Trends in doctors' early career choices for general practice in the UK: longitudinal questionnaire surveys

Response

Overall, of those for whom deliverable addresses were available, the response rate 1 year after qualification was 57% (15103/26481 doctors); and 3 years after qualification it was 63% (8425/13 311). Response varied across cohorts in year 1 from 69% of qualifiers of 2000 to 47% of those of 2009; and, in year 3, from 71% of qualifiers of 2000 to 56% of those of 2005.

Untied first choices for general practice 1 year after graduation

An untied first choice is a first choice for general practice without the responder offering an equal first choice for another specialty. In the 2009 cohort, 20.4% of responders specified an untied first choice in 2010 for general practice (95% finite population confidence interval [CI] = 19.3% to 21.5%). Equivalent percentages for the cohorts of 2000, 2002, 2005, and 2008 were, respectively, 22.2% (95% CI = 21.3% to 23.1%), 20.2% (95% CI = 19.2% to 21.1%), 23.2% (95% CI = 22.3% to 24.1%), and 21.3% (95% CI = 20.3% to 22.3%). There was no trend over time (Table 1a; χ² test for linear trend: χ^2₁ = 1.2, P = 0.28).

Whole cohort estimates of the percentage choosing general practice, after adjusting for the higher percentage of women than men who responded to the surveys, were slightly lower at 21.6%, 19.5%, 22.6%, 20.9%, and 20.1%, respectively, for the cohorts of 2000, 2002, 2005, 2008, and 2009.

All first choices for general practice (whether or not shared with another specialty) 1 year after graduation

When those who chose general practice as their first choice jointly with another specialty (‘tied first choices’) were added (Table 1b), the percentages who chose general practice as a first choice, in the cohorts of 2000, 2002, 2005, 2008, and 2009 were, respectively, 29.6% (95% CI = 28.7% to 30.6%), 26.3% (95% CI = 25.3% to 27.3%), 33.0% (95% CI = 31.9% to 34.0%), 31.0% (95% CI = 29.9% to 32.1%), and 30.6% (95% CI = 29.3% to 31.8%). Whole cohort estimates were slightly lower (Table 1b). There was evidence of a modest upward linear trend in all first choices for general practice (χ^2₁ = 5.9, P = 0.015).

Any choice for general practice 1 year after graduation

Table 1c shows the percentages in each graduation year that chose general practice, whether as first, second or third choice. In the 2008 and 2009 cohorts, approximately half the responders chose general practice: among graduates of 2008 the percentage was 49.5% (95% CI = 48.3% to 50.8%) and among graduates of 2009 it was 49.9% (1457/2917, 95% CI = 48.6% to 51.3%). In the cohorts of 2000, 2002, and 2005 the corresponding percentages were 46.5% (95% CI = 45.4% to 47.5%), 43.4% (95% CI = 42.3% to 44.6%), and 52.6% (95% CI = 51.5% to 53.7%). Whole cohort estimates were slightly lower (Table 1c). The upward linear trend from 2000 to 2009 was significant (χ^2₁ = 19.0, P<0.001).

As Table 1 shows, the percentage of women who chose general practice was considerably, and consistently, higher than that of men.

Variation by clinical medical school attended

When cohorts were combined and responders grouped by medical school attended, large differences by medical school were evident in the level of choices for general practice 1 year after graduation (Table 2). The highest rates of choice for general practice were found in most of the newly established English medical schools, and the lowest rates among graduates of Oxford and Cambridge. Schools whose levels of choice for general practice were higher or lower than the overall average are highlighted in Table 2.

Variation by graduate entry

Untied first choices for general practice were made by 23.5% (382/1628) of graduate entrants in the 2005, 2008, and 2009 cohorts, compared with 19.5% (1407/7220) of non-graduate entrants (χ^2₁ = 12.8, P<0.001). The percentages of graduate entrants and non-graduate entrants making any first choice for general practice (that is, including choices tied with another choice) were 34.9% (568/1628) and 28.7% (2075/7220) respectively (χ^2₁ = 23.7, P<0.001); the percentages making any choice for general practice (whether first, second or third choice) were 52.0% (846/1628) and 47.6% (3438/7220) respectively (χ^2₁ = 9.9, P = 0.002).

Choices 3 years after graduation

The percentages of responders who chose general practice as an untied first choice 3 years after graduation (Table 3a) were 27.9%, 26.1%, and 35.1% in the cohorts of 2000, 2002, and 2005 respectively, a substantial increase between years 1 and 3 in each cohort. The percentages for all first choices, whether tied or untied, were also higher in year 3 than in year 1 (Table 3b). By contrast, considering all choices for general practice (whether first, second or third choices), year 3 figures were lower than in year 1 (Table 3c).

Changes of choice between years 1 and 3

Changes of choice were examined for doctors who responded in both years 1 and 3 (Table 4). Of those who chose general practice as an untied first choice in year 1, 82% did so again in year 3. Women were a little more likely than men to do so. Fifteen per cent of doctors who did not choose general practice in year 1 did so in year 3.

In all, 773 responders chose general practice in year 3 but not year 1, while 271 chose general practice in year 1 but not year 3. The percentage who chose general practice increased from 22.4% (1531/6820) in year 1 to 29.8% (2033/6820) in year 3 (McNemar's test, P<0.001). The net change represented 7.4% (502/6820) of responders, and meant that 38.0% (773/2033) of those who made a first choice for general practice in year 3 turned to it between years 1 and 3.

Women were a little more likely than men to switch to general practice between years 1 and 3.

Table 5 shows numbers who chose general practice as a first, second or third choice of specialty. For first, second or third choices combined, between years 1 and 3, there were net reductions, both for men and women, in the total numbers who chose general practice (all P<0.001). For example, in year 1 3315 doctors gave a choice — first, second or third — for general practice. In year 3, the number had fallen to 2941; 946 doctors who considered general practice in year 1 were no longer considering it, and 572 doctors were now considering general practice who had not considered it in year 1 (Table 5).

Summary

About one-fifth of UK graduates choose general practice as their sole first choice of career 1 year after graduation. This study's data show that this fraction has remained largely static across the past decade. If second and third choices are included, the figure rises to approximately a half in the most recent two cohorts of 2008 and 2009 studied in 2009 and 2010, in line, at first sight, with current UK policy targets. However, many doctors choosing general practice as a second or third choice in their first postgraduate year will eventually train in their first choice of another specialty rather than in general practice.

Data 3 years after graduation for the 2000, 2002, and 2005 cohorts give further insights. The cohort estimate of 34.3% of 2005 graduates who regard general practice as their sole first choice after 3 years is indicative of the percentage which may be expected to embark upon specialist training in general practice at that stage; many more than the 22.6% expressing the same intention in year 1. Some of the rise no doubt reflects increasing acceptance of the likely availability of training and career grade posts in general practice compared with opportunities in other specialties. For others, change of choice towards general practice no doubt follows favourable changes in views about the desirability of a career in general practice. In the same span of time, between years 1 and 3, the total percentage of the 2005 qualifying year who chose general practice fell to 43.9%, compared with 51.9% who had expressed an interest in general practice at year 1, indicating a softening of initial choices for general practice which were uncertain. At least a partial explanation for this decline will be that doctors, originally content to enter general practice but preferring another specialty, will in fact have been on track for a career in the other specialty.

There was continued evidence of substantial differences in the level of choices for general practice among graduates of different medical schools, in line with the authors' previous study of the graduates of 1999 and 2000.10 Doctors from new medical schools in England showed a higher level of choices for general practice than doctors from other schools.

Graduate entrants were slightly more likely than non-graduate entrants to choose careers in general practice, as in earlier cohorts.11 However, to date, doctors have not been asked whether they undertook a specific fast-track graduate entry course; many of these courses are recently established and it is perhaps too soon to tell to what extent they will affect career choices among their graduates.

Strengths and limitations

Strengths are that the study is large-scale (over 15 000 responders in year 1), includes all UK medical schools, covers five cohorts of newly qualified doctors, and its longitudinal design means that each doctor's replies can be compared in years 1 and 3. Its main weakness is that some level of non-response is inevitable and the possibility of responder bias must be considered. Non-response varied between cohorts; but, in year 1 first choices (that is, the firmest early decision), the percentage who chose general practice was stable. This suggests that trends in non-responder bias are unlikely. Choices for general practice of doctors who responded to the first two mailings were compared with those who responded to subsequent mailings. Late responders were used as a proxy for non-responders (because the latter would have been non-responders, if the study had been confined to two mailings). The percentages who chose general practice were very similar in the two groups.

Comparison with existing literature and implications for future research

Year 1 results for the cohorts of 2008 and 2009 were broadly similar to results in earlier cohorts. It seems likely that the 2008 and 2009 cohorts will therefore have similar aspirations at year 3 as those in earlier cohorts. This suggests that the 50% target, for recruiting to general practice, is a long way off the actual aspirations of doctors when they leave medical school.

The English Centre for Workforce Intelligence, reporting to the Department of Health on recommendations for medical specialty training for 2011, specifies that 2800 doctors were accepted for GP training in 2010 and recommends ‘… that moderate expansion is allowed from the existing GP training posts numbers for 2011 towards 3000 trainees’.12 This number would represent 46% of the registered UK medical graduates of 2009.

Data from the 2008 and 2009 cohorts provide some evidence that the higher level of preference for general practice expressed by women than men, in the first year after qualification, is falling. If the gender gap is truly in decline, it may not be possible to rely for much longer on the ‘demographic uplift’ of an increasing percentage of women among the graduates resulting in more choosing general practice.

Information about early career preferences matters because early choices for general practice are highly predictive of eventual careers. For example, in previous cohorts 82% of doctors who specified that they wanted to become GPs in year 1, and 92% who specified this in year 3, eventually practised as GPs.13

Extrapolating from the responder data to whole cohorts, and expressing the findings as numbers rather than percentages, the total number of young doctors choosing general practice 1 year after graduation, whether as first, second or third choice, has increased substantially since the year 2000. This is due partly to the expansion of medical schools, and partly to increased numbers of women in medicine, rather than any substantial change in the underlying level of choice for general practice by either sex.

The doctors in previous cohorts regard general practice highly.14,15 Nonetheless, a much smaller percentage of doctors leave medical school wanting a career in general practice than the NHS needs. In this context the high level of choices for general practice among the first graduates from the new English schools is interesting.

This mirrors experience in the US, where the American Academy of Family Physicians has recently reported16 just a very small increase in interest in primary care careers among trainees in 2010, citing the following factors ‘Multiple forces including student perspectives of the demands, rewards, and prestige of the specialty; national dialogue about health care reform; turbulence in the economic environment; lifestyle issues; the advice of deans; and the impact of faculty role models continue to influence medical student career choices.’ Its report16 continues ‘Despite matching the highest number of US seniors into family medicine residencies since 2004, in 2010 the production of family physicians remains insufficient to meet the current and anticipated need to support the nation's primary care infrastructure.’

In depth study of why this is so in the UK, and consideration of policy changes to encourage more doctors to enthusiastically embrace general practice as their first career choice, is required. Such study might consider the extent to which the level of choices for general practice in different schools reflects the different entry characteristics and aspirations of students, or whether it reflects the degree to which intentions are moulded by undergraduate experiences.