Geographical movement of doctors from education to training and eventual career post: UK cohort studies

Abstract

Objective

To investigate the geographical mobility of UK-trained doctors.

Design

Cohort studies conducted by postal questionnaires.

Setting

UK.

Participants

A total 31,353 UK-trained doctors in 11 cohorts defined by year of qualification, from 1974 to 2008.

Main outcome measures

Location of family home prior to medical school, location of medical school, region of first training post, region of first career post. Analysis for the UK divided into 17 standard geographical regions.

Results

The response rate was 81.2% (31,353/45,061; denominators, below, depended on how far the doctors’ careers had progressed). Of all respondents, 36% (11,381/31,353) attended a medical school in their home region and 48% (10,370/21,740) undertook specialty training in the same region as their medical school.

Of respondents who had reached the grade of consultant or principal in general practice in the UK, 34% (4169/12,119) settled in the same region as their home before entering medical school. Of those in the UK, 70% (7643/10,887) held their first career post in the same region as either their home before medical school, or their medical school or their location of training. For 18% (1938/10,887), all four locations – family home, medical school, place of training, place of first career post – were within the same region. A higher percentage of doctors from the more recent than from the older cohorts settled in the region of their family home.

Conclusion

Many doctors do not change geographical region in their successive career moves, and recent cohorts appear less inclined to do so.

Introduction

The geographical distribution of doctors is an important factor in the equitable distribution of health services. It is also an important outcome from medical education and training: medical schools and training programmes should have an interest in the geographical destinations of their graduates. Where doctors obtain posts is a compromise between the availability of posts and where they want to work. Little is known about how doctors move around for work and how they locate from home to medical school, from medical school to training posts and from training to career posts.

The UK Medical Careers Research Group (MCRG) has undertaken surveys of all medical qualifiers from all United Kingdom (UK) medical schools in selected year-of-qualification cohorts for many years, with follow-up at regular time intervals. We used data from these cohorts to study the geographical movement of doctors.

Method

Our questionnaires are mailed one, three and five years after qualification, at 10 years and at longer time intervals thereafter. We send up to five reminders to non-responders. We ask doctors about the location of their family home, clinical medical school and jobs after qualification. In this study we analysed a range of ‘pairwise’ comparisons. For example, we compared movement between location of home and medical school, medical school and specialty training, specialty training and career post, family home and career post and medical school and career post.

We limited the analysis to medical students who had a family home in the UK prior to beginning medical school (referred to as ‘UK-based’). The year-of-qualification cohorts we surveyed were those of 1974, 1977, 1983, 1988, 1993, 1996, 1999, 2000, 2002, 2005 and 2008, and they contribute to comparisons depending on the career stage they have reached as follows. All cohorts were used for the comparison of family home with medical school. Since the 2008 cohort had not yet progressed far enough, its doctors were excluded from comparisons of training location and for the same reason doctors belonging to the 2002, 2005 and 2008 cohorts were excluded from comparisons of career location.

Locations were coded initially by UK county and were re-grouped for analysis into UK regions and UK countries. To achieve continuity over the period covered by the study, the regions we used were those of the National Health Service (NHS) following the reorganization of 1974 into 14 English Regional Health Authorities (combining the four Thames regions, that included London, as one), four regions in Scotland, Wales and Northern Ireland. In the era of the NHS Regional Health Authorities, it was national policy that there must be at least one medical school and postgraduate specialty training programme per region. Accordingly, all regions described by us in this paper had medical students and specialty training programmes. For analysis the London medical schools were combined into one. For the main findings, we show movement at each career stage between regions and each of the four UK countries (England, Scotland, Wales and Northern Ireland). We then report, briefly, on doctors who received their specialty training and/or first career post outside the UK. The geographical groupings of regions and medical schools are summarized in The Appendix and further geographical detail is available from the authors.

As the structure of specialty training has altered over time,¹ so have grades and terminology. For ease of wording in the text, those in training posts for general practice are described as general practitioner (GP) trainees; those in hospital training posts as hospital trainees. Full details of grades regarded as training or career posts are available from the authors (Supplementary data 1). In brief, the ‘first training post’ for the purpose of the analysis was defined as the training post following, but not including, the house officer year (previously known as the preregistration house officer) for the early cohorts, or the F2 Foundation year for the recent cohorts. The ‘first career post’ was the first post as a hospital consultant, hospital associate specialist, hospital staff grade doctor, public health consultant, GP principal or salaried GP (all including academic appointments with honorary NHS contracts at equivalent level).

All statistical analyses were performed in SPSS version 15.0 for Windows. Chi-square tests were used to compare results between men and women, and between those with career posts in hospital medicine and general practice. Linear-by-linear association was used to assess trends over time.

Direction of analysis

Percentages have been calculated, for many of the comparisons, in two directions: percentages looking forward from region at the early stage of career to region of later destination, and percentages looking backwards from the destination to the region at earlier career stages.

Further information from the authors

This study is based on a very large data-set, which can be analysed with many permutations and the paper is necessarily a summary. Further tables detailing the numbers on which percentages are based, additional breakdowns of the results, and figures detailing the results for each region, are available from the authors (Supplementary data 2).

Results

Response

A total of 45,616 doctors were surveyed, generally at one, three and five years after qualification and at various intervals afterwards. We excluded from the overall total 25 who had never registered to practise, 360 who had declined to participate and 170 known to be deceased by the time of the most recent survey of each cohort, leaving an effective total of 45,061. Of these, 36,585 (81.2%) had responded to at least one survey of their cohort. Family home location was provided by 33,425 respondents and of these, 31,353 (94.6%) were UK-based. This group, representing 85.7% of all respondents (31,353/36,585), formed the basis for the analysis.

Information on training posts was available for 22,115 of the UK-based doctors of the 1974–2005 cohorts, and information on career posts was available for 12,224 of the UK-based doctors of the 1974–2000 cohorts. We compared locations by UK region and separately by UK country (England, Scotland, Wales and Northern Ireland). Comparison was on the basis of paired locations. The number of doctors in each of the pairs, e.g. location of family home and location of eventual career post, depends both on response rates and on how far doctors in each cohort have progressed in their careers; see totals in Table 1.

Table 1.

Percentage of doctors who were in the same region in various combinations of their family home, medical school, training post and first career post: UK-based doctors by year of qualification

Year(s) of qualification	Medical school in region of family home1	Training in region of medical school2	Career in region of training4	Training in region of family home2	Career in region of medical school3	Career in region of family home3	Career in region of training, med school & family home5	Career in region of at least one previous location5
1974	40.5	50.7	50.6	36.2	35.6	34.0	16.6	61.5
1977	43.2	51.6	53.5	36.7	39.2	32.6	19.2	64.1
1983	40.8	48.2	52.8	35.3	38.0	33.3	17.7	66.0
1988	36.1	47.4	58.1	32.2	38.5	32.5	16.8	70.0
1993	33.6	44.1	66.3	31.8	40.7	35.0	17.6	77.1
1996	33.2	41.7	70.4	31.9	39.4	35.0	15.9	79.7
1999	32.6	45.5	74.6	31.2	44.5	39.4	19.5	84.4
2000	33.5	45.7	78.3	34.4	49.5	43.9	23.1	86.6
2002	33.9	48.5		34.1
2005	36.5	55.9		39.2
2008	38.0
1974–2008 Total	36.3	47.7	59.1	34.1	39.4	34.4	17.8	70.2
Men	37.6	47.0	54.7	34.6	37.3	33.0	16.7	66.4
Women	35.1	48.4	65	33.7	42.2	36.3	19.3	75.3
Total N	31,353	21,740	10,887	21,740	12,119	12,119	10,887	10,887
Linear trend (χ21)	70.8	4.7	270.7	2.3	26.2	16.9	0.4	273.4
P value	<0.001	0.03	<0.001	0.1	<0.001	<0.001	0.5	<0.001

Note Pearson χ² P values for all columns are <0.001. Except for career in region of training, med school and family home P value = 0.01

*See following notes for cohorts combined in each comparison

¹1974–2008 cohorts combined, locations of family home and medical school known

²1974–2005 cohorts combined, training job present, locations of training and family home known

³1974–2000 cohorts combined, career post present, both locations for comparison known

⁴1974–2000 cohorts combined, career post and training post present, locations of career and training known. 274 doctors with training post abroad are excluded

⁵1974–2000 cohorts combined, all four locations known

Comparisons of locations (Table 1)

Table 1 shows locations along the respondents’ career stages. Overall, 36% of respondents went to medical school in the same region as their family home; 48% undertook specialty training in the region of their medical school; 59% obtained their first career post in the region in which they undertook specialty training, 34% undertook specialty training in the region of their family home, 39% obtained their first career post in the region of their medical school, and 34% obtained their first career post in the region of their family home.

For 18% of all doctors, all stages of their student life and medical career – home before medical school, medical school, training and career post – were in the same region. For 70%, their first career post was in the same region as at least one of their previous three stages.

There were no major trends over time in the percentages of doctors who obtained their first training post in the region of their family home; or in those who obtained their first training post in the same region as their medical school.

The percentage who obtained their first career post in the region of their training rose from 51% of the qualifiers of 1974 to 78% of the qualifiers of 2000; the percentage who obtained their first career post in the region of their medical school rose from 36% of the qualifiers of 1974 to 50% of the qualifiers of 2000; and the percentage who obtained their first career post in the region of their family home rose from 34% of the qualifiers of 1974 to 44% of the qualifiers of 2000. The percentage of doctors who obtained their first career post in a region which had been at least one of their previous locations increased, up from 62% of the qualifiers of 1974 to 87% of the qualifiers of 2000 (respectively, 66% of women and 60% of men, up to 89% of women and 82% of men). A higher percentage of women than men held a career post in their region of training (women 66%, men 55%, Table 1). Women were also more likely than men to hold a career post in a region that had been at least one of their previous locations (75% women, 66% men, Table 1).

Hospital doctors and GPs

The regions of career and regions of training were compared separately for hospital doctors and GPs. Overall 49.6% (2304/4643) of hospital trainees and 67.9% (4051/5970) of GP trainees obtained first career posts in the same region as their training (χ²₁ = 361.4, P < 0.001). Women were significantly more likely to do so than men, regardless of whether they were GPs or hospital doctors. For men and women GPs, the respective percentages were 70.1% (2117/3019) and 65.5% (1934/2951, χ²₁ = 14.4, P < 0.001). For men and women hospital doctors, the respective percentages were 57.8% (842/1458) and 45.9% (1462/3185, χ²₁ = 56.2, P < 0.001).

Comparison of countries within the UK

The direction of the analyses in Table 2 shows the progression of doctors, by UK country, from the location of their family home through to medical school, from medical school to first training post and from training post to first substantive career post. English medical students and doctors tended to stay within England. For example, 91% of medical students whose homes were in England went to medical school in England and 90% obtained their first career post in England. By contrast, 64% of medical students whose homes were in Northern Ireland went to medical school in Northern Ireland and 50% obtained their first medical career post in Northern Ireland. The corresponding percentages for Wales were 39% and 45% and for Scotland they were 88% and 63%.

Table 2.

Percentage of doctors who were in the same country, looking forward from region of family home to medical school to training post to first career post: combined cohorts of 1974–2008*

Country at previous status†	Med school in country of family home1	Training in country of family home2	Training in country of medical school2	Career in country of family home3	Career in country of medical school3	Career in country of training4	Career in country of training, medical school & family home5
England	90.8	89.0	90.9	89.7	90.4	93.2	83.8
Scotland	88.2	68.0	61.6	63.4	56.4	73.7	58.3
Wales	38.5	38.0	52.4	45.0	49.3	66.5	23.8
N Ireland	64.3	54.3	77.9	50.3	68.6	80.4	41.4
Total	86.5	82.2	84.1	82.1	82.4	88.9	75.5
Total N	31,353	21,740	21,740	12,119	12,119	10,613	10,613

Note Pearson χ² P values for all columns are <0.001

*Cohorts combined as follows:

¹1974–2008 cohorts combined, locations of family home and medical school known

²1974–2005 cohorts combined, training job present, locations of training and family home known

³1974–2000 cohorts combined, career post present, both locations for comparison known

⁴1974–2000 cohorts combined, career post and training post present, locations of career and training known

⁵1974–2000 cohorts combined, all four locations known

^†Previous status in each comparison is the earlier of the two locations; e.g. family home is the origin in Column 2, and medical school in Column 4

The direction of analysis in Table 3 starts with the later ‘destination’ and looks back to where the student or doctor came from. It shows, for example, that 93% of respondents who went to medical school in England had their original family home in England; and 89% of those who got their first career post in England came from an English medical school. The corresponding values for Scotland were 64% and 73%, for Wales were 44% and 49% and for Northern Ireland were 97% and 93%.

Table 3.

Percentage of doctors who were in the same country, looking backwards from career post to training post, training post to medical school, medical school to home: combined cohorts of 1974–2008*

Country of destination†	Family home in country of medical school1	Family home in country of training2	Medical school in country of training2	Family home in country of career3	Medical school in country of career3	Training in country of career4	Training, medical school & family homein country of career5	Career in same country as at least one previous location5
England	93.3	90.8	90.5	89.5	89.1	92.4	81.3	97.2
Scotland	64.2	68.7	83.7	72.8	82.8	85.7	65.2	93.1
Wales	43.7	43.8	52.9	49.2	42.9	65.9	24.6	76.2
N Ireland	97.0	93.1	88.4	92.9	88.1	85.3	76.6	95.8
Total	86.5	86.0	88.0	85.7	86.0	90.1	76.4	95.6
Total N	31,353	20,774	20,774	11,614	11,614	10,483	10,483	10,483

Note Pearson χ² P values for all columns are <0.001

*Cohorts combined as follows:

¹1974–2008 cohorts combined, locations of family home and medical school known

²1974–2005 cohorts combined, training job, locations of training and family home all known

³1974–2000 cohorts combined, career post present, both locations for comparison known

⁴1974–2000 cohorts combined, career post and training post present, locations of career and training known

⁵1974–2000 cohorts combined, all four locations known

^†Destination in each comparison is the later of the two locations; e.g. medical school is the destination in Column 2, and training location in Column 4

Comparison of regions within the UK

Table 4 summarizes the maximum and minimum percentage range for doctors whose successive career stages were in the same region. As expected, the variations were large. For example, looking forward from earlier to later, 64% of medical students whose homes were in Northern Ireland went to medical school in Northern Ireland, while only 5% of students from family homes in the Oxford region went to medical school in Oxford. Looking backward, 97% of medical students who graduated from Belfast medical school in Northern Ireland came from family homes in Northern Ireland; while only 8% of Cambridge graduates came from family homes in East Anglia.

Table 4.

Percentage of doctors whose successive moves were in the same region

Looking forward: Percentage of doctors for whom ‘paired locations’ were in the same region, showing maxima and minima
Comparison of locations	Minimum (%)	Maximum (%)	Average (%)
Medical school in family home region	5.5	64.3	36.3
Region	Oxford	N. Ireland
Specialty training in medical school region	20.3	77.9	47.7
Medical school	Cambridge	Belfast
First career post in training post region	34.1	80.4	59.9
Region	E. Scotland	N. Ireland
First career post in family home region	14.6	50.3	34.4
Region	Oxford	N. Ireland
First career post in medical school region	12.3	68.6	39.4
Medical school	Oxford	Belfast

Looking backward: Percentage of doctors for whom ‘paired locations’ were in the same region, showing maxima and minima
Comparison of locations	Minimum (%)	Maximum (%)	Average (%)
Family home in medical school region	7.8	97.0	36.3
Medical school	Cambridge	Belfast
Medical school in region of training post	14.0	88.4	50.0
Region	E. Anglia	N. Ireland
Training post in region of first career post	37.8	85.3	60.6
Region	N. Ireland	E. Scotland
Family home in first career post region	14.8	92.9	35.9
Region	E. Anglia	N. Ireland
Medical school in first career post region	6.7	88.1	41.0
Region	Oxford	N. Ireland

Specialty training and career posts outside the UK (Table 5)

Table 5.

Percentage of doctors from medical school in each UK country whose location of training and of first career post were outside the UK

Country of medical school	Location of training			Location of first career post
	Outside the UK		Total	Outside the UK		Total
	Percentage	Number		Percentage	Number
England	4.3	715	16,531	4.0	366	9240
Scotland	5.3	190	3559	5.0	98	1971
Wales	4.3	39	909	3.6	18	507
Northern Ireland	3.0	22	741	5.7	23	401
Total	4.4	966	21,740	4.2	505	12,119

In all, 4.4% of the UK-based doctors who had progressed as far as a specialty training post did their specialty training outside the UK. The percentage variation by country was borderline significant (England 4.3%, Scotland 5.3%, Wales 4.3%, Northern Ireland 3.0%; χ²₃ = 11.1, P = 0.01). The range among the English regions was from 3.1% (Mersey) to 7.0% (Oxford).

Among those who had attained a career grade post 4.2% had done so outside the UK. The percentage did not vary significantly by country (England 4.0%, Scotland 5.0%, Wales 3.6%, Northern Ireland 5.7%; χ²₃ = 7.1, P = 0.07). The range among the English regions was from 2.5% (West Midlands) to 8.4% (Wessex).

Discussion

Principal findings and interpretation

Medicine is, in principle, a profession that permits a high level of geographical mobility: the need for doctors is universal. In practice, many doctors stay reasonably local in their successive career moves. Fifteen years ago we reported similar data on the mobility of doctors in the graduation cohorts who graduated between 1974 and 1993.² The data since 1997 to date, reported here, indicate that relationships between location of career post and training post, career post and medical school and career post and original family home have strengthened in recent UK cohorts. This may reflect increasing moves to structure specialist training programmes in non-teaching hospitals with training relationships with their local medical school. We do not have figures from our studies for the numbers of doctors who move between the end of medical school and the first year of postgraduate training (now termed the F1 year). However, it is known from other research that it is uncommon for doctors to move during their two year Foundation training. For example, the latest (2011) annual report of the UK Foundation Programme³ reports that 91% of F2 doctors in August 2011 started the second year of their Foundation Programme in the same foundation school as in the first year. A study by the British Medical Association followed a subgroup of 431 doctors from the 2006 cohort of UK medical graduates and found that 55% of the 339 respondents who were in UK specialty training remained in the same geographic region as their medical school,⁴ a figure which is very similar to our figure of 56% for the graduates of 2005.

The increase in percentages of doctors who stay local might also reflect shorter periods of training such that doctors are less inclined to move to career posts afar from training posts. However, one of the most striking characteristics in the trends was the increased likelihood that doctors from more recent than older cohorts settled, for their first career post, in the broad location of their original family home.

The predominance of dual income households means that considerations of job location need to take account of the careers of both partners more than was the case in the past. On balance, this may reduce the ease with which doctors can move between distant locations; and increase the tendency for doctors, and their partners, to co-locate more locally. We considered the possibility that the greater tendency in recent years of doctors to stay close to previous locations might have been a function of the increased intake of women into medicine; but it was not. As shown above, the increase in the percentage of men who stayed local was similar to that of women. Recent research suggests that career expectations and practice patterns of younger doctors differ from those of older generations.⁵ Younger generations are more likely to take into account the preferences of their spouses than older generations. Greater emphasis in recent years on ‘work-life balance’ may have caused more doctors to stay close to parental family.

Strengths and limitations

We present findings from a unique data-set, comprising multiple cohorts of medical graduates and multiple surveys of each cohort, over 35 years. The study is restricted to UK-trained medical graduates. Non-responder bias is possible, as with all self-completed surveys; for example, doctors who leave the UK may be more difficult to reach. We did not consider doctors’ motivations for selecting their location. We do not know whether doctors who settled close to, or far from, their previous location were currently in a preferred location, an acceptable one or one they regretted.

There are circumstances that are not easily factored into comparisons between regions in the movement of doctors. The geographical proximity of regions and other factors such as the quality of transport links between them suggest that it may be more likely for a doctor to move from, say, the North East to the North West than to the South West. As expected, there was much variation between the medical schools both in attracting prospective students from outside their local region and in retaining them in the same region after graduation. For example, Oxford and Cambridge have a very wide demographic spread – national and international – in their student intake. Northern Ireland, with Belfast as the only medical school, is the most self-sufficient part of the UK in terms of attracting local students to its medical school and subsequently supplying its own doctors.

Implications & conclusions

Migration of doctors, and the equitable distribution of health workers, is a major issue in the equitable distribution of healthcare. Much of the literature on medical migration focuses on international rather than internal migration.⁶ No studies were found reporting comprehensive information about doctors’ mobility between their family home, medical school, place of training and eventual career appointment. On a more limited basis, over the years there have been a number of policy initiatives to address shortages of doctors in particular geographical areas within countries, notably in remote areas. For example, in Missouri an initiative running since the mid-1990s called the Rural Track Pipeline Programme (MU-RTPP) increased the supply and retention of rural physicians state-wide.⁷ A similar project in New Brunswick, Canada⁸ has operated since 1981 and it was found that exposure to a different geographical area during postgraduate training, but not during undergraduate training, increased the likelihood that doctors would be recruited and retained in the province. A comparison of the distribution of primary care physicians in Great Britain and Japan⁹ found that the more equitable distribution in Britain was associated with the more controlled and incentivized allocation of posts within the NHS, compared with the Japanese system. However, a recent UK study¹⁰ concluded that equity of distribution of general practitioners in England has fallen since 2002 and that new ‘targeted area level policies’ may be needed.

In summary, a substantial percentage of UK doctors stay reasonably close to their previous location, at least at the level of the region. Doctors in younger generations are more likely than older generations to remain close to their previous location. Reduced mobility may not be sustainable: doctors have to go where the jobs are.

DECLARATIONS

Competing interests

All authors declare no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work; no other relationships or activities that could appear to have influenced the submitted work

Funding

This is an independent study commissioned and funded as part of a wider work programme by the Policy Research Programme in the Department of Health. The views expressed are notnecessarily those of the funding body

Ethical approval

This study was approved by the UK National Research Ethics Service (ref 04/Q1907/48)

Guarantors

MG and TL

Contributorship

MG and TL designed and implemented the cohort studies. JD undertook the primary analysis and interpretation of the data and wrote the first draft of the paper. JM undertook further analysis with assistance from TL. All authors contributed to the intellectual content of the paper and to further drafts of the manuscript, and had access to all of the data. All authors approve the final version

Acknowledgements

We would like to thank Emma Ayres who administered the surveys, Janet Justice and Alison Stockford for their careful data entry, and all the doctors who participated in this survey

Appendix

Regions and medical schools used in the analysis

The United Kingdom comprises England, Scotland, Wales and Northern Ireland.

The regions used in the analysis, with the names of their medical schools in parentheses, were as follows.

England comprised 11 regions: London/Thames regions (London schools, Brighton and Sussex Medical School); East Anglia (Cambridge and East Anglia); Mersey (Liverpool); North Western (Manchester); Northern (Newcastle); Oxford (Oxford); South Western (Bristol, Peninsula); Trent (Leicester, Nottingham, Sheffield); West Midlands (Birmingham, Keele, Warwick); Wessex (Southampton); and Yorkshire (Leeds, Hull York).

Scotland comprised four regions: East Scotland (Dundee); North/North East Scotland (Aberdeen); South East Scotland (Edinburgh); West Scotland (Glasgow).

Wales was treated as one region (Cardiff Medical School).

Northern Ireland was treated as one region (Belfast Medical School).