of women in positions of authority may reflect some systematic, underlying effect whereby women simply do not choose to take on heavily committed senior positions. Because it is difficult to get away from the traditional linking of age to grade so prevalent in medicine5 women face a trade off between career posts associated with power and influence and family and emotional responsibilities. Although marriage and children are traditional constraints on the careers of female physicians, other factors, such as long hours and on call, impose considerable strain on relationships because careers of partners also have to be taken into account. 5 19 Even in dual career families, women are traditionally in charge of caring work.19 The present study suggests that female physicians give priority to emotional and family commitments at the expense of power and influence on medical decisions. A reassessment of the traditional linkage between career advantage and age is needed in the medical career structure. More part time opportunities and flexibility of careers would make it possible to give familial commitments precedence and subsequently make a preference shift towards influential medical positions.
Conclusion
Despite an equal opportunities policy with positive sex discrimination in favour of women and a liberal labour market encouraging female participation, Norwegian women still have difficulties in reaching medical career posts. Leadership jobs for women were associated with public health, where working hours are regulated. A promising finding is that the probability of women's achieving senior positions seems to increase with the proportion of women in the specialty. The results suggest that women face a trade off between family commitments and influence in medical leadership.
Acknowledgments
We thank professor Tore Schweder for helpful advice on the statistical analyses.
Contributors: GSB and KJK initiated the study and formulated the hypothesis; OGA made data available from the masterfile of the Norwegian Association. KJK and OGA were responsible for the data analysis. The paper was written jointly by OGA, GSB, and KJK. KJK is guarantor for the study.
Funding: University of Oslo.
Conflict of interest: None declared.