Prevalence and Characteristics of Non–US-Born and US-Born Health Care Professionals, 2010-2018

IMPORTANCE Immigration to the US results in greater racial/ethnic diversity. However, the contribution of immigration to the diversity of the US health care professional (HCP) work force and its contribution to health care are poorly documented. OBJECTIVE To examine the sociodemographic characteristics and workforce outcomes of non–USborn and US-born HCPs. DESIGN, SETTING, AND PARTICIPANTS This cross-sectional study used national US Census Bureau data on US-born and non–US-born HCPs from the American Community Survey between 2010 and 2018. Demographic characteristics and occupational data for physicians, advanced practice registered nurses, physician assistants, registered nurses, licensed practical nurses or licensed vocational nurses, and other HCPs were included for analysis. Data were analyzed between December 2020 and February 2021. EXPOSURES Nativity status, defined as US-born HCP vs non–US-born HCP (further stratified by <10 years or 10 years of stay in the US). MAIN OUTCOMES AND MEASURES Annual hours worked, proportion of work done at night, residence in medically underserved areas and populations, and work in skilled nursing/home health settings. Inverse probability weighting of 3 nativity status groups was carried out using logistic regression. F test statistics were used to test across-group differences. Data were weighted using American Community Survey sampling weights. RESULTS Of a total of 657 455 HCPs analyzed (497 180 [75.5%] women; mean [SD] age, 43.7 [13.0] years; 518 317 [75.6%] White, 54 233 [10.8%] Black, and 60 680 [9.6%] Asian), non–US-born HCPs (105 331 in total) represented 17.3% (95% CI, 17.2%-17.4%) of HCPs between 2010 and 2018. They were older (mean [SD] age, 44.7 [11.6] years) and had more education (75 227 [70.1%] HCPs completed college) compared with US-born HCPs (mean [SD] age, 43.4 [13.3] years; 304 601 [55.2%] completed college). Nearly half of non–US-born HCPs (47 735 [43.0%]) were Asian. In major metropolitan areas, non–US-born HCPs represented 40% or more of all HCPs. Compared with US-born HCPs, non–US-born HCPs with less than 10 years and 10 or more years of stay worked 32.3 hours (95% CI, 19.2 to 45.4 hours) and 71.6 hours (95% CI, 65.1 to 78.2 hours) more per year, respectively. Compared with US-born HCPs, non–US-born HCPs were more likely to reside in areas with shortages of health care professionals (estimated percentage: <10 years, 75.3%; 10 years, 62.8% vs US-born, 8.3%) and work in home health settings (estimated percentage: <10 years, 17.5%; 10 years, 13.1% vs US-born, 12.8%). (continued) Key Points Question How large is the non–US-born health care professional (HCP) work force in the US, what occupations are non–US-born HCPs more likely to hold, and how do their work conditions compare with US-born HCPs? Findings In this cross-sectional study, 17.3% of HCPs were born outside the US. Overall, non–US-born HCPs worked more hours, were more likely to work at night and in skilled nursing and/or home health settings, and were more likely to reside in medically underserved areas than US-born HCPs. These associations differed by health care occupation and the length of stay in the US. Meaning These findings suggest that non–US-born HCPs are making significant contributions to health care in the US. Author affiliations and article information are listed at the end of this article. Open Access. This is an open access article distributed under the terms of the CC-BY License. JAMA Network Open. 2021;4(4):e218396. doi:10.1001/jamanetworkopen.2021.8396 (Reprinted) April 29, 2021 1/13 Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 08/20/2021 Abstract (continued)continued) CONCLUSIONS AND RELEVANCE The contributions of non–US-born HCPs to US health care are substantial and vary by profession. Greater efforts should be made to streamline their immigration process and to harmonize training and licensure requirements. JAMA Network Open. 2021;4(4):e218396. doi:10.1001/jamanetworkopen.2021.8396


Introduction
In 2018, more than 45 million immigrants lived in the US, accounting for approximately 14% of the population. 1 Immigrants were most commonly from Mexico (25%), followed by India (6%) and China (5%). 1 However, immigrants are increasingly coming from South Asia, Central America, and Africa in the past decade. 2 Health disparities among racial/ethnic minority groups in the US are well established. 3,4 Patientlevel factors (such as mistrust and treatment refusal), health system-level factors (eg, language barriers, access to health care), and structural social determinants related to segregation and poverty contribute to these disparities. 4,5 Patients who receive care from race-concordant health professionals rate their care as more participatory than patients who receive care from racediscordant health professionals. 6 Expanding and diversifying the health workforce may improve access to care and reduce health disparities. The ongoing COVID-19 pandemic has further amplified longstanding health disparities. 7 Throughout the pandemic, even as the disease spread from urban to rural areas, 8 racial and ethnic minority populations experienced a disproportionate disease burden. 9 As the US diversifies, the health workforce must adapt to meet the population's needs.
Immigrants help address health workforce shortages while furthering needed racial/ethnic diversity. 10 A more diverse workforce is associated with higher patient satisfaction owing to shared language, cultural sensitivity, and a better appreciation of patients' socioeconomic realities. 10,11 Diversity among health care professionals (HCPs) has not kept pace with demographic changes in the US population. In 2018, only 5% of physicians 12 and only 6.2% of nurses 13 were Black, compared with 13.4% of the US population. The shortage of Hispanic physicians has worsened over the past 30 years 14 ; only 6% of physicians in 2018 12 and only 5% of nurses in 2017 were Hispanic. 13 It is unclear to what extent immigrants constitute the current share of HCPs, and how they differ from US-born HCPs in employment outcomes. Thus, we examined the unique contributions of non-US-born HCPs by comparing their sociodemographic characteristics and employment outcomes with US-born HCPs to inform ongoing efforts to diversify the health workforce.

Design
The 2010 to 2018 American Community Survey (ACS) is an ongoing, nationally representative annual cross-sectional study of US-born and non-US-born adults residing in the 50 US states and the District of Columbia (excluding US territories). 15 The ACS, described in detail elsewhere, 15 obtains information on socioeconomic, housing, and demographic characteristics recorded on an online or paper survey from an invited sample of 1 in 38 US households. We extracted data for HCPs who participated in the survey. The initial sample included 718 507 HCPs whose primary job was in a health care delivery industry based on North American Standard Industry Classification System codes, 16 including hospitals, nursing and residential care facilities, outpatient care, and home and other health care services. Professional roles were identified based on self-reported occupation using 2010 US Census occupation codes. 17 Six groups of health care professionals were examined: registered nurses (RNs) with at least an associate degree; advanced practice registered nurses (APRNs; including nurse practitioners, certified nurse midwives, and certified registered nurse anesthetists with at least a master's degree); licensed practical nurses (LPNs) and licensed vocational nurses (LVNs); physician assistants (PAs); physicians; and other health care professionals (eg, therapists, dietitians, paramedics, acupuncturists, medical technicians, pharmacists, sonographers, optometrists). A total of 61 052 observations for individuals who had not worked in the past 12 months were excluded, leaving a sample of 657 455 HCPs. The Johns Hopkins School of Nursing institutional review board deemed this study exempt from approval because of the use of deidentified public use data files. Participants in the ACS provided informed consent prior to participating in the survey. 16 This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Measures Dependent Variables
Four dependent variables were used as indicators of potential contributions to US health care by non-US-born HCPs: annual hours worked, working at night, residence in medically underserved areas and populations (MUA/P), and employment in a skilled nursing, residential, institutional, or home health setting (henceforth, skilled nursing/home health setting). Annual hours worked were calculated by multiplying weeks worked by reported usual hours worked per week in the past 12 months. An individual was classified as working at night if they reported arriving at work between 6:00 PM and 1:59 AM. We identified MUA/P, geographic areas and populations with a lack of access to primary care services, 18 through 2020 county-level records obtained from the Health Resources and Services Administration, 19 which we then matched to the ACS data set based on county of residence using a crosswalk file from Integrated Public Use Microdata Series (IPUMS) USA. 20

Independent Variable
The chief exposure of interest was nativity status, classified as US-born (referent), non-US-born with under 10 years of stay in the US, and non-US-born with 10 years or more of stay in the US. Among non-US-born, the year of US entry was used to calculate the length of stay, which was dichotomized at 10 years because it is a commonly used indicator of acculturation among immigrants. 21 Countries of birth were grouped using a modified version of standardized world regions 22

Statistical Analysis
Descriptive statistics for the full sample were weighted using ACS sampling weights. Nativity status was the main independent variable. An F test was used to test for trends in the percentage of geographic location (ie, residence in metropolitan vs nonmetropolitan areas, US census region), and survey years. Inverse probability weighting of the 3 immigration groups was carried out using logistic regression. The weighted means of the dependent variables within each immigration group and their 95% confidence intervals were then estimated, and F test statistics tested across-group differences.
Significance levels were adjusted for multiple comparisons using the Benjamini-Hochberg procedure. 24 The false discovery rate was set at P < .05 using a 2-sided test. All the dependent variable means were estimated overall and stratified by health care occupation because we anticipated qualitative differences in outcomes based on profession. All statistical analyses were performed with Stata version 16 (StataCorp).

Non-US-Born Percentage of US Health Care Workforce
Rates of non-US-born HCPs differ by profession ( Table 2)
Differences across professions were found. Among non-US-born physicians, only those with less than 10 years of stay worked more (74.7 hours, 95% CI, 36.6-112.9 hours), whereas those with 10 or more years of stay were 9.8% less likely to reside in MUA/P (95% CI, −11.0% to −8.6%) and those with less than 10 years of stay were 0.3% less likely to work in skilled nursing/home health settings (95% CI, −0.5% to −0.1%) than US-born physicians. Non-US-born PAs with less than 10 (21.8%; 95% CI, 15.5%-28.1%) years of stay and APRNs with less than 10 (19.1%, 95% CI, 8.6%-29.7%) and 10 or more years of stay (5.3%; 95% CI, 1.6%-9.1%) were more likely to reside in MUA/P than their US-born peers, and APRNs with 10 or more years of stay were also more likely to work at night (1.3%; 95% CI, 0.2%-2.4%). Among RNs and LPN/LVNs, inferences were the same as those obtained in the total sample except that only non-US-born RNs with 10 or more years of stay worked more per year (67.2

Discussion
This analysis of national data offers fresh evidence that non-US-born HCPs (physicians, APRNs, PAs, RNs, LPNs/LVNs, and other HCPs) provide valuable skills and services in our health care workforce.
During the study period, 17.3% of all working HCPs were non-US-born. These estimates are slightly higher than the analyses of ACS data by Patterson et al, 25 which found that 15.7% of HCPs were non-US-born from 2011 to 2013. Non-US-born HCPs worked more hours per year, were more likely to work at night, reside in MUA/P, and work in the skilled nursing/home health setting than US-born HCPs. Also, since non-US-born HCPs also speak languages other than English, they possess language skills to improve the care of diverse patients. While 78.3% of US-born HCPs were women, only 68.3% of non-US-born HCPs were women, which may be explained by gender-based educational opportunities and skills-based migration policies that favor men. 26 The percentage of non-US-born   HCPs in major metropolitan areas in California, Texas, New York, and New Jersey was almost twice the percentage nationwide. Notably, these metropolitan areas were hard-hit during the peak of the COVID-19 pandemic in 2020. 8,27,28 Studies have shown that immigrants to the US tend to be highly educated. 1,29 Thirty-seven percent of non-US-born HCPs had at least a master's degree compared with 27.2% of US-born HCPs.
The Immigration and Naturalization Act of 1965 30 has contributed to the presence of highly educated immigrants who are accepted through temporary visa programs for high-skilled workers. However, "brain-waste," 31 severe waste of human capital resulting from the unemployment or underemployment of highly skilled college-educated immigrants, is common. The nonrecognition or discounting of foreign academic credentials and the costly processes for obtaining US-based credentials limit the effective integration of non-US-born HCPs. 31 Policies to eliminate brain waste among non-US-born HCPs may strengthen the health workforce.
While the proportion of non-US-born HCP increased over the study interval, there was a significant decline in the proportion of recently immigrated non-US-born HCPs (<10 years of stay).
This trend mirrors national estimates, showing that legal migration to the US declined 7.3% from 2016 to 2018. 32 Processing delays and policy changes, including "heightened screening and vetting" 33 processes have been blamed for the decline in employment-based immigration.
Non-US-born HCPs were more likely to come from Southeast Asia (24.6%) than other world regions. Most HCPs from Southeast Asia were RNs and born in the Philippines. The sizable presence of Filipino RNs is no accident. In addition to the Exchange Visitor Program, 34 the implementation of American nursing programs in the Philippines in 1898 resulted in an "Americanized" nursing curriculum taught in English. 35 Thus, Filipino nurses are actively recruited to fill the shortage of US nurses. 36 During the COVID-19 pandemic, about a third of nurses who have died from COVID-19 have been Filipino nurses, although they make up only 4% of the nursing workforce. 37 The Hispanic population constituted about 18% of the US population and surpassed 60 million in 2019, 38 yet they are underrepresented in US-born (5.9%) and non-US-born HCPs (15.5%).
Hispanic people are less likely to receive preventive health care or guideline-based health care than non-Hispanic people because of several factors, including low rates of health insurance and poor access to affordable care. 39 The COVID-19 pandemic has further amplified these longstanding health   Predicted margins for US-born vs non-US-born groups in US <10 years and Ն10 years were estimated using the margins command in STATA version 16; 95% CIs were estimated using the δ method. Significance levels were adjusted for multiple comparisons using the Benjamini-Hochberg procedure; all P values were considered significant at <.05 using a 5% false discovery rate and 8 contrasts.
b An individual was classified as working at night if they reported arriving at work between 6:00 PM and 1:59 AM. c MUA/P county-level records were obtained from the Health Resources and Services Administration 19 and matched to the American Community Survey data set based on county of residence using a crosswalk file from Integrated Public Use Microdata Series USA. 20