Health screenings administered during the domestic medical examination of refugees and other eligible immigrants in nine US states, 2014–2016: A cross-sectional analysis

Background Refugees and other select visa holders are recommended to receive a domestic medical examination within 90 days after arrival to the United States. Limited data have been published on the coverage of screenings offered during this examination across multiple resettlement states, preventing evaluation of this voluntary program’s potential impact on postarrival refugee health. This analysis sought to calculate and compare screening proportions among refugees and other eligible populations to assess the domestic medical examination’s impact on screening coverage resulting from this examination. Methods and findings We conducted a cross-sectional analysis to summarize and compare domestic medical examination data from January 2014 to December 2016 from persons receiving a domestic medical examination in seven states (California, Colorado, Minnesota, New York, Kentucky, Illinois, and Texas); one county (Marion County, Indiana); and one academic medical center in Philadelphia, Pennsylvania. We analyzed screening coverage by sex, age, nationality, and country of last residence of persons and compared the proportions of persons receiving recommended screenings by those characteristics. We received data on disease screenings for 105,541 individuals who received a domestic medical examination; 47% were female and 51.5% were between the ages of 18 and 44. The proportions of people undergoing screening tests for infectious diseases were high, including for tuberculosis (91.6% screened), hepatitis B (95.8% screened), and human immunodeficiency virus (HIV; 80.3% screened). Screening rates for other health conditions were lower, including mental health (36.8% screened). The main limitation of our analysis was reliance on data that were collected primarily for programmatic rather than surveillance purposes. Conclusions In this analysis, we observed high rates of screening coverage for tuberculosis, hepatitis B, and HIV during the domestic medical examination and lower screening coverage for mental health. This analysis provided evidence that the domestic medical examination is an opportunity to ensure newly arrived refugees and other eligible populations receive recommended health screenings and are connected to the US healthcare system. We also identified knowledge gaps on how screenings are conducted for some conditions, notably mental health, identifying directions for future research.


Introduction
Between 2014 and 2016, the US welcomed an annual average of 75,000 refugees, 9,900 Special Immigrant Visa holders (SIVHs), and 23,000 persons granted asylum (asylees) [1]. Before US resettlement, these populations may have spent years without access to routine, preventive healthcare [2]. To address healthcare gaps and ensure healthy resettlement in the US, the immigration process includes multiple health screening exams and interventions [3].
Under section 325 of the Public Health Service Act (42 USC 252) and sections 212 and 232 of the Immigration and Nationality Act (8 USC 1182(8 USC , 1222, refugees, SIVHs, and derivative asylees (persons applying outside of the US) are required to undergo a medical exam before admission to the US to detect the presence of inadmissible health conditions. These medical exams are conducted by physicians contracted by the US Department of State and are performed according to US Centers for Disease Control and Prevention (CDC) Technical Instructions for the Medical Examination of Aliens [3]. In addition, US-bound refugees (but not other visa types) are eligible for overseas health interventions such as predeparture screening, the voluntary Vaccination Program for US-bound Refugees, and the presumptive parasite treatment program [3]. Asylees granted their status after US arrival do not receive overseas screening or interventions [4].
After US arrival or certification, refugees, SIVHs, asylees, Cuban/Haitian entrants (parolees), Amerasians (a visa status available to children born in Vietnam of US citizen fathers after January 1, 1962, and before January 1, 1976, and their eligible dependents), and certified victims of human trafficking are eligible for benefits, including a recommended medical screening examination that addresses health conditions diagnosed overseas, identifies additional health concerns, and connects the individuals with the US healthcare system [3,5]. CDC provides Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees, but state or local refugee health programs independently coordinate the examination, and these guidelines may be adapted for local use [6,7]. The comprehensive examination includes a medical history and physical examination, screening for select communicable diseases, nutritional status assessment, blood lead screening for children <17 years old, and immunization administration according to US schedules [7].
Examinations and interventions provided during the initial resettlement period ensure that refugees and other eligible populations are healthily integrated into receiving communities and are ready to begin work and school. Further, the domestic examination is usually the first contact the individual has with the US healthcare system and possibly the first time he/she is screened for certain health conditions or receives preventive healthcare [8,9].
Assessing the impact of the domestic examination on the recipient as well as on the receiving community has been challenging. Data from these examinations are not routinely collected at the national level, and state capacity to collect and analyze these data varies. In a review of the literature, we found that most analyses based on domestic medical examination data occur at the clinic, state, or local levels [10] and may not represent all recipients of refugee health program services, given the heterogeneity of these populations. Where domestic examination data are pooled across states or sites, analyses are generally limited to a particular condition (e.g., hepatitis B, lead) or population (e.g., children) [11][12][13][14]. Finally, we identified few analyses that reported the proportion of persons screened by condition during the domestic medical examination [10][11][12], which, to our knowledge, prevents assessment of the impact of the domestic examination on screening coverage. Accordingly, we conducted, to our knowledge, a novel analysis to assess screening coverage during the domestic medical examination on screening coverage for various conditions by calculating and comparing screening proportions among refugees and other visa holders by select demographic characteristics.

Methods
We conducted a retrospective, cross-sectional analysis of data derived from domestic medical examinations administered to refugees, SIVHs, asylees, parolees, and victims of trafficking between January 1, 2014, and December 31, 2016, in seven states (California, Colorado, Minnesota, New York, Kentucky, Illinois, and Texas); one county (Marion County, Indiana); and one academic medical center in Philadelphia, Pennsylvania. These sites provided their data as part of the CDC-funded CK12-1205 Strengthening Surveillance for Diseases among Newly Arrived Immigrants and Refugees nonresearch cooperative agreement administered by CDC. At least one representative of each site participated in the data collection and analysis of data. Although this is a convenience sample and not generalizable to all refugees in the US, the data set includes data from three states (California, Texas, and New York) with the highest volume of refugee, asylee, and SIVH arrivals between 2014 and 2016 and represents 28% of all refugee, asylee, and SIVH arrivals to the US during the same period. As part of a larger data collection effort, this analysis was not guided by a specific prospective analysis plan. However, the variables collected for analysis were outlined in a protocol used for ethical determination and shared with site partners contributing data. After data collection, we assessed the quality and availability of our data and identified appropriate analyses, including this descriptive analysis to assess screening coverage. This activity received a nonresearch determination from a CDC human subjects adviser and was exempt from further ethics review.
We suggested the variables and formats to compile the data to the nine participating site data partners. If partners could not provide those variables, they shared comparable raw data with explanations for missing variables or data. We then recoded the raw data for inclusion in the combined data set. Most data were collected from the state or local refugee health program databases, but some partners obtained data from other state or local programs (e.g., tuberculosis [TB] control or lead poisoning) because those data were not in the refugee program database. Some sites' restrictions on sharing personally identifying information prevented them from sharing birth and arrival dates; instead, they provided age at arrival and arrival year as proxy variables. Sites either provided nationality and country of last residence or shared a unique identifier we used to link to the US Department of State's Worldwide Refugee Admissions Processing System (78.1% of records linked) and CDC's Electronic Disease Notification system (74.0% linked) to obtain these two variables. Nationality or country of last residence were left blank or imputed from country of birth for records that did not match. Other than the unique identifier and birth date, we did not collect or store personally identifying information. Data were transmitted to CDC via secure file transfer and stored behind CDC's firewall and physical security measures; access to raw data was limited to a few CDC refugee health program staff.
Demographic variables were age (in years), sex, arrival year, visa status, nationality, and country of last residence (generally the country of asylum for refugees). For our outcome of interest, screening status, we created a bivariate variable of "screened," which included all persons a site reported to be screened, regardless of outcome, and "not screened," which included persons who were not screened; screening status was known for all persons. We examined results for conditions with screening recommended by CDC's Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees (Table 1) , blood lead levels (among children <17 years old), and mental health. For most conditions, we were unable to collect detailed information on the method of screening used by partners. We assumed the sites conducted screening or testing in accordance with the CDC's Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees (Table 1).
We used SAS 9.4 (SAS Institute, Cary, NC) to combine, clean, and standardize data from each site, resulting in an initial data set of 107,162 records meeting our inclusion criteria. A unique identifier was available for six sites (n = 90,209, 84.2%); data sharing laws in three sites (n = 16,953, 15.8%) prevented sharing of unique identifiers. We used the unique identifier to remove 1,621 (1.8%) duplicate records. Most duplicate records (n = 1,430, 88.2%) were from one site and related to a reporting error unlikely to have been repeated in other sites, in which the corrected rate of duplicate records was 0.2%. Therefore, we estimate the potential number of duplicate records remaining in the portion of data set without unique identifiers to be low (approximately 0.2%). We presented data individually for the top 10 countries of nationality and last residence by arrival volume and grouped the remaining countries into an "other" category. Denominators for each condition consisted of the total number of persons for whom screening status (e.g., screened or not screened) was available, which resulted in different denominators by condition. For each condition, we excluded data from sites that were only able to provide positive testing results without providing data on how many people were screened for the condition or who did not routinely screen or report data on a specific condition. The denominators used to calculate proportions screened are reported in "Total" column of each table. We calculated frequencies and percentages and their 95% confidence intervals (CIs) for demographic and screening variables and cross-tabulations for the number and proportion of persons screened with valid results for each condition by age, sex, visa type, examination year, countries of nationality, and last residence and for TB only, the screening test type. We used χ 2 and Fischer exact tests to compare proportions screened by age, sex, visa

Gonorrhea
Anyone with symptoms or leukocyte esterase positive on urine sample and women and children with history or risk of sexual assault Laboratory-confirmed testing (not specified)

Chlamydia
Sexually active women <25 years old, women >25 years old with risk factors, women or children with sexual assault history or risk, anyone with symptoms or are leukocyte esterase positive on urine sample Laboratory-confirmed testing (not specified)

Human Immunodeficiency Virus
All groups (opt-out screening) Laboratory-confirmed testing (not specified)

Blood Lead
All children 6 months to 16 years old Blood lead level from capillary or venous specimen (venous preferred)

Mental Health
Individuals >16 years old should be screened for symptoms of major depression and posttraumatic stress disorder RHS-15 or diagnosis or notation of mental condition including posttraumatic stress disorder, depression, anxiety, adjustment disorder, alcohol or substance abuse, and/or generalized anxiety disorder a This type, exam year, nationality, and country of last residence. We considered differences in proportions statistically significant at the 0.05 alpha level and examined 95% CIs to assess stratum-specific differences in screening by characteristic. This analysis is reported per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 STROBE Checklist).

Results
Our analysis data set included 105,541 refugees and other eligible persons who received a domestic medical examination between January 1, 2014, and December 31, 2016, in one of the nine sites. Characteristics of the data set are reported by visa status in Table 2. All nine sites provided screening data for TB, hepatitis B, HIV, and hepatitis C, resulting in a denominator of 105,541 persons screened for those conditions (Table 3). Eight sites provided screening data for malaria, schistosomiasis, strongyloidiasis, intestinal parasites, and sexually transmitted infections (STIs), resulting in a denominator of 100,677 persons screened for these infections (Tables 4 and 5). All nine sites provided blood lead screening data for children <17 years old, which is the age recommendation for lead screening in CDC's Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees (lead screening denominator: 35,118, Table 6). Finally, five sites provided mental health screening data for 27,712 persons (Table 6). Four sites did not share mental health screening data because they either did not conduct mental health screening at the time of the domestic screening examination or did not retain mental health screening information in the data sources used for this analysis. Approximately 51% of persons in our initial data set were between 18 and 44 years old, and 34.8% were under age 18 (Table 2). Fifty-three percent were male. Burma (16.8%), Iraq (15.8%), Cuba (14.7%), Afghanistan (11.1%), and Somalia (8.0%) contributed the largest volumes of arrivals by nationality ( Table 2). The data set mostly comprised refugees (70.2%), Cuban/Haitian parolees (14.1%), SIVHs (11.2%), and asylees (3.8%) ( Table 2). Fifty percent of examinations took place within 30 days of US arrival ( Table 2). Timing of screening after arrival varied by visa type (refugees and SIVH: median 27 days; parolees: 49 days; asylees: 57 days,     Col% refers to proportion of persons with characteristics using the total population (in first column) as denominator, and Row% refers to proportion of persons screened using stratum-specific (row) total as denominator.
Among the five sites that provided mental health screening data (denominator: 27,712 persons), screening was reported for 10,208 (36.8%, 95% CI 36.3-37.4) persons (Table 6). Mental health screening was low among children <16 years old (3.9% 95% CI 3.5-4.3, data collapsed from Table 6); 16 is the age at which mental health screening is recommended in CDC's Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees. Screening among Col% refers to proportion of persons with characteristics using the total population (in first column) as denominator, and Row% refers to proportion of persons screened using stratum-specific (row) total as denominator.
a Excludes data from one site that did not collect information on total numbers of persons screened for these conditions. b Excludes 10 individuals missing age.

Discussion
This analysis used a multistate data set to describe disease screenings conducted during the domestic medical examination for refugees and other eligible populations. These results provide new evidence that most persons eligible for refugee health benefits are screened for several health conditions after US arrival through the domestic medical examination, particularly for TB (91.6%), HBV (95.8%), HIV (80.3%), and blood lead (88.6% of persons <17 years old). Screening coverage for other health conditions, including STIs, parasites, and mental health, was lower. For most conditions, screening coverage increased with age and differed by nationality and country of last residence. Differences in proportions screened by sex and year of health assessment, although statistically significant, were not of clinical or public health significance for most conditions. To our knowledge, this is the first analysis of a large, multisite data set to assess screening coverage resulting from the domestic medical examination recommended for newly arrived refugees and other eligible visa holders in the US. Our data are consistent with prior studies that have reported high screening coverage for TB (93%) [16] and HBV (92.5%) during the domestic medical examination [11,13,16]. Compared with two studies reporting screening coverage for intestinal parasites among refugees in Minnesota (64%-80%) [16,17], our analysis found a lower screening coverage among refugees (50.2%). However, the Minnesota studies predate the current overseas presumptive parasite treatment program for US-bound refugees. Now, screening for intestinal parasites may be considered based on clinical symptoms and results of diagnostic testing (e.g., eosinophilia). Reports on screening coverage during the domestic medical examination for other conditions were lacking for comparison.
Our findings indicate that most persons undergoing the domestic medical examination receive several routine health screenings recommended in the US, resulting in higher screening coverage in this population compared with the broader US population for several conditions. In the US, TB and HBV screening are recommended for populations at highest risk of infections [18,19], including foreign-born persons from countries with moderate to high burden of the diseases [20,21]. In our analysis, 91.6% of persons who received a domestic medical examination were screened for TB compared with roughly 63% of all foreignborn respondents in the 2011-2012 National Health and Nutritional Examination Survey (self-reported data representative of persons residing in the US) [22]. Almost 96% of persons in our data set were screened for HBV compared with 19% of patients enrolled in four US healthcare organizations regardless of risk factors (N = 1,248,558) [23]. CDC recommends screening all adults for HIV at least once in their lifetime [24]. We found that 85% of adults were tested for HIV during the domestic screening examination, which is higher than the testing rate among the US adult population (<40% in 2016-2017) [24]. Finally, during the period of this analysis, one-time HCV screening was recommended for all US adults born between 1945 and 1965, regardless of risk factors [25,26]. The HCV screening rate among persons in our data set falling into this birth cohort (>44 years old at the time of their examination) was 56.1% (95% CI 55.3-56.9, data collapsed from Table 3), which is higher than the HCV screening rate among the general US population born between 1945 through 1965 (13% ever screened in 2015) [27]. The relatively high screening rates for TB, HBV, HIV, and HCV during the domestic examination demonstrates that the domestic examination is an effective opportunity to identify eligible new arrivals with these conditions and link them to care in the US. However, efforts to explore why screening opportunities for these conditions are missed during the domestic medical examination may further improve screening coverage. Screening coverage for malaria and intestinal parasites varied widely by visa type, nationality, and country of last residence. Screening coverage for malaria and intestinal parasites was lower among refugees than among other visa holders, which is likely the result of the overseas presumptive parasite treatment; if complete overseas treatment is documented, domestic screening is not required [28]. Ukrainians are likely overscreened for Strongyloides infection based on unclear guidance for screening for refugees originating from Europe, where Strongyloides is not hyperendemic [29]. Domestic screening is still recommended for persons who are symptomatic or have signs of infection such as eosinophilia, those who were ineligible for overseas treatment because of contraindications, and those from countries where presumptive treatment is not available.
The CDC Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees recommend screening for certain STIs based on whether overseas screening was done (for syphilis), age, and other risk factors, including history of sexual assault [7]. The proportions of refugees and SIVH screened for syphilis (53% and 58%, respectively) were lower compared with other visa types, likely reflecting the impact of overseas immigration screening requirements for syphilis. Domestic clinicians are recommended to review overseas syphilis testing results and to repeat testing in the US if overseas results are unavailable. For most of the period of this analysis, gonorrhea testing was not required overseas (overseas testing began October 1, 2016), and chlamydia testing has never been required overseas. For these infections, the domestic guidelines recommend screening for patients with certain risk factors (age, sexual history, or symptoms). Domestic gonorrhea and chlamydia screening among these populations could reflect presence of risk factors or symptoms, availability of overseas testing results, clinic screening policies for these conditions, or cultural barriers that discourage refugees from disclosing risk factors [30].
Blood lead screening was reported for 88.6% of all children <17 years old; children �2 years old had the lowest screening rates (82.0%). CDC's Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees recommend screening children aged 6 months through 16 years for lead upon arrival, but our data did not permit us to uniformly identify children younger than 1 year. Consequently, 82.0% is likely an underestimate of the proportion of children appropriately screened in this youngest age group. However, children aged 1-2 still had lower screening coverage compared with older children. Prior investigations have found a high prevalence of elevated blood lead levels among refugees [12,[31][32][33], and infants and young children are at greatest risk of both lead exposures and the resulting negative health outcomes [33]. Understanding and addressing barriers to lead screening among infants and young children is critical to ensuring appropriate care. Moreover, results from blood lead screening performed during the domestic medical exam serve as a baseline for the recommended repeat screening 3-6 months following resettlement that may highlight potential domestic exposures.
CDC's Guidelines for the U.S. Domestic Medical Examination for Newly Arriving Refugees encourage screening for depression and posttraumatic stress disorders (PTSDs) among people over age 16. In our analysis, mental health screening coverage among adults was between 50% and 55% and was lowest among refugees (31.8%) and SIVHs (30.8%). The relatively low rate of mental health screening could be the result of multiple factors. First, immigration-related screening has historically prioritized infectious disease screening in order to reduce disease risks to refugees and their resettlement communities. Second, mental health screening may be deferred from the domestic medical examination, which is typically administered within the first 90 days of US arrival, because refugees may be in a period of adjustment in which mental health symptoms are not as apparent [34,35], or they may not feel comfortable disclosing symptoms during the initial examination [7]. After building a rapport with clinicians, both care providers and resettled refugees may be better positioned to screen for and address mental health issues. Finally, research has found that screening is often informally administered, which may have made it difficult to capture in our data source, and screening may be less likely to be administered if culturally and linguistically competent mental health services are unavailable in the screening jurisdiction for referrals [36,37]. However, we were not able to ascertain why mental health screening was not done in this analysis. Meta-analyses of mental health among refugees and other conflict-affected populations have estimated prevalence of depression (>23%) and PTSDs (also >23%) [38,39] that exceed the estimated prevalence of these conditions among other populations in the US, such as veterans (PTSD 8%) and the US household population (depression 7.6%). The high prevalence of mental health conditions coupled with the low screening coverage in our analysis highlights the need for increased investigation of when and how mental health screening is being conducted among refugees and other populations receiving the domestic medical examination, barriers to mental health screening and services, and effective treatments for these populations.
This analysis was subject to limitations. The data upon which we relied were often collected primarily for programmatic rather than surveillance purposes, which yielded gaps in information (e.g., screening test type). Sites used screening tests and algorithms that may differ in sensitivity and specificity to detect conditions, and they collected and shared data differently. Therefore, our findings only reflect those sites contributing data and not all refugee and other persons receiving the domestic medical screening examination in the US. Some sites did not routinely collect information on each health condition included in the analysis, but this may not mean that they did not screen for these conditions; conversely, it also is possible that sites reported testing done outside the domestic medical examination, depending on their source of data and method of data extraction. One site captured only positive screening outcomes for parasitic infections and STIs without a denominator of all persons tested; we excluded these results from our calculations. Screening for some conditions is based on patient age, region of origin, or other characteristics (e.g., blood lead, malaria, STIs), so not all patients should have received screening for each health condition. Screening is done at the discretion of the examining clinician, with some healthcare providers being more likely to screen for certain conditions than others; and all screening tests can be declined by the recipients, which may be influenced by sex, age, cultural factors, and type of testing. We were not able to capture information on why an individual may not have been screened. Finally, we were unable to collect data on linkage to care in this population and could not evaluate how well the domestic medical screening examination connected people to the US healthcare system for further care.
This analysis found that refugees and other recipients of the domestic screening examination are well screened for several health conditions shortly after arrival in the US. This indicates that the domestic medical examination is effective at its intended purpose to screen newly resettled persons for health conditions, protecting both their health and the health of the receiving community. Future directions for research could include exploration of screening practices during the domestic medical examination, particularly mental health screening, and assessment of the domestic examination's effectiveness at connecting recipients to the US healthcare system.
Our assessment of a large, multisite data set found that screening coverage for several key health conditions-TB, hepatitis B and C, and HIV-was high during domestic screening examination, particularly when compared with screening rates among the US population. The domestic medical examination provides a good opportunity to improve the health of newly arrived refugees, SIVHs, asylees, and parolees by ensuring they receive several routine recommended health screenings.