Correlates of African Health Worker Migration to the United States

Metrics and Correlates of African Health-Worker Migration to the United States, United Kingdom and Seven Other Wealthy Destinations O. A. Arah, C. E. Okeke, and U. C. Ogbu Academic Medical Center University of Amsterdam, the Netherlands & UCLA School of Public Health, Los Angeles AcademyHealth Annual Research Meeting Orlando, June 4, 2007 INTRODUCTION  Some 57, mostly African, countries are said to face crippling health-worker shortages: global deficit of 2.4 million  Migration from developing to developed countries seen as a major contributor  United States, United Kingdom, Canada, Australia, and other wealthy destinations  International medical/nursing graduates make up at least a quarter of the health workforce in some destination countries  Health-worker migration studies often lump source countries together as poor, without checking if these countries display a systematic migration differential  Unclear from the literature if source countries proportionately lose physicians and nurses alike to the same destinations  To quantify these flows, studies often rely on two common migration metrics – Absolute number of émigrés – Emigration fraction: émigrés as a proportion of the source country’s workforce prior to emigration  Aims of this study: – To study the effect of the choice of migration metric on the relative magnitude of health-worker migration from Africa to the US, UK, and other wealthy nations – To correlate magnitude of migration with source countries’ workforce capacities, health, health spending, economic and social (developmental) profiles – To investigate if African countries that lost more physicians also lost more nurses to the same destinations METHODS  Study design and data – Cross-country design using new database adjusted for bilateral net migration flows between African countries and: eight western countries and South Africa (circa 2002) – Migration data courtesy of the Center for Global Development  Study population – 53 Source countries: (North and Sub-Saharan) African countries – 9 Destination countries: US, UK, Canada, Australia, France, Belgium, Portugal, Spain, South Africa  Metrics – – Absolute number of émigrés Emigration fraction  number of émigrés ÷ (number of émigrés + number of health-workers left at home) number of émigrés per 100,000 source country’s population (Source: Arah et al, Am J Public Health 2007 – in press) – Health-worker migration density   Explanatory variables – – – Health workforce capacities: current densities Health indicators: infant and under-5 mortalities Health spending: total health spending and share of health spending from international aid – Economic and social progress: GDP per capita, poverty (percentage population living under I$1 PPP-a-day); female literacy rates; human development index (HDI)  Analysis – Ranking of source countries using the three migration metrics – Estimating correlations between migration metrics and source countries’ health workforce, health system, economic and social profiles – Correlations between nurse and physician migration metrics from Africa to same destinations: for US, UK and other seven countries separately and then combined – Multiple regressions modeling the metrics of migration (not presented due to space constraints) RESULTS •19% of African-born physicians work in UK, US, Canada, Australia, France, Portugal, Belgium, and Spain combined Australia Destination Countries 3 4 8 11 Canada France UK USA 0 5 6 36 23 20 23 25 30 35 40 36 10 15 20 Percentage of African-born Physicians in Nine Destination Countries % African Physicians Abroad % Sub-Saharan African Physicians Abroad • US and UK are primary destinations for African nurses, mostly from Sub-Saharan Africa Australia Destination countries 3 3 Canada France UK USA 0 3 3 8 25 30 30 5 10 15 20 25 30 35 37 38 40 45 Percentage African-born nurses in 5 out of 9 nine destination countries % African Nurses Abroad % Sub-Saharan African Nurses Abroad   12,813 African physicians in US; 52,128 in other destinations; 280,808 left at home Top five African physician sources for the US: Number of émigrés Egypt Nigeria South Africa Emigration fraction Liberia (53%) Gambia (35%) Ghana (29%) Physician migration density Egypt (5.6 per 100,000) South Africa (4.5) Ghana (4.4) Kenya Ghana Ethiopia (23%) Eritrea (20%) Liberia (3.6) Cape Verde (3.5) Source: Arah, BMC Public Health 2007 (in press)   20,983 African nurses in the US; 48,606 in other destinations; 758,698 left at home Top five African nurse sources for the US: Number of émigrés Nigeria Ghana Ethiopia South Africa Emigration fraction Liberia (77%) Gambia (52%) Eritrea (29%) Sierra Leone (23%) Nurse migration density Seychelles (35 per 100,000) Liberia (26.5) Cape Verde (21.3) Gambia (17) Liberia Guinea-Bissau (16%) Sierra Leone (15.8)  Correlations between physician migration metrics and African countries’ profiles Source countries’ profile Physician density Infant mortality Total health spending GDP per capita Number of émigrés 0.353* -0.213 0.189 0.140 Emigration fraction -0.352* 0.152 -0.072 -0.144 Physician migration density 0.694** -0.654** 0.583** 0.534** Human development 0.180 0.063 0.703**  Correlations between nurse migration metrics and African countries’ profiles Source countries’ profile Nurse density Infant mortality Total health spending GDP per capita Number of émigrés 0.192 -0.187 0.122 0.073 Emigration fraction -0.210 0.042 -0.138 -0.172 Nurse migration density 0.462** -0.512** 0.411** 0.375** Human development 0.177 0.055 0.578**  Correlations between nurse and physician migration to same destinations – to the US: 0.944 (P < 0.001) – to the UK: 0.850 (P < 0.001) – to other destinations: 0.849 (P < 0.001)  Correlations of physician migration between destinations – US versus UK: 0.679 (P < 0.001) – US versus others: 0.109 (n.s.) – UK versus others: 0.011 (n.s.)  Similar results for nurse migration SUMMARY Different metrics rank countries’ magnitude of migration differently  Migration density metrics have the most consistent correlations with profiles  Migration either has no recognizable development-related pattern contrary to what is often implied or seems to rise with developmental improvements  African countries which lose more physicians also lose relatively more nurses to the US and UK  POLICY RELEVANCE   Policies can be expected to be sensitive to the choice of migration metrics More attention to be paid to the less poor among source countries who appear to suffer more migration than is perhaps recognized (Arah, Med J Australia 2007 – in press)   Policies addressing health-worker migration and shortages in Africa must be integrated across the entire workforce since when nurses leave, physicians also leave or vice versa Paradigm shift from national to global health workforce policy-making, collaboration and alignment; US and UK can easily take lead Contact: Onyebuchi A. Arah, MD, DSc, MPH, PhD Department of Social Medicine Academic Medical Center, University of Amsterdam PO Box 22700, Amsterdam 1100 DE, Netherlands & UCLA School of Public Health Box 951772, Los Angeles, CA 90095-1772 E-mail: o.a.arah@amc.uva.nl arah@ucla.edu