Health Policy and Planning

journal article

LitStream Collection

South African physician emigration and return migration, 1991–2017: a trend analysis

Nwadiuko, Joseph; Switzer, Galen E; Stern, Jaime; Day, Candy; Paina, Ligia

2021 Health Policy and Planning

doi: 10.1093/heapol/czaa193pmid: 33778873

Abstract Although critical for understanding health labour market trends in low- and middle-income countries (LMICs), longitudinal LMIC health worker emigration and return migration trends are not routinely documented. This article seeks to better understand SA’s trends in physician emigration and return migration and whether economic growth and related policies affect migration patterns. This study used physician registry data to analyse patterns of emigration and return migration only among SA-trained physicians registered to practice in top destination countries such as Australia, Canada, New Zealand, the USA or the UK between 1991 and 2017, which represent the top five emigration destinations for this group. A linear regression model analysed the relationship between migration trends (as dependent variables) and SA’s economic growth, health financing and HIV prevalence (as independent variables). There has been a 6-fold decline in emigration rates from SA between 1991 and 2017 (from 1.8% to 0.3%/year), with declines in emigration to all five destination countries. About one in three (31.8% or 5095) SA physicians returned from destination countries as of 2017. Annual physician emigration fell by 0.16% for every $100 rise in SA GDP per capita (2011 international dollars) (95% confidence interval −0.60% to −0.086%). As of 2017, 21.6% (11 224) of all SA physicians had active registration in destination nations, down from a peak of 33.5% (16 366) in 2005, a decline largely due to return migration. Changes to the UK’s licensing regulations likely affected migration patterns while the Global Code of Practice on International Recruitment contributed little to changes. A country’s economic growth might influence physician emigration, with significant contribution from health workforce policy interventions. Return migration monitoring should be incorporated into health workforce planning. Sub-Saharan Africa, South Africa, human resources for health, physician migration, circular migration, return migration, global health, health policy KEY MESSAGES South African physician emigration has fallen precipitously since 2003, likely due to an improvement in the South African economy and UK changes in physician licensing regulations. Approximately one in three South African physicians returns to South Africa from abroad, largely from the UK and after the 1994 South African elections. Human resources for health (HRH) migration policies interact closely with home country conditions in determining migration outcomes. HRH-related policies should account for the integration of returnees and anticipate sudden shifts in emigration trends. Introduction Physician emigration from low- and middle-income countries (LMICs) to high-income countries is a topic of great concern to health policymakers (O’Brien and Gostin, 2011). It is estimated that 17.1% of all physicians in Organization for Economic Co-operation and Development (OECD) nations were of foreign origin in 2000 (Organization for Economic Co-operation and Development, 2016), many of whom have come from countries with physician shortages (Organization for Economic Co-operation and Development, 2007). An analysis of physician migration from nine sub-Saharan African countries estimated that the training costs of physicians who went to medical school in those countries but were practicing in OECD nations as of 2010 were more than $2 billion (Mills et al., 2011). This concern around health worker migration has brought about several policy level responses, including codes of practice discouraging international recruitment of health professionals adopted by the UK and the World Health Organization (WHO) (see Table 1). However, testing the effectiveness of such policies is limited by a paucity of published data on year-to-year (as opposed to single-year cross-sectional) physician emigration trends. The latest comprehensive trends of LMIC physician emigration cover the years 1991–2004 (Bhargava et al., 2011), followed by a few destination-country specific analyses focusing on immigration trends to the USA and UK as late as 2013 (Blacklock et al., 2012; Tankwanchi et al., 2015). This leaves unanswered the question of whether declines in LMIC emigration to one nation are compensated by rises in another. Table 1. Policies and Political changes potentially affecting South African Physician Emigration (Blacklock et al., 2012). Policies with explicit intent to curb SA or LMIC physician emigration 2003: List of countries published by the UK National Health Service from which active recruitment was prohibited, which predominantly included LMIC nations. UK recruitment ads in SA rose 191% between 2000 and 2004 before falling but were persistent as of 2010 (Rogerson, 2007; Dambisya and Mamabolo, 2012). 2004: Memorandum of Understanding between the UK and SA affirmed: Short term arranged mutual clinical exchanges were agreed upon between both countries, with re-affirmed commitment by the UK to curb active recruitment. 2008: Initiation of occupation specific dispensation (OSD): SA reorganization of salaries and career pathways within the public medical sector designed in part to decrease physician emigration. It was supposed to have increased physician salaries between 30% and 50% and effectively raised salaries of medical officers above their counterparts in UK and Australia by between USD $7000 and USD $12 000 per year (George and Rhodes, 2012). 2010: Adoption of World Health Organization (WHO) Global Code of Practice on the International Recruitment of Health Personnel, a voluntary code signed by member states to reduce recruitment of physicians from LMIC nations. Other policies with potential effect on SA physician emigration 1991–94: SA transition of power to post-apartheid government. 2003: Abolishment of automatic full registration status for physicians trained in SA, New Zealand, Australia, Hong Kong and the West Indies applying in the UK (General Medical Council, 2003) 2004: First funds disbursed from the US President’s Emergency Fund for AIDS Relief to South Africa. 2006: A 2003 Amendment to UK Medical Act is implemented for physicians, requiring European Economic Area (EEA) Migrants be given first priority where possible for medical positions and requiring work permits for non-EEA physicians. Policies with explicit intent to curb SA or LMIC physician emigration 2003: List of countries published by the UK National Health Service from which active recruitment was prohibited, which predominantly included LMIC nations. UK recruitment ads in SA rose 191% between 2000 and 2004 before falling but were persistent as of 2010 (Rogerson, 2007; Dambisya and Mamabolo, 2012). 2004: Memorandum of Understanding between the UK and SA affirmed: Short term arranged mutual clinical exchanges were agreed upon between both countries, with re-affirmed commitment by the UK to curb active recruitment. 2008: Initiation of occupation specific dispensation (OSD): SA reorganization of salaries and career pathways within the public medical sector designed in part to decrease physician emigration. It was supposed to have increased physician salaries between 30% and 50% and effectively raised salaries of medical officers above their counterparts in UK and Australia by between USD $7000 and USD $12 000 per year (George and Rhodes, 2012). 2010: Adoption of World Health Organization (WHO) Global Code of Practice on the International Recruitment of Health Personnel, a voluntary code signed by member states to reduce recruitment of physicians from LMIC nations. Other policies with potential effect on SA physician emigration 1991–94: SA transition of power to post-apartheid government. 2003: Abolishment of automatic full registration status for physicians trained in SA, New Zealand, Australia, Hong Kong and the West Indies applying in the UK (General Medical Council, 2003) 2004: First funds disbursed from the US President’s Emergency Fund for AIDS Relief to South Africa. 2006: A 2003 Amendment to UK Medical Act is implemented for physicians, requiring European Economic Area (EEA) Migrants be given first priority where possible for medical positions and requiring work permits for non-EEA physicians. Open in new tab Table 1. Policies and Political changes potentially affecting South African Physician Emigration (Blacklock et al., 2012). Policies with explicit intent to curb SA or LMIC physician emigration 2003: List of countries published by the UK National Health Service from which active recruitment was prohibited, which predominantly included LMIC nations. UK recruitment ads in SA rose 191% between 2000 and 2004 before falling but were persistent as of 2010 (Rogerson, 2007; Dambisya and Mamabolo, 2012). 2004: Memorandum of Understanding between the UK and SA affirmed: Short term arranged mutual clinical exchanges were agreed upon between both countries, with re-affirmed commitment by the UK to curb active recruitment. 2008: Initiation of occupation specific dispensation (OSD): SA reorganization of salaries and career pathways within the public medical sector designed in part to decrease physician emigration. It was supposed to have increased physician salaries between 30% and 50% and effectively raised salaries of medical officers above their counterparts in UK and Australia by between USD $7000 and USD $12 000 per year (George and Rhodes, 2012). 2010: Adoption of World Health Organization (WHO) Global Code of Practice on the International Recruitment of Health Personnel, a voluntary code signed by member states to reduce recruitment of physicians from LMIC nations. Other policies with potential effect on SA physician emigration 1991–94: SA transition of power to post-apartheid government. 2003: Abolishment of automatic full registration status for physicians trained in SA, New Zealand, Australia, Hong Kong and the West Indies applying in the UK (General Medical Council, 2003) 2004: First funds disbursed from the US President’s Emergency Fund for AIDS Relief to South Africa. 2006: A 2003 Amendment to UK Medical Act is implemented for physicians, requiring European Economic Area (EEA) Migrants be given first priority where possible for medical positions and requiring work permits for non-EEA physicians. Policies with explicit intent to curb SA or LMIC physician emigration 2003: List of countries published by the UK National Health Service from which active recruitment was prohibited, which predominantly included LMIC nations. UK recruitment ads in SA rose 191% between 2000 and 2004 before falling but were persistent as of 2010 (Rogerson, 2007; Dambisya and Mamabolo, 2012). 2004: Memorandum of Understanding between the UK and SA affirmed: Short term arranged mutual clinical exchanges were agreed upon between both countries, with re-affirmed commitment by the UK to curb active recruitment. 2008: Initiation of occupation specific dispensation (OSD): SA reorganization of salaries and career pathways within the public medical sector designed in part to decrease physician emigration. It was supposed to have increased physician salaries between 30% and 50% and effectively raised salaries of medical officers above their counterparts in UK and Australia by between USD $7000 and USD $12 000 per year (George and Rhodes, 2012). 2010: Adoption of World Health Organization (WHO) Global Code of Practice on the International Recruitment of Health Personnel, a voluntary code signed by member states to reduce recruitment of physicians from LMIC nations. Other policies with potential effect on SA physician emigration 1991–94: SA transition of power to post-apartheid government. 2003: Abolishment of automatic full registration status for physicians trained in SA, New Zealand, Australia, Hong Kong and the West Indies applying in the UK (General Medical Council, 2003) 2004: First funds disbursed from the US President’s Emergency Fund for AIDS Relief to South Africa. 2006: A 2003 Amendment to UK Medical Act is implemented for physicians, requiring European Economic Area (EEA) Migrants be given first priority where possible for medical positions and requiring work permits for non-EEA physicians. Open in new tab Furthermore, the migration of skilled professionals is best understood as a continuum, encompassing permanent emigration, return migration and diaspora effects (Newland, 2010; Docquier and Rapoport, 2012; Zimmermann, 2014; Del Carpio et al., 2016; Rapoport, 2018; Azose and Raftery, 2019; Constant, 2020). According to one analysis, one out of every four migration events worldwide is a return migration event (Azose and Raftery, 2019). Despite a growing body of research focused on return migration in other professions (Meyer et al., 1997; Docquier and Rapoport, 2012), only a small amount of this work has focused on health professionals, using survey data or qualitative interviews to describe return migration among sampled physicians from Ghana, the South Pacific, and Peru (Gaviria and Wintrob, 1975; Brown and Connell, 2004; Adzei and Sakyi, 2014). Existing large scale studies on migration have primarily focused on permanent emigration abroad instead of return migration as researchers typically only access registry data in OECD destination nations as opposed to those of physician’s origin nations (Bhargava et al., 2011). However, as many LMICs are undergoing ambitious economic transitions and health sector reforms, understanding return migration trends will be increasingly important in informing LMIC health worker staffing projections, as well as policies regarding physician retention and returnee integration. South Africa (SA) historically has had a substantial proportion of its physician workforce registered outside the country (estimated as high as 37% by earlier studies), largely in the UK (World Health Organization, 2006). SA has responded with several specific policies to limit health worker emigration, including a bilateral agreement with the UK and a rise in public health sector salaries (including occupation specific dispensation, an incentive-based pay scheme) (Blacklock et al., 2012; George and Rhodes, 2012). SA has also been affected by the UK’s recruitment ban (although UK-driven recruitment efforts targeting South Africans persisted as late as 2010) (Rogerson, 2007; Dambisya and Mamabolo, 2012) and a 2003 change in UK licensing regulations that ended automatic full registration status for SA, Australia, New Zealand, Hong Kong and the West Indies (General Medical Council, 2003). See Table 1 for further descriptions of all policies. SA has also recently experienced a wave of return migration in non-health sectors, with some sources estimating that as many as 350 000–500 000 skilled professionals across various industries had returned to SA from overseas as of 2014, though some controversy regarding the methodology producing these estimates remains (Kaplan and Höppli, 2017). One recently published paper on SA physician migration, using OECD data drawn from member country physician registries, estimated that 3449 SA-trained physicians had exited the registries of OECD member nations between 2010 and 2014, attributing it to possible return migration (Tankwanchi et al., 2019). However, the dataset does not allow any correction for retirement of SA-trained physicians nor does it account for physicians who might have registered across multiple nations (which might lead to inflated counts in the above estimation). Furthermore, a 2007 survey of SA-based physicians found that 48% of 1064 responding physicians had practiced outside SA in the past. This study was unfortunately limited by a low response rate (7%) (Crush et al., 2014). A registry-based analysis of trends and predictors of SA physician emigration and return migration is therefore timely and important. This article will first describe emigration and return migration trends amongst SA-trained physicians across five destination countries from 1991 to 2017. Second, it will determine whether SA economic growth, health financing increases, or HIV prevalence was correlated with changes in emigration or return migration, reflecting data that have shown a relationship between source country economic growth and HIV prevalence with physician emigration (Bhargava and Docquier, 2008; Arah et al., 2008; Okeke, 2013). Finally, it will examine the potential impact of health workforce policies and SA political changes (Table 1) on emigration and return migration trends. Methods In this article, we estimate SA physician migration patterns to and from top destination countries, including Australia (AUS), Canada (CAN), New Zealand (NZ), UK and the USA between 1991 and 2017, representing the period covered by physician registries made available to us. This approach included SA-trained physicians who had emigrated to these countries and SA-trained physicians who had returned from these countries to SA. We then fit these trends to a time-lagged ordinary least squared (OLS) model with economic and health systems data from SA to determine the relationship of GDP, HIV prevalence and health financing growth with emigration and return migration. Finally, we reflect on whether emigration and return migration were affected by the end of SA’s apartheid regime or a series of policies targeting health worker migration and registration [described in Table 1, with a particular focus on the 2003 UK National Health Service active international recruitment ban & GMC abolishment of automatic full registration status, and the 2010 adoption of the World Health Organization (WHO) Global Code of Practice on the International Recruitment of Health Personnel]. Data sources Table 2 summarizes the datasets that we used for the analysis. Prior investigations have indicated that between 1991 and 2014 five countries represented the destination of 88-99% of all emigrated SA trainees, (Bhargava et al., 2011; Tankwanchi et al., 2019). We requested registration data for all active physicians (as of July 2017) from the Health Professions Council of SA (HPCSA) as well as physicians with at least undergraduate medical training in SA and current or past practice in Australia, Canada, New Zealand, the UK and USA (‘destination countries’) from each country’s respective physician registry. We chose to identify physicians by their place of education instead of their place of birth given the financial costs of training to source countries (Mills et al., 2011). Table 2: Datasets for physician-level data Dataset source . Data provided . Australia (AUS) Medical Directory of Australia Data on all physicians registered between 1999 and 2017 (n = 1480) Canada (CAN) Canadian Institute of Health Information De-identified annual physician entry counts (1968–2016) Scott’s Medical Database Data on all physicians actively registered as of 2017 (n = 2541) Canadian Medical Association De-identified annual physician return migration counts, gathered from surveys (1995–2017) New Zealand (NZ) Medical Council of New Zealand Data on all physicians registered between 1989 and 2017 (n = 1656) South Africa (SA) Health Professions Council of South Africa Data on all physicians actively registered as of 2017 (n = 44 892) UK General Medical Council Data on all physicians registered between 1991 and 2017 (n = 17 680) USA American Medical Association MASTERFILE Data on all physicians registered between 1989 and 2017 (n = 2152) Dataset source . Data provided . Australia (AUS) Medical Directory of Australia Data on all physicians registered between 1999 and 2017 (n = 1480) Canada (CAN) Canadian Institute of Health Information De-identified annual physician entry counts (1968–2016) Scott’s Medical Database Data on all physicians actively registered as of 2017 (n = 2541) Canadian Medical Association De-identified annual physician return migration counts, gathered from surveys (1995–2017) New Zealand (NZ) Medical Council of New Zealand Data on all physicians registered between 1989 and 2017 (n = 1656) South Africa (SA) Health Professions Council of South Africa Data on all physicians actively registered as of 2017 (n = 44 892) UK General Medical Council Data on all physicians registered between 1991 and 2017 (n = 17 680) USA American Medical Association MASTERFILE Data on all physicians registered between 1989 and 2017 (n = 2152) Open in new tab Table 2: Datasets for physician-level data Dataset source . Data provided . Australia (AUS) Medical Directory of Australia Data on all physicians registered between 1999 and 2017 (n = 1480) Canada (CAN) Canadian Institute of Health Information De-identified annual physician entry counts (1968–2016) Scott’s Medical Database Data on all physicians actively registered as of 2017 (n = 2541) Canadian Medical Association De-identified annual physician return migration counts, gathered from surveys (1995–2017) New Zealand (NZ) Medical Council of New Zealand Data on all physicians registered between 1989 and 2017 (n = 1656) South Africa (SA) Health Professions Council of South Africa Data on all physicians actively registered as of 2017 (n = 44 892) UK General Medical Council Data on all physicians registered between 1991 and 2017 (n = 17 680) USA American Medical Association MASTERFILE Data on all physicians registered between 1989 and 2017 (n = 2152) Dataset source . Data provided . Australia (AUS) Medical Directory of Australia Data on all physicians registered between 1999 and 2017 (n = 1480) Canada (CAN) Canadian Institute of Health Information De-identified annual physician entry counts (1968–2016) Scott’s Medical Database Data on all physicians actively registered as of 2017 (n = 2541) Canadian Medical Association De-identified annual physician return migration counts, gathered from surveys (1995–2017) New Zealand (NZ) Medical Council of New Zealand Data on all physicians registered between 1989 and 2017 (n = 1656) South Africa (SA) Health Professions Council of South Africa Data on all physicians actively registered as of 2017 (n = 44 892) UK General Medical Council Data on all physicians registered between 1991 and 2017 (n = 17 680) USA American Medical Association MASTERFILE Data on all physicians registered between 1989 and 2017 (n = 2152) Open in new tab Physician-level data To capture the physicians registered in destination countries, we obtained historical and current physician-level data from AUS, NZ, the US and the UK and current physician-level and historical annual count data from medical registries in CAN (given the lack of available historical physician-level data in CAN). Data sources are detailed in Table 2. Physician-level data of active SA physicians as of July 2017 was obtained from the Health Professions Council of SA (HPCSA). All physician-level databases at a minimum included physician first and last names, medical school and medical school year of graduation. Historical counts of SA-registered physicians were gathered from HPCSA’s annual reports, published literature and the WHO Global Health Workforce Database for 1992, 1996 and 1999–2017 with linear interpolation in-between and linear extrapolation to 1991 (Health Professions Council of South Africa, 2017; World Health Organization, 2018). National level predictor data We drew SA GDP per capita and health expenditure per capita data (adjusted for inflation) from national account data via the World Bank Data Catalog and WHO Global Health Expenditure Database. We drew SA HIV prevalence data from the Joint United Nations Programme on HIV and AIDS (UNAIDS). Primary outcomes Emigration We derived emigration rates from AUS, NZ, UK and US physician registries and de-identified Canadian physician entry counts (1969–2016) with extrapolation to 2017. We excluded physicians who did not achieve full registration status in those countries (n = 13). For physicians registered in more than one destination country, we attributed only the earliest year of full registration to them and they were assigned as having gone only to the first country of registry to avoid artificially inflating emigration totals. We calculated the annual SA emigration rate (i.e. annual new physician flows) as: Annual emigration rate=DD+S × 100,(1) where D represented the number of SA-trained physicians who newly registered in destination countries each year and S represented the number of physicians registered with the HPCSA the same year. We describe the estimation method for determining the total number of existing physician emigrants (i.e. total annual physician stocks) registered in destination countries in the Supplementary material. Return migration Figure 1 summarizes our return migration determination method. In the first step of our return migration analysis, we cross-referenced registration data from physicians in NZ, US and UK databases with the HPCSA database by first and last name, medical school, and year of graduation using probabilistic matching with a required minimum match of 80%. Probabilistic (or fuzzy) matching relies on the approximate (rather than exact) comparison of paired fields from two separate datasets (Wasi and Flaaen, 2015). Matches were reviewed by the first author manually for fidelity. (We could not use data from Australia given unreliable registration expiry dates for SA-trained physicians in its database; we estimated return migration from Canada using separate survey data described below). Figure 1: Open in new tabDownload slide Identifying return migrants amongst South African-trained physicians. Figure 1: Open in new tabDownload slide Identifying return migrants amongst South African-trained physicians. SA-trained physicians in NZ/US/UK databases were considered to have returned if they (1) were successfully matched with the 2017 HPCSA database and (2) if they were determined to have voluntarily withdrawn their registration from all five destination countries by a retirement age of 65 (estimated by a medical school graduation year of 1976, assuming entry into a 6-year medical school at the age of 18). We excluded physicians who were reported deceased (n = 205); administratively suspended by medical boards (n = 20); or who shared an identical first name, last name, medical school and graduation date as someone else in the same destination country database (n = 10). If an individual had missing medical school information (n = 234) they were kept for matching only if they had listed their nationality as South African and had a year of graduation listed (resulting in 46 kept, 188 excluded). The retirement age of 65 is supported by cross-sectional and qualitative data on physician retirement in Australia, Canada, New Zealand, the USA and the UK which suggested a median age of retirement between 60 and 69 years of age (Silver et al., 2016). For physicians registered in multiple destination countries, departure dates and countries were assigned based on the latest date of registration expiry. Since the Canadian physician-level dataset excluded physicians who had dropped out, we relied on national survey data to determine return counts. The Canadian Medical Association from 1995 onward conducted annual surveys of its physicians to determine the destination of physicians who they determined to have departed from Canada with a self-reported 75% response rate; this data is reported in other literature (Watanabe et al., 2008) and was requisitioned by our team to determine return migration from Canada. Return migration rates were calculated as: Annual return rate=RR+P × 100,(2) where R represented the number of newly returned physicians in a given year, and P represented the total number of SA-trained physicians registered to practice in destination countries for that same year. As stated earlier, we describe the estimation method for determining the annual total number P (i.e. physician stocks) registered in destination countries in the Supplementary material. Analysis Regression analysis We tested the relationships of annual SA Gross Domestic Product (GDP) (1991–2017), total annual SA health expenditures per capita (2000–16) and annual SA HIV prevalence (1991–2017) with emigration and return migration rates in a linear regression model: Model 1: Yt=β0+ β1Xt-1+β2Yeart controlling for the year as a linear covariate (⁠ Yeart, to try to control for unobserved confounding in annual trends) and a 1-year lag of all independent variables (⁠ Xt-1 ⁠, with the assumption that migration decisions are not made immediately in relationship to market or health system factors). Annual emigration and return migration rates represented dependent variables (Yt). All regressions were done with robust variance estimates. All analyses were done in STATA 14 (Stata Corp; College Station, TX). Policy analysis We explored the effect of policies (see Table 1) on emigration and return migration trends, with a particular focus on the 1994 end of the SA apartheid regime, 2003 changes in UK GMC registration and National Health Service (NHS) recruitment policy and the 2010 adoption of the World Health Organization (WHO) Global Code of Practice on the International Recruitment of Health Personnel. Ethical approval The protocol was determined to not meet criteria for human subject research by the Institutional Review Board of the authors’ institute. Results As of 2017, there were 44 892 Health Professional Council of SA (HPCSA) registered physicians, of which 188 were excluded from matching. We obtained identifying historical records in OECD destination nations for 21 838 SA-trained physicians, of which 248 were excluded from all analysis. Of the 21 590 remaining identified SA-trained physicians in OECD destination nation datasets, 11 224 (52%) had active registrations to practice outside of SA as of the end of 2017 (see Table 3), 7216 (33%) had withdrawn their registration before the age of 65; and 9871 (46%) were successfully matched with the 2017 HPCSA database, with an average match score of 99.6% (range 85.6–100%). Table 3. Distribution of emigrated South African physicians, by country of practice (if actively registered in 2017) Australia 1312 Canada 2540 New Zealand 1079 UK 6002 USA 1681 History of registration in two destination countries 2980 History of registration in three destination countries 155 History of registration in four destination countries 3 Australia 1312 Canada 2540 New Zealand 1079 UK 6002 USA 1681 History of registration in two destination countries 2980 History of registration in three destination countries 155 History of registration in four destination countries 3 Given presence of multi-nation registrants, physician total for nation specific counts will equal greater than total number of physicians. Destination countries: Australia, Canada, New Zealand, UK, USA. Open in new tab Table 3. Distribution of emigrated South African physicians, by country of practice (if actively registered in 2017) Australia 1312 Canada 2540 New Zealand 1079 UK 6002 USA 1681 History of registration in two destination countries 2980 History of registration in three destination countries 155 History of registration in four destination countries 3 Australia 1312 Canada 2540 New Zealand 1079 UK 6002 USA 1681 History of registration in two destination countries 2980 History of registration in three destination countries 155 History of registration in four destination countries 3 Given presence of multi-nation registrants, physician total for nation specific counts will equal greater than total number of physicians. Destination countries: Australia, Canada, New Zealand, UK, USA. Open in new tab Emigration Overall annual emigration rates rose from 1.8% (405 registrations) in 1991 to a peak of 9.7% (2947 registrations) in 2003, fell to 1.3% in 2004 and continued to decline to 0.3% in 2017 (145 registrations) (Figure 2). This included a relative decline of 98.5% in the number of registrations to the UK between 2003 and 2004. Inside non-UK destination countries there was a 64.3% relative fall of new registrations between 1991 and 2017, with falls beginning between 2000 and 2003 (−100% AUS, −47.4% CAN, −82.1% NZ, −100% USA) (Supplementary Figures S1 and S2). As of 2017, 21.6% (11 224) of all SA physicians had active registration in destination nations, down from a peak of 33.5% (16 366) in 2005 (Figure 3). Over 1991–2017, UK bound emigration accounted for an average of 50% of annual SA physician emigration. Figure 2: Open in new tabDownload slide South African emigration and return rates by year. Figure 2: Open in new tabDownload slide South African emigration and return rates by year. Figure 3: Open in new tabDownload slide Percent of South African physicians registered in destination countries, 1991–2017. Figure 3: Open in new tabDownload slide Percent of South African physicians registered in destination countries, 1991–2017. In a year-adjusted OLS model (i.e. with a covariate for year), declining emigration from 1991 to 2017 was correlated with SA GDP per capita, with a decline in annual migration rate of 0.16% for every $100 rise in GDP per capita (2011 international dollars) (95% CI −0.309 to −0.021%) (Table 4, Figure 4). There was no significant correlation between emigration trends and SA Total Health Expenditure or HIV prevalence. There was a significant immediate decline in SA-to-UK emigration rates after 2003 changes in UK recruitment and registration policy and no immediate change in emigration rate trends after adoption of the WHO Global Code (Figure 2). Figure 4: Open in new tabDownload slide South African physician emigration rate and South African GDP per capita. Figure 4: Open in new tabDownload slide South African physician emigration rate and South African GDP per capita. Table 4: National level factors and their relation to emigration/return migration . ANNUAL emigration rate (%) . ANNUAL return migration rate (%) . . OLS, year adjusted . OLS, year adjusted . SA GDP per capita, PPP ($100 Intl 2011) (1991–2017) −0.165* [−0.309, −0.0210] −0.0431 [−0.137, 0.0508] SA total health expenditure per capita ($100 PPP Intl 2011) (2000–16) 0.0138 [−0.00833, 0.0359] −0.0134 [−0.0299, 0.03] SA HIV prevalence (per 100 000) population (1991–2017) 0.000220 [−0.000461, 0.000901] 0.000515** [0.000141, 0.000889] N See footnote See footnote . ANNUAL emigration rate (%) . ANNUAL return migration rate (%) . . OLS, year adjusted . OLS, year adjusted . SA GDP per capita, PPP ($100 Intl 2011) (1991–2017) −0.165* [−0.309, −0.0210] −0.0431 [−0.137, 0.0508] SA total health expenditure per capita ($100 PPP Intl 2011) (2000–16) 0.0138 [−0.00833, 0.0359] −0.0134 [−0.0299, 0.03] SA HIV prevalence (per 100 000) population (1991–2017) 0.000220 [−0.000461, 0.000901] 0.000515** [0.000141, 0.000889] N See footnote See footnote 95% confidence intervals in brackets. * P < 0.05, ** P < 0.01, *** P < 0.001. N=varied depending on years of data available for independent predictors; 27 for GDP per capita and HIV Prevalence; 17 for SA Total Health expenditure per capita. Limitations in date range reflect limitations in data availability. Each model is adjusted for year and the individual predictor in question and represents a 1-year time lag in independent variables (Yt=β0 + β1Xt-1+ β2Yeart). GDP, gross domestic product; OLS, ordinary least squares; PPP, purchasing power parity; SA, South Africa. Open in new tab Table 4: National level factors and their relation to emigration/return migration . ANNUAL emigration rate (%) . ANNUAL return migration rate (%) . . OLS, year adjusted . OLS, year adjusted . SA GDP per capita, PPP ($100 Intl 2011) (1991–2017) −0.165* [−0.309, −0.0210] −0.0431 [−0.137, 0.0508] SA total health expenditure per capita ($100 PPP Intl 2011) (2000–16) 0.0138 [−0.00833, 0.0359] −0.0134 [−0.0299, 0.03] SA HIV prevalence (per 100 000) population (1991–2017) 0.000220 [−0.000461, 0.000901] 0.000515** [0.000141, 0.000889] N See footnote See footnote . ANNUAL emigration rate (%) . ANNUAL return migration rate (%) . . OLS, year adjusted . OLS, year adjusted . SA GDP per capita, PPP ($100 Intl 2011) (1991–2017) −0.165* [−0.309, −0.0210] −0.0431 [−0.137, 0.0508] SA total health expenditure per capita ($100 PPP Intl 2011) (2000–16) 0.0138 [−0.00833, 0.0359] −0.0134 [−0.0299, 0.03] SA HIV prevalence (per 100 000) population (1991–2017) 0.000220 [−0.000461, 0.000901] 0.000515** [0.000141, 0.000889] N See footnote See footnote 95% confidence intervals in brackets. * P < 0.05, ** P < 0.01, *** P < 0.001. N=varied depending on years of data available for independent predictors; 27 for GDP per capita and HIV Prevalence; 17 for SA Total Health expenditure per capita. Limitations in date range reflect limitations in data availability. Each model is adjusted for year and the individual predictor in question and represents a 1-year time lag in independent variables (Yt=β0 + β1Xt-1+ β2Yeart). GDP, gross domestic product; OLS, ordinary least squares; PPP, purchasing power parity; SA, South Africa. Open in new tab Return migration 5095 physicians met criteria for return migration, 4966 from the UK, 133 from Canada, 116 from NZ and 13 from the USA, representing a return rate of 35.7%, 0.3%, 8.3% and 1.1% of active & historical SA-trained physicians returning from the UK, New Zealand, Canada and the USA, respectively (Table 5); returning physicians composed 11.3% of all 2017 HPCSA licensed physicians. Mean time between destination country registration and registration cancellation among return migrants was 6.7 years, with a median estimated age of 34 at return. The total annual return rate rose from a 1992 to 1994 average of 0.2% to a 1995–2009 average of 1.8% (after the apartheid transition period), after which it declines. Between 2005 and 2017, the annual count of return migrants closely matches and at times overcomes the number of emigrants (Supplementary Figure S3). In fact, 52.2% of the overall decline in SA-trained physicians in destination countries (2937/5144 physicians) since 2005 was due to return migration. Table 5. South African-trained physician return rate, by destination country of registration . Number of returned physicians . Number of ‘eligible’ physiciansa . Return migration rate (%) . UK 4966 13 918 35.7 Canada 133 0.3b New Zealand 116 1399 8.3 USA 13 1135 1.1 Total 5095 15 999 31.8 . Number of returned physicians . Number of ‘eligible’ physiciansa . Return migration rate (%) . UK 4966 13 918 35.7 Canada 133 0.3b New Zealand 116 1399 8.3 USA 13 1135 1.1 Total 5095 15 999 31.8 a All SA-trained physicians with history of training in destination countries under the age of 65 (estimated by graduation 1976 or after). b Averaged from annual counts between 1995 and 2017. Open in new tab Table 5. South African-trained physician return rate, by destination country of registration . Number of returned physicians . Number of ‘eligible’ physiciansa . Return migration rate (%) . UK 4966 13 918 35.7 Canada 133 0.3b New Zealand 116 1399 8.3 USA 13 1135 1.1 Total 5095 15 999 31.8 . Number of returned physicians . Number of ‘eligible’ physiciansa . Return migration rate (%) . UK 4966 13 918 35.7 Canada 133 0.3b New Zealand 116 1399 8.3 USA 13 1135 1.1 Total 5095 15 999 31.8 a All SA-trained physicians with history of training in destination countries under the age of 65 (estimated by graduation 1976 or after). b Averaged from annual counts between 1995 and 2017. Open in new tab There was a substantial discontinuity in return rate among UK emigrants between those who first registered in the UK during 1991–2005 (30.9%) and those who first registered there between 2006 and 2017 (8.7%), a difference not seen in return rates from New Zealand registrants (Figure 5). In the OLS model, there was a minimally positive correlation between HIV prevalence and return migration rates, with a 0.000515% increase in return per additional 1/100 000 cases (95% CI 0.000141–0.000889). Figure 5: Open in new tabDownload slide Return rate by year of registration in UK and NZ. Figure 5: Open in new tabDownload slide Return rate by year of registration in UK and NZ. Discussion SA physician emigration rose in the 1990’s until the early 2000’s when it fell precipitously in the UK and more gradually in other countries, reaching a new equilibrium between 2004 and 2017, below its 1991 annual rate. This is accompanied by a rise of return migration rates starting in 1995. Overall, the data support a shift in migration trends between SA-trained physicians and SA over the past 25 years with fewer physicians leaving the country and increasing numbers returning after practicing abroad. Emigration SA GDP per capita was inversely associated with emigration in the regression model, suggesting that economic growth in SA contributed to the decline in emigration. The negative correlation between economic growth and physician emigration has been recorded elsewhere (Arah et al., 2008; Okeke, 2013). SA is likely benefiting from its status as an upper-middle income country with greater economic power to retain its medical professionals, augmented by a rise in economic growth in the early 2000s. Emigration declines in SA do not appear to signal broader declines in emigration of Sub-Saharan African physicians to high-income countries. In fact, emigration of Sub-Saharan African physicians to the USA has increased and has remained constant to the UK (Blacklock et al., 2012; Tankwanchi et al., 2013). The strength of this apparent correlation might suggest that future severe economic downtowns, particularly in the time of the SARS-COV-2 pandemic, might lead to a rebound in emigration desire. Whether that translates to an actual increase in emigration is unknown and will depend in no small part in the immigration policies of high-income country destinations. The 2001–03 rise in SA physician emigration to the UK is contemporaneous with a rise in UK migration by all South Africans and a sharp rise in physician-specific recruitment to the UK (Supplementary Tables S4 and S5) (Rogerson, 2007; Dambisya and Mamabolo, 2012; Determinants of International Migration, 2015). The rise in overall emigration from South Africa had been linked to increased inequality and crime starting in the late 1990s (Sveinsson and Gumuschian, 2008). However, the 2003 National Health Service recruitment ban alone is unlikely to have caused the drastic fall in SA–UK physician emigration rates between 2003 and 2004: emigration rates to the UK took a 87% relative fall between 2003 and 2004, despite an 191% rise in recruitment ads from the UK between 2000 and 2004 (Rogerson, 2007; Dambisya and Mamabolo, 2012). Based on news reports and qualitative studies, the end of automatic registration for physicians trained in SA and other selected Commonwealth nations has been linked to the decline in emigration rates in 2004 (Smetherham, 2003; Labonté et al., 2015) (and in this context there is at some possibility that at least a portion of the 2001–3 rise might have been anticipatory of this policy change). The 2004 launch of the US President's Emergency Plan for AIDS Relief (PEPFAR) funding to SA might have also contributed to post 2004 downward emigration trends, although not likely to the 2003–4 decline itself. The 2010 Global Code of Practice on the International Recruitment of Health Personnel seems to have had little additional impact on already decreasing SA emigration rates, which might be due to limited implementation (Edge and Hoffman, 2013). Likewise, public-sector health worker salary raises enacted in SA in 2008 (i.e. occupational specific dispensation) (George and Rhodes, 2012) (see Table 1) likely did not result in significant shifts in emigration trends. This likely reflects the diversity of factors that impact migration decisions besides income, as has been shown with other segments of South African health workforce (Blaauw et al., 2010). On the other hand, there is some qualitative evidence that occupational specific dispensation might have decreased internal and international migration among nurses (Labonté et al., 2015). Return migration Compared to other host nations, elevated return rates from the UK are consistent with the SA-to-UK circular migration hypothesis suggested by previous studies (Sveinsson and Gumuschian, 2008; Andrucki, 2010). This hypothesis argues that due to close historical linkages in part due to South Africa’s status as a former settler colony, the UK has offered more favourable visa policies for South African nationals. This has led to increased emigration of SA nationals to the UK, followed after brief periods by return migration to SA. In a survey of health professional return migrants, (Crush et al. (2014) noted the most common contributors to return by physicians, dentists and pharmacists was related to the temporary nature of their position (61.6%), familial desire to return (61.9%), a permanent job offer in SA (63.8%) and lifestyle, cultural and environmental factors in SA (60.9–86.5%). It is possible that increased political stability in SA, particularly after the 1994 presidential election, contributed to the 1995 rise in return migration rates, which was sustained until 2010. The decline of return migration after 2010 might be an unintended consequence of a change in the UK’s immigration policy regarding physicians. In 2006, the UK applied a previously passed revision to Medical Act to international medical graduates, giving priority to European Economic Area (EEA) graduates for positions and later requiring work permits from non EEA graduates for the first time (Blacklock et al., 2012) (see Table 1). These increased requirements for certification are what likely caused the discontinuity in return rates amongst UK-based emigrants between those who registered from 1991 to 2005 (30.9%) and those who registered there between 2006 and 2017 (8.7%) (Figure 5), perhaps indicating a lower willingness to surrender licenses gained under a more competitive environment. No such shift was witnessed in the return rate from NZ. This phenomenon is not dissimilar to other scenarios where raised entry barriers paradoxically decreased return migration, such as the US Southern Border or within Europe (Zimmermann, 2014). There is a need for further investigation for other causes of the variation in return migration trends. While we found a positive correlation between HIV rates and return migration, it was of small magnitude. While it might reflect increased return for charitable purposes or in response to demand, further investigation is needed. Limitations This is a retrospective observational study and causal claims cannot be made with certainty. Emigration data from Australia before 1999 might represent an underestimate of registration counts, although Australia-destined emigration accounted for a small fraction of SA emigration overall. More importantly, ‘registration’ carries heterogeneous significance across countries: while it is synonymous with licensure to practice in Australia, Canada and the USA, in NZ registration is only a prerequisite to licensure, and registration and licensure were similarly separated into two separate categories by the UK General Medical Council in 2009 (General Medical Council, 2009). Both cases might lead to emigration rates being slightly inflated over time. It was not possible to determine previous registration or licensure status after it had been cancelled in NZ or the UK, however the fact that UK return rates did not rise after 2009 or significantly differ amongst those who registered in the UK before or after 2009 (Figure 5) suggests that registrants without licenses likely do not account for a significant number of returnees. However, there is the possibility that a physician might have retired before the age of 65 with differential cancellation of registration status (i.e. cancellation across OECD destination countries but not South African registries), leading to an possible overestimate of return migration. On the other hand, physicians might not register in OECD destinations when relocating, effectively leaving the clinical workforce, leading to underestimates of emigration rates. Finally, they might also maintain registration in those countries after leaving them to return to South Africa, leading to underestimates of return migration through our approach. Conclusion and implications Between 1991 and 2017, SA emigration rates declined 6-fold and return migration rates rose, with the net effect that one of three of all SA expatriates had returned as of 2017. Market, policy and political factors contributed to changes in emigration and return migration trends, most prominent being SA’s economic growth, the 1994 SA general election and regulatory licensing changes by the UK. The interplay between market and policy effects on migration trends is important: SA GDP trends alone likely do not explain the 2003–4 fall in emigration or the depressed return migration rate prior to 1995, for example. Likewise, it is possible SA’s prosperity allowed policy changes to be effective in reducing emigration in ways not possible for physician emigrants from poorer source countries. A deeper understanding of the complexity of factors that guide LMIC physician migration decisions is important for more effective management of migration flows by policymakers in both destination and recipient countries. The WHO Global Code of Practice includes circular migration as a policy option, stating that: ‘Member States should facilitate circular migration of health personnel, so that skills and knowledge can be achieved to the benefit of both source and destination countries’ and that ‘Member States are encouraged to…assess the scope and impact of circular migration’ (Joint Action Health Workforce Planning and Forecasting, 2016). This analysis suggests how emigration and return migration trends can be successfully monitored as policy outcomes. Other nations wishing to assess return migration may utilize registry-based analyses such as those introduced in this article or employ surveys (like those sent by the Canadian Medical Association) (Watanabe et al., 2008). Of course, either scenario will require health workforce planners to invest significant resources in data collection and data sharing across countries. Supplementary data Supplementary data are available at Health Policy and Planning online. Acknowledgements We would like to acknowledge the following for feedback: Krisda Chaiyachati (University of Pennsylvania), Caglar Ozden (World Bank), Mark Pauly (University of Pennsylvania), Marilyn Schapira (University of Pennsylvania), Caroline Theoharides (Amherst College), Harsha Thirumurthy (University of Pennsylvania), as well as agencies who provided data for this project (American Medical Association, Medical Directory of Australia, Canadian Institute of Health Information, Scott’s Directories, Canadian Medical Association, Medical Council of New Zealand, Health Professional Council of South Africa, UK General Medical Council). Funding Funding has been provided by the Migration and Remittances Unit of the World Bank Group. Dr. Nwadiuko was supported by training grant T32HP1002623 from the US National Institutes of Health. Conflict of interest statement. All authors have completed the disclosure form at www.icmje.org/coi_disclosure.pdf and declare funding from the World Bank (related to this project), but no other relevant conflicts of interest, financial or otherwise. Ethical approval. This project was deemed to be not human subjects research by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board. References Adzei FA Sakyi EK. 2014 . Drivers of return migration of Ghanaian health professionals: perspectives from doctors and nurses in urban Ghana . International Journal of Migration, Health and Social Care 10 : 102 – 20 . Google Scholar OpenURL Placeholder Text WorldCat Andrucki MJ. 2010 . The visa whiteness machine: transnational motility in post-apartheid South Africa . Ethnicities 10 : 358 – 70 . Google Scholar OpenURL Placeholder Text WorldCat Arah O. A Ogbu UC Okeke CE. 2008 . Too poor to leave, too rich to stay: developmental and global health correlates of physician migration to the United States, Canada, Australia, and the United kingdom . American Journal of Public Health 98 : 148 – 54 . Google Scholar OpenURL Placeholder Text WorldCat Azose JJ Raftery AE. 2019 . Estimation of emigration, return migration, and transit migration between all pairs of countries . Proceedings of the National Academy of Sciences of the United States of America 116 : 116 – 22 . Google Scholar OpenURL Placeholder Text WorldCat Bhargava A Docquier F. 2008 . HIV Pandemic, Medical Brain Drain, and Economic Development in Sub-Saharan Africa . The World Bank Economic Review 22 : 345 – 66 . Google Scholar OpenURL Placeholder Text WorldCat Bhargava A Docquier F Moullan Y. 2011 . Modeling the effects of physician emigration on human development . Economics & Human Biology 9 : 172 – 31 . Google Scholar OpenURL Placeholder Text WorldCat Blaauw D Erasmus E Pagaiya N et al. 2010 . Policy interventions that attract nurses to rural areas: a multicountry discrete choice experiment . Bulletin of the World Health Organization 88 : 350 – 6 . Google Scholar OpenURL Placeholder Text WorldCat Blacklock C Heneghan C Mant D Ward AM. 2012 . Effect of UK policy on medical migration: a time series analysis of physician registration data . Human Resources for Health 10 : 35 . Google Scholar OpenURL Placeholder Text WorldCat Brown RPC Connell J. 2004 . The migration of doctors and nurses from South Pacific Island Nations . Social Science & Medicine 58 : 2193 – 210 . Google Scholar OpenURL Placeholder Text WorldCat Del Carpio X Ozden C Testaverde M Wagner M. 2016 . Global Migration of Talent and Tax Incentives: Evidence from Malaysia’s Returning Expert Program. World Bank Policy Research Working Paper Series No. 7875. World Bank, Washington, DC. Constant AF. 2020 . Time Space Dynamics of Circular and Return Migration: Theories and Evidence. GLO Discussion Paper No. 446, Global Labor Organization, Essen. Crush J Chikanda A Bourgeault IL Labonté R Murphy GT. 2014 . Brain Drain and Regain: The Migration Behaviour of South African Medical Professionals. South African Migration Programme Working Paper Series, Cape Town. Dambisya YM Mamabolo MH. 2012 . Foreign advertisements for doctors in the SAMJ 2006 - 2010 . South African Medical Journal 102 : 669 . Google Scholar OpenURL Placeholder Text WorldCat Determinants of International Migration . 2015 . DEMIG C2C, Version 1.2 , Limited Online Edition. Oxford : International Migration Institute, University of Oxford . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Docquier F Rapoport H. 2012 . Globalization, brain drain, and development . Journal of Economic Literature 50 : 681 – 730 . Google Scholar OpenURL Placeholder Text WorldCat Edge JS Hoffman SJ. 2013 . Empirical impact evaluation of the WHO Global Code of Practice on the International Recruitment of Health Personnel in Australia, Canada, UK and USA . Globalization and Health 9 : 60 . Google Scholar OpenURL Placeholder Text WorldCat Gaviria M Wintrob R. 1975 . Foreign medical graduates who return home after US residency training: the Peruvian case . Academic Medicine 50 : 167 – 75 . Google Scholar OpenURL Placeholder Text WorldCat General Medical Council . 2003 . No Changes to the registration process for doctors who have qualified in Australia, Hong Kong, New Zealand, Singapore, South Africa, West Indies before 31st October 2003 or the University of Malaya before 31st December 1989. https://web.archive.org/web/20040109214126/http://www.gmc-uk.org:80/register/regprocessfaq.htm (Accessed 14 December, 2018). General Medical Council . 2009 . 213,000 doctors respond to GMC ‘Licensing; it’s time to choose’ campaign. https://web.archive.org/web/20100206185314/http://www.gmc-uk.org/news/4893.asp (19 June 2019, last accessed). George G Rhodes B. 2012 . Is there really a pot of gold at the end of the rainbow? Has the Occupational Specific Dispensation, as a mechanism to attract and retain health workers in South Africa, leveled the playing field? BMC Public Health 12 : 1 – 9 . Google Scholar OpenURL Placeholder Text WorldCat Health Professions Council of South Africa . 2017 . Annual Reports, 2006-2017. Pretoria, South Africa. Kaplan D Höppli T. 2017 . The South African brain drain: an empirical assessment . Development Southern Africa 34 : 497 – 18 . Google Scholar OpenURL Placeholder Text WorldCat Labonté R Sanders D Mathole T et al. 2015 . Health worker migration from South Africa: causes, consequences and policy responses . Human Resources for Health 13 : 92 . Google Scholar OpenURL Placeholder Text WorldCat Meyer J-B Charum J Bernal D et al. 1997 . Turning brain drain into brain gain: the Colombian experience of the diaspora option . Science Technology & Society 2 : 285 – 315 . Google Scholar OpenURL Placeholder Text WorldCat Mills EJ Kanters S Hagopian A et al. 2011 . The financial cost of doctors emigrating from sub-Saharan Africa: human capital analysis . BMJ 343 : d7031 – d7031 . Google Scholar OpenURL Placeholder Text WorldCat Newland K (ed). 2010 . Diasporas: New Partners in Global Development Policy . Washington : Migration Policy Institute . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC O’Brien P Gostin LO. 2011 . Health Worker Shortages and Global Justice. Milbank Memorial Fund, New York. Okeke EN. 2013 . Brain drain: do economic conditions ‘push’ doctors out of developing countries? Social Science and Medicine 98: 169 – 178 . Google Scholar OpenURL Placeholder Text WorldCat Organization for Economic Co-operation and Development . 2007 . Immigrant Health Workers in OECD Countries in the Broader Context of Highly Skilled Migration. International Migration Outlook Annual Report, 161 – 228 . Organization for Economic Co-operation and Development . 2016 . Health Workforce Policies in OECD Countries: Right Jobs, Right Skills, Right Places . OECD , Paris. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rapoport H. 2018 . Diaspora externalities: a view from the South. UNU-WIDER Working Paper 2018/25, Helsinki. Rogerson CM. 2007 . Medical Recruits: The Temptation of South African Health Care Professionals. Southern African Migration Programme, Working Paper Series, Cape Town. Silver MP Hamilton AD Biswas A Warrick NI. 2016 . A systematic review of physician retirement planning . Human Resources for Health 14 : 67 Google Scholar OpenURL Placeholder Text WorldCat Smetherham J. 2003 . UK work restrictions a blow to SA doctors. IOL. https://www.iol.co.za/news/south-africa/uk-work-restrictions-a-blow-to-sa-doctors-106938 (Accessed 15 December 2018). Sveinsson KP Gumuschian A. 2008 . Understanding Diversity: South Africans in Multi-Ethnic Britain. Runnymede Community Studies, The Runnymede Trust, London. Tankwanchi ABS Ozden C Vermund SH. 2013 . Physician emigration from sub-Saharan Africa to the United States: analysis of the 2011 AMA physician masterfile . PLoS Medicine 10 : e1001513 . Google Scholar OpenURL Placeholder Text WorldCat Tankwanchi ABS Vermund SH Perkins DD. 2015 . Monitoring Sub-Saharan African physician migration and recruitment post-adoption of the WHO code of practice: temporal and geographic patterns in the United States . PLoS One 10 : e0124734 . Google Scholar OpenURL Placeholder Text WorldCat Tankwanchi AS Hagopian A Vermund SH. 2019 . International migration of health labour: monitoring the two-way flow of physicians in South Africa . BMJ Global Health 4 : e001566 . Google Scholar OpenURL Placeholder Text WorldCat The Joint Action on Health Workforce Planning and Forecasting . 2016 . Circular Migration of the Health Workforce. Wasi N Flaaen A. 2015 . Record linkage using Stata: preprocessing, linking, and reviewing utilities . The Stata Journal: Promoting Communications on Statistics and Stata 15 : 672 – 97 . Google Scholar OpenURL Placeholder Text WorldCat Watanabe M Comeau M Buske L. 2008 . Analysis of international migration patterns affecting physician supply in Canada . Healthcare Policy 3 : e129–38 . Google Scholar OpenURL Placeholder Text WorldCat World Health Organization . 2006 . Working together for health: World Health Report 2006. Geneva. World Health Organization . 2018 . Global Health Workforce Database. Geneva. Zimmermann K 2014 . Circular migration . IZA World of Labor, IZA, Bonn, Germany. Google Scholar OpenURL Placeholder Text WorldCat © The Author(s) 2021. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine. All rights reserved. For permissions, please e-mail: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

1986

Related Journals: