International nurse recruitment and NHS vacancies: a cross-sectional analysis

Background Foreign-trained nurse recruits exceeded the number of new British-trained recruits on the UK nurse register for the first time in 2001. As the nursing shortage continues, health care service providers rely increasingly on overseas nurses to fill the void. Which areas benefit the most? And where would the NHS be without them? Methods Using cross-sectional data from the 2004 Nursing and Midwifery Council register, nurse resident postcodes are mapped to Strategic Health Authorities to see where foreign recruits locate and how they affect nurse shortages throughout the UK. Results Areas with the highest vacancy rates also have the highest representation of foreign recruits, with 24% of foreign-trained nurses in the UK residing in the London area and another 16% in the SouthEast (comparable numbers for British-trained nurses are 11% and 13%, respectively). Without foreign recruitment, vacancy rates could be up to five times higher (three times higher if only Filipino recruits remained). Conclusion The UK heavily relies on foreign recruitment to fill vacancies, without which the staffing crisis would be far worse, particularly in high vacancy areas.


Background
The National Health Service (NHS) has been suffering the effects of a nursing shortage for the past decade as fewer women train or remain in the nurse workforce, favoring improved job market opportunities in other sectors. The same is true of many other industrialised nations. Over the next 5-10 years, the nurse shortfall is predicted to be 275,000 in the US; 53,000 in the UK and 40,000 in Australia by 2010 [1]. By 2020, the US shortfall may be as high as 800,000 [2]. Efforts to address the shortage include return-to-work initiatives, improved pay, better working environment and flexible hours, and attracting more students to nurse training programs. Even so, the negative aspects of a nursing career discourage many people from training or remaining in the nurse workforce. Perceptions of the NHS as a poor employer are particularly acute and wages remain below other professions, even other jobs within the public sector [3,4].
In this age of globalisation, many countries have turned to overseas recruitment to fill the vacancies caused by a limited or unwilling locally trained workforce. Foreigntrained nurses accounted for 23% of the nurse workforce in New Zealand in 2002; 6% in Canada (2001); 8% in Ireland (2002) and the UK (2001); and 4% in the US in 2000 [1]. And these numbers are likely to grow if domestic hiring continues to lag. In fact, in 2001, more than half of newly registered nurses were trained overseas [5]. Even doctors and pharmacists are being recruited from overseas [6,7]. But foreign recruitment comes at a price, particularly for the source countries.
The broadsheets are increasingly reporting on the 'global nursing crisis,' the 'healthcare brain drain' from developing countries and the effects the 'nurse exodus' has on quality of care in poor countries. This is especially true of African nations which not only subsidise countries like England (by investing in the training of healthcare professionals who move to industrialised nations), but suffer from growing nursing shortages of their own, exacerbating problems of existing staff shortages, high infant and maternal mortality rates, the HIV epidemic, poor nutrition and a host of other public health concerns [8-10]. As one of the world's largest importers of healthcare professionals from developing countries, the UK is in need of adopting and enforcing standards of ethical conduct in recruitment.
In 1999, the Department of Health issued guidelines on international recruitment intended to curtail poaching from poor countries already suffering from healthcare staffing shortages, such as South Africa and the Caribbean. Unless foreigners provide unsolicited applications or their governments have established programmes for professional development with the UK, the NHS was advised to avoid recruiting [11]. The guidelines were updated in 2001 to include recruitment agencies working for the NHS, but the guidelines do not apply to the independent sector and are not monitored or enforced [12]. Still, the recommendations help explain the strong presence of Filipino nurses who are explicitly encouraged to train overseas (and send income home) as part of the 2001-2004 Medium Term Philippines Development plan. Even the Philippines, however, is starting to feel the crunch of a nursing shortage and trying to find ways of ensuring that sufficient numbers of nurses, particularly nurse educators, are trained and retained in the future [1,13]. As conditions worsen in source countries, many believe that it is neither realistic nor ethical for developed nations to continue relying on foreign recruitment from disadvantaged nations [8,12,[14][15][16].
If overseas recruitment becomes more difficult, it will have a detrimental effect on vacancy rates in the UK in future, at least in the short-term. (In the long run, perhaps staff shortages even more extreme than today could provide the impetus for major changes that would attract British-trained nurses into the profession. The changing wage structure under Agenda for Change may also help, but it is far too soon to know.) In order to understand that future effect, it helps to study the current and historic impact of overseas recruitment on the NHS. Using data from the Nursing and Midwifery Council (NMC) register from April, 2004, this paper identifies where nurses reside, by country of origin, to see how foreign recruitment affects different Strategic Health Authorities (SHAs) across England. Hypothetical vacancy rates for 2004 are estimated (based on a worst case scenario) as if the UK had no overseas recruits to determine what the shortage would look like across the country if the UK could rely only on its own stock of nurses. Though hypothetical, these numbers help shed light on the extent to which the British workforce has been unwilling to enter or remain in the nursing profession, leading to reliance on foreign sources of labour.

Methods & Data
The Nursing and Midwifery Council (NMC) maintains the register of qualified nurses, midwives and health visitors for the UK. Because any qualified nurse wishing to work in the UK must register with the NMC, it represents the entire pool of potential nurses that could be recruited into the NHS (overseas residents or retired nurses still registered notwithstanding). The NMC register is updated on a daily basis, containing over 600,000 records. An extract of data from mid-April, 2004, was provided by the NMC, containing counts of all registered nurses by 5-digit postcode by country of qualificiation (where known). While there is certainly error in assigning postcode data to areas within England, the method produces aggregate numbers comparable to the NMC's own (unpublished) analysis by reported country of residence. The NMC estimated that, as of the end of March, 2004, roughly 632,000 (or 96%) of the 660,215 registrants resided in the UK, of whom roughly 509,000 were in England, 64,000 in Scotland, 32,000 in Wales and 22,000 in Northern Ireland (with an additional 4,000 in the UK unspecified). These are roughly comparable with the counts from April, 2004, presented here (Table 1). They found almost 29,000 registrants either reside overseas or did not provide resident country or postcode, which is comparable to the estimated 31,000 unknown or foreign addresses found here and to previous estimates that roughly 5% of the register reside overseas [17]. These numbers are similar to a report on the 2002-2003 data [18]. Because the register is updated on a daily basis, published numbers will not match the current analysis exactly.
To verify country of training, the NMC data was compared with a publication on London-based nurses. A 2002 analysis by Buchan, Finlayson and Gough found that roughly 12% of nurses reporting a London postcode were from overseas, similar to the 10% found here [19]. The lower count may be attributable to one of several reasons. First, a small number of registrants may have been misassigned to a country, or allocated to 'unknown postcode,' due to partial or inaccurate postcode reporting (which is the only geographic identifier available in this analysis). These cases would not have been assigned to a London Strategic Health Authority. Furthermore, newly registered foreign recruits may list their recruitment agency address initially, introducing error in SHA assignment for newly registered immigrants (though this should disappear upon reregistration). Finally, in previous work done by Buchan ana-lyzing the NMC data, all foreign recruits appeared to enter the UK after 1995, despite known foreign recruitment before then. The loss of data on country of training may be attributable to a change in computer systems used by the NMC to track nurse registrations. But whatever the explanation, it appears the NMC data undercount foreigntrained nurses, excluding those registering prior to 1996 and perhaps undercounting since then as well. So the current analysis focuses on recent recruitment, subject to the limitations of the NMC data itself. (Table 2 shows that of the 67,176 new non-UK admissions to the NMC register since 1993, only 7335, or 11%, registered prior to 1996. Some of them may not have entered the UK or could have left by now. But to the extent they remain in the UK, the majority of foreign recruits will have been captured and properly attributed to their source country in the 2004 data.) Postcode data was mapped to SHA code using the XYZ Digital Map Company's PostZon file, mapping 7-digit postcodes to SHAs (based on data supplied by Royal Mail) [20]. For NMC postcodes with multiple SHAs (particularly where only the first two postcode digits were reported in the register), nurse counts were split across SHAs according to the proportion of 7-digit postcodes within the NMC postcode assigned to an SHA. 2. There were no foreign-trained nurses working in the UK, and; 3. Wages and job characteristics would be no different today without foreign recruits than they are with (notwithstanding the added stress of higher vacancies, which would probably further increase vacancies).
Obviously these assumptions are severe and improbable, but they are needed to present a worst case scenario and to calculate the upper boundary of vacancies in the absence of foreign recruitment. Table 1 shows staffing levels (wte and headcounts), NMC register counts, proportion of the register working in the NHS (headcount divided by register count), and vacancy data by SHA in 2004, for all qualified nurses, midwives and health visitors. The register is based on postcode of residence, rather than postcode of work, so the estimated 'proportion working in NHS' will be biased upwards by commuters coming to work in an SHA (the effect of which may be muted somewhat if resident nurses work in the private sector, commute out of the SHA or do not work, in comparable numbers to the incoming commuters). For example, in North Central London, obviously many   Obviously, vacancy rates are highest in London, which also sees a higher proportion of the registered nurse population working in the NHS (though this statistic is biased upwards by commuters from outside the London area). But these aggregate numbers mask the problem of encouraging British citizens to train and work in nursing, as vacancies are increasingly being filled by foreign-trained recruits. The next step is to determine the extent to which shortage areas rely on overseas recruitment. Using the NMC mid-April data, broken down by country of training by postcode, an SHA of residence is assigned based on the reported postcode sector (the 5-digit postcode, assuming a valid or partial postcode is provided and is within the UK). Table 4 presents data on the NMC register, broken down by SHA of resident postcode, by training source (foreign or UK). Scotland has the lowest number of registered foreign recruits as a fraction of all registered nurses with only 1.3%, while England has the highest at 5.5%. SHAs outside of London and the Southeast experienced average foreign representation of only 4.3% (of all registered nurses within the SHA). SHAs with high fractions of foreign recruits (5.5% or more of all registered nurses), are either in London (over 10%) and the Southeast (6.2%) or contain some of the largest cities in England (namely Manchester, Birmingham and Bristol). Leeds (in the West Yorkshire SHA) has a slightly lower foreign presence (5%) with only Liverpool and Sheffield (the Cheshire&Merseyside and South Yorkshire SHAs) among Britain's seven largest cities with very low shares of foreign-trained nurses (less than 4.5% of their potential workforce). Unsurprisingly, the SHAs with higher proportions of overseas recruits are also the ones with the highest vacancy rates (Table 1). Apparently, the worse the nursing shortage, the more active the foreign recruitment efforts (assuming high foreign representation does not drive vacancies, but vacancies drive foreign recruitment).

Results & Discussion
Of the roughly 30,000 foreign-trained nurses residing in the UK, 24% are in the London area and another 16% in the SouthEast (with 49% in the rest of England and the remaining 11% divided among Wales, Scotland and Northern Ireland). Comparable numbers for Britishtrained registrants living in the UK are 11% in London, 13% in the Southeast and 56% in the rest of England.
While there are problems with using NMC registration address to assign the SHA of work, the data can roughly determine the distribution of nurses across England (with some error for commuters and misassigned postcodes), and provide reasonable estimates of the pool of potential nurses located within SHA boundaries.
Clearly a disproportionate share of foreign-trained nurses live (and probably work) in the London area, helping to lower London vacancy rates beyond what they would be were international recruitment not possible. Table 5 presents upper bound estimates for hypothetical vacancy rates in the absence of international recruitment. Without any foreign recruits, vacancy rates could be as high as 12% for England, but higher for individual SHAs, particularly in London where aggregate vacancies could rise above 20%. These vacancy rates are three to five times greater than current estimates. Even allowing for Filipino recruitment, vacancies would still be two to three times greater. Of course, given the assumption that all foreign-trained recruits work full-time for the NHS (rather than the independent sector that probably recruited them, for example), these numbers are an upperbound, with more reliable predicted vacancies lying somewhere between current rates and those in

Conclusion
There is a growing literature concerning nurse labour shortages and foreign recruitment. This paper identifies where foreign recruits move in the UK and estimates that vacancy rates could be three to five times higher without such a strong foreign-trained presence. This is especially true of the highest vacancy areas, like London and the SouthEast, which could otherwise have double digit vacancy rates (excluding all international recruitment other than from the Philippines). An estimated 100,000 nurses on the NMC register were over the age of 54 and less than 12% (fewer than 66,000) are under the age of 30 [17]. This lopsided age distribution will cause further problems as the 55+ cohort retires over the next ten years. Without overseas recruits to rely on, it is not clear the domestic market can supply the necessary staff. So without major changes, the UK's reliance on foreign-trained nurses will continue (in the forseeable future), and nursing is likely to remain a safe career choice for foreigners hoping to emigrate to the UK.
Reliance on foreign recruitment (not just in the UK, but the USA and other industrialised nations) poses two important questions for policymakers and researchers. The first of course is how to mitigate the impact of nurse emigration on source countries. Developing countries cannot hope to compete with the higher salaries and better working conditions offered by the UK or other developed economies, leaving their health services (especially in rural areas) with labour shortages of their own [24]. For example, an estimated two-thirds of the Jamaican nurse population has emigrated, leaving Jamaica to fill the void from Cuba [25]. Restricting emigration or taxing leavers (as was common in the 1970s) have a variety of problems and will not eliminate individual workers' desire to leave [25]. Any long run solution should involve strategies to encourage workers to stay, through better working conditions, improved wages, or other positive inducements. These efforts could have the added benefit of encouraging even more people to train as healthcare providers in developing countries. To better understand and address these problems, further research is needed to quantify the effect of nurse migration on developing countries; estimate what fraction of the nurse workforce emigrates from each country; determine what fraction of emigrating nurses remain permanently overseas or return home (and in what timeframe); and study the effect of various policy options to encourage more nurses to stay in their home countries. So far, data from source countries is limited, and even industrialised nations have little information tracking skilled workforce migrants [24].
The second question to address is how to encourage more people in industrialised nations to train as nurses. Just as developing countries have difficulty competing with developed economies in the nurse labour market, the NHS cannot effectively compete with other employers (in healthcare or other sectors) in the domestic labour mar-ket. There are several reasons for the declining number of nurses in the UK, including working conditions, low wages, the cost of living, the changing nature of the job, feeling valued and of course, outside employment opportunities. The labour supply elasticity literature from the UK suggests nurses are relatively unresponsive to wage increases, with a 10% increase in wages leading to an estimated 4% increase in hours worked or 6% increase in the probability of working [27,28]. However, survey data suggests pay does drive decisions (or intentions) to quit. And a comparison of nurses' wages with those of other nonmanual female workers in the UK suggests nurse wages increased in real terms over the past 20 years, but fell relative to other workers (presumably making other careers more attractive by comparison).
Furthermore, geographic variation in vacancy rates across the UK is partially driven by variation in housing costs. This also suggests low wages may be the culprit, since the relatively flat pay structure in the NHS does not adequately adjust wages in high cost areas (there is an adjustment, but it is small compared with London housing costs, for example), driving nurses away [29]. Another reason behind the nursing shortage in England is poor labour force planning in the early 1990s, during which the number of training posts was intentionally reduced. Since 1994, training positions have increased with the Department of Health and now, Strategic Health Authorities, setting the number of training positions needed. However, by failing to take into account demographic trends and increased competition from the private sector (in terms of rising demand from an ageing patient population, an ageing nurse workforce which will need to be replenished, and growth in private sector employment opportunities for nurses), the targets continued to fall short of demand, at least in the 1990s [30]. Ideally, future demand will be met from the domestic labour supply, but this requires better educational planning and improved working conditions and pay to train and retain appropriate numbers of nurses.
Additional research is also needed on the labour supply of nurses in industrialised nations. The UK has a good starting point as the Nursing and Midwifery Council registry provides an invaluable resource containing geographic and basic training data for all nurses. With some effort, it could be used to track migration patterns of nurses within the UK as nurses update their information every few years (when they re-register). A long-term strategy might also entail adding employment information to the database, to determine where nurses work, whether in the NHS or private sector, and how far a commute they experience depending on their SHA of employment. And any of these analyses could be broken down by source country of training, to see whether foreigners choose different career paths or locations from the domestically-trained nurse population. Coupled with vacancy and turnover rate data from the Department of Health, more detailed information about the student nurse population (particularly dropout rates and job choice following graduation), and better information about working conditions and wages across Strategic Health Authorities, this data could help the NHS better understand what policies to develop and where to target them (demographically or geographically) in order to improve nurse recruitment within the UK.

Competing interests
Financial support from Bristol-Myers Squibb is gratefully acknowledged. I have no competing interests.