Post-migration HIV acquisition among African Immigrants in the U.S.

Abstract

Background:

African immigrants in the U.S. are more likely to have a late HIV diagnosis than U.S.-born people, potentially leading to onward transmission. We sought to determine the proportion of African-born people living with HIV (APLWH) who 1) had tested HIV negative prior to diagnosis, and 2) likely acquired HIV in the U.S.

Methods:

We interviewed APLWH from 2014–2017 and estimated the proportion with post-migration HIV acquisition based on clinical data, HIV testing history, immigration date, and behavioral data.

Results:

Of 179 participants, 113 (63%) were women. Less than half (44%) reported a negative HIV test prior to diagnosis. Among 142 (79%) participants with sufficient data to evaluate post-migration HIV acquisition, we estimate that 29% acquired HIV post-migration. Most APLWH acquire HIV prior to immigration.

Discussion:

Approximately one-quarter of APLWH acquire HIV post-migration and HIV testing is infrequent, highlighting the need for prevention efforts for African immigrants in the U.S.

Introduction

African immigrants are disproportionately affected by HIV in the U.S.,[1–3] in some cases experiencing rates much higher than that observed among U.S.-born Blacks.[3–5] This disparity is particularly pronounced among African-born Black women, who in 2016 experienced an HIV diagnosis rate that was almost 85-fold higher than that of White women in the U.S.; in comparison, this figure was 14.5 for US-born Black women compared to U.S. White women.[5] Although African immigrants tend to fare better than their U.S.-born counterparts in later stages of the HIV care continuum, with high levels of engagement in care and viral suppression[3, 6], they are more likely to present to care late in the course of infection, leading to significant HIV-related morbidity and possibly onward HIV transmission.[1, 3, 6–10]

The population of African-born persons with HIV (APLWH) is heterogeneous in terms of when and where they acquired and were diagnosed with HIV infection; this heterogeneity has important implications. Some people with HIV born outside the U.S. were diagnosed before immigrating to the U.S. However, the National HIV Surveillance System - which requires documentation of a prior HIV positive test to define a case as a “previous” diagnosis - is not well designed to capture information about which immigrants arrive in the U.S. with an established HIV diagnosis.[9, 11] A second group of immigrants with newly diagnosed HIV acquired before immigration, but were first diagnosed in the U.S.; ascertainment of pre-immigration HIV acquisition is not possible with routinely collected HIV surveillance or partner services data. Finally, a third group of immigrants acquires HIV in the United States. At present, the proportion of APLWH in each of these three groups is unknown.

The location of HIV acquisition and diagnosis among immigrant populations is important for two reasons: 1) Including previously diagnosed HIV cases in local and national HIV incidence estimates inflates HIV incidence, particularly in areas with large immigrant populations, and may contribute to estimated racial disparities, particularly insofar as data on US-born and immigrant Black population are conflated; [9] and 2) Failure to clearly define the place or timing of HIV acquisition and diagnosis inhibits effective public health programmatic work. Local health departments seek both to prevent HIV and to engage and retain people with HIV in care, but the strategies adopted to reach these goals are different for persons who acquire HIV before or after immigration. For immigrants who arrive in the U.S. with previously acquired HIV infection, the goal of public health efforts is to ensure that they test as soon as possible after arrival in the U.S., then successfully link to HIV care. In contrast, infections acquired post-migration are potentially avertable through increased condom use, PrEP, and frequent, repeated HIV testing after arrival in the U.S. and initial negative HIV testing. Therefore, taking a one-size-fits-all approach to HIV prevention – for example, focusing primarily on condom use, PrEP and routine HIV testing - may be less effective among immigrants who acquired HIV pre-migration, and therefore actually exacerbate existing disparities between immigrant and U.S.-born populations.

While several groups of European investigators have estimated the proportion of sub-Saharan African immigrants with HIV with postmigration acquisition, few studies have examined this question among any immigrant population in the U.S.[2, 12–16] We undertook the current study to evaluate previous HIV testing among APWLH in the U.S., and to test the use of an algorithm originally developed in the United Kingdom to calculate the degree of post-migration HIV acquisition among Black African migrants in that country.[12] Using this algorithm, we estimate the proportion of APLWH in the U.S. with post-migration HIV acquisition.

Methods

We recruited participants at four sites: King County, WA, New York City, NY, Chicago, IL, and Philadelphia, PA. We began recruitment in King County in February, 2014, followed by New York City in September, 2014; Chicago in September, 2015; and Philadelphia in April, 2016. All sites continued recruiting participants through January, 2018. Recruitment occurred primarily in HIV care clinics, health department clinics, and community-based organizations. Recruitment methods varied by site; in King County and Chicago recruitment occurred primarily via university (the University of Washington and Northwestern University) HIV care clinics. In King County, we also recruited participants through Public Health – Seattle and King County’s HIV partner services program, a high risk pregnancy clinic, a community-based organization that provides HIV case management, and via other research studies. In Philadelphia, participants were recruited through the Philadelphia Department of Public Health’s city health centers. In New York City, participants were recruited among people receiving HIV case management at a community-based organization. Eligibility criteria included birth in an African country, age 18 or greater, having an HIV diagnosis in 2000 or later, and the ability to provide informed consent. Initially we enrolled only participants who had been diagnosed with HIV in 2005 or later in an attempt to limit recall bias, but revised this in an attempt to increase enrollment (30 months after enrollment began). Language criteria varied by site; we based language criteria on the prevalence of languages spoken among potential participants at all four sites. In King County and New York City, we initially began recruiting only participants who spoke English, but after translating consent materials into five languages (Amharic, French, Tigrinya, Swahili, and Oromo) and working with the University of Washington Institutional Review Board to assure that procedures were in place to appropriately provide interpretation for participants who spoke other languages, we began recruiting participants regardless of language. In Philadelphia, we recruited English- and French-speaking participants. In Chicago, we recruited English-speaking participants only.

In King County, New York City, and Philadelphia, we interviewed participants in person, by telephone, or via an online survey using a shared web-based data collection instrument. Face to face interviews took place in recruitment settings, or in Philadelphia and King County, in participants’ homes or other settings chosen by the participant. To complete the online survey, which was available in English only, participants provided an email address and a link to the survey was emailed to them. We initially provided a $35 gift card as an incentive, but we increased this to $50 within a year of beginning recruitment in response to slower than anticipated enrollment.[17]

For our analysis of pre-/post-migration HIV acquisition, we utilized an algorithm originally developed by Paine et al., and revised by Burns, et al. in 2004, to perform similar analyses in the United Kingdom (Table 1) using a combination of interview and clinical information.[12, 18] We modified the original algorithm to better conform to our current understanding of the progression of HIV infection. We grouped participants into five categories: 1) Definite pre-migration HIV acquisition, 2) Probable pre-migration acquisition, 3) Probable post-migration acquisition, 4) Definite post-migration acquisition, and 5) Inadequate information/unknown location of acquisition (Table 1). The interview included questions regarding demographics, timing and sequence of immigration, immigration-related medical exams, access to medical care in the U.S., previous HIV testing, and a variety of risks for HIV, including sexual behaviors, drug use, exposure to blood, and both male and female circumcision. We included a group of questions regarding sexual violence developed for the National Violence Against Women Survey, and modified them for use with both genders.[19] We asked participants for authorization to access their medical, case management, or public health surveillance records to obtain a limited amount of clinical information, including date of diagnosis or establishment of HIV care, AIDS diagnosis ever, AIDS diagnosis date, any AIDS-defining conditions, and CD4 and viral load at the most recent medical appointment and at the time of diagnosis.

Table 1:

Algorithm for estimating post-migration HIV acquisition¹

Definitely acquired HIV outside the U.S.

Tested HIV positive before arrival

No unprotected sex in U.S. before testing HIV positive

In the U.S. for less than 45 days before HIV diagnosis

In the U.S. for less than 1 year before AIDS diagnosis

Had a child who was born with HIV before U.S. arrival

Probably acquired HIV outside the U.S.

Had a known positive partner prior to arrival in the U.S. and negative partners in the U.S.

In the U.S. for less than 2 years before diagnosis with AIDS

In the U.S. for less than 1 year before HIV diagnosis

Had sex with commercial sex workers prior to departure from Africa, no other suspected positive partners

Probably acquired HIV in the U.S.

Diagnosis in the U.S., known positive partner in the U.S., CD4 >=500 cells at diagnosis

In the U.S. for at least 12 years when diagnosed with AIDS (with no unprotected sexual activity during travel to Africa before diagnosis)

A negative test in the year before arrival in the U.S.

Definitely acquired HIV in the U.S.

No sex or no unprotected sex before living in the U.S.

Prior negative test in the U.S. with no overseas travel after

Inadequate information to determine location of acquisition

In the U.S. for less than 5 years when diagnosed with AIDS or CD4<200

¹Adapted from Burns FM, Arthur G, Johnson AM, Nazroo J, Fenton KA. United Kingdom acquisition of HIV infection in African residents in London: more than previously thought. AIDS. 2009;23(2):262–266.

Study procedures were approved by the University of Washington, Philadelphia Department of Public Health, and Northwestern University institutional review boards.

Results

Participant characteristics

We enrolled 179 participants: 74 (41%) in King County, 34 (19%) in New York City, 17 (10%) in Chicago, and 54 (30%) in Philadelphia (Table 2). Of those enrolled, 151 (84%) completed an in-person interview, 21 (12%) completed a phone interview, and 7 (4%) completed the online survey. Almost all participants (174, 97%) provided authorization to access their medical or case management records to obtain clinical data. The majority (113, 63%) of participants were women, and 136 (76.8%) were age 35 or older. While 52 (30%) participants had less than a high school education, a similar proportion (N=54, 31%) had a college or advanced degree. The largest share of participants (83, 49%) arrived in the U.S. via a visitor visa. The median number of years spent living in the U.S. prior to the interview was 8 years (range: 2 months-33 years). The most frequently reported region of birth was West Africa (83, 46%), followed by East Africa (72, 40%), Middle Africa (16, 9%), North Africa (4, 2%), and South Africa (3, 2%). Three-quarters of participants (135, 75%) completed an interview or online survey in English, while 29 (16%) and 8 (4%) completed the interview in French and Amharic, respectively. A smaller proportion (7, 5%) completed the interview in Tigrinya, Somali, Arabic, Oromo, and Kinyarwanda. Among the 110 (110/174, 63%) participants who reported being diagnosed with HIV in the U.S., median time in the U.S. before HIV diagnosis was 2 years (range: 1 month-31 years). Among seven participants who had lived in the U.S. less than 1 year before interview, HIV diagnosis occurred a median of 3 months after arrival, compared to a median of 7.6 years after arrival for people who had lived in the U.S. for 15 years or more (N=37). Reflecting heterogeneous African immigrant populations, participant characteristics varied by site; a substantial proportion of participants in New York City and Philadelphia were interviewed in French (21% and 33%, respectively, data not shown), while in King County, the predominant interview language outside of English was Amharic (11%). The majority of participants in Chicago were women (94%), compared to 57% in King County and 65% in Philadelphia.

Table 2:

Participant characteristics by study site, African-born people with a new HIV diagnosis, 2000–2017, King County, WA, New York City, Chicago, and Philadelphi^

	King County		New York City		Chicago		Philadelphia		Total
	N=74		N=34		N=17		N=54		N=179
	N	%	N	%	N	%	N	%	N	%
Gender
Women	42	56.8	20	58.8	16	94.1	35	64.8	113	63.1
Men	32	43.2	14	41.2	1	5.9	19	35.2	66	36.9
Age, years
18–25 years	1	1.4	2	5.9	1	5.9	5	9.4	9	5.1
26–35 years	13	17.8	7	20.6	5	29.4	7	13.2	32	18.1
36–45 years	30	41.1	15	44.1	8	47.1	15	28.3	68	38.4
45 years and older	29	39.7	10	29.4	3	17.7	26	49.1	68	38.4
Education
< High school	21	29.2	8	24.2	3	17.7	20	38.5	52	29.9
High school	16	22.2	8	24.2	6	35.3	9	17.3	39	22.4
Some college	10	13.9	4	12.1	4	23.5	6	11.5	24	13.8
College grad or higher	25	34.7	11	33.3	4	23.5	14	26.9	54	31.0
Other	0	0	2	6.1	0	0.0	3	5.8	5	2.9
Region of birth
West Africa	7	9.5	26	76.5	5	31.3	45	83.3	83	46.4
South Africa	1	1.4	1	2.9	1	6.3	0	0.0	3	1.7
East Africa	58	78.4	1	2.9	6	37.5	7	13.0	72	40.2
Middle Africa	5	6.8	6	17.7	3	18.8	2	3.7	16	8.9
North Africa	3	4.1	0	0.0	1	6.3	0	0.0	4	2.2
Visa status at immigration
Refugee	13	18.6	3	10.0	4	23.5	10	19.6	30	17.9
Asylee	10	14.3	2	6.7	1	5.9	1	2.0	14	8.3
	8	11.4	0	0.0	2	11.8	5	9.8	15	8.9
Legal permanent resident
Student	5	7.1	2	6.7	3	17.7	2	3.9	12	7.1
Visitor	27	38.6	22	73.3	3	17.7	31	60.8	83	48.5
Other	7	10.0	1	3.3	4	23.5	2	3.9	13	8.2
Year of immigration to U.S.
1980–1989	3	4.1	0	0.0	0	0.0	2	3.9	5	2.9
1990–1999	7	9.6	3	9.1	1	5.9	10	19.6	21	12.1
2000–2009	28	38.4	11	33.3	13	76.5	19	37.3	71	40.8
2010–2016	35	48.0	19	57.6	3	17.7	20	39.2	77	44.3
Year of HIV diagnosis
1994–2004	9	12.7	4	11.8	5	29.4	17	32.7	35	20.1
2005–2009	26	36.6	8	23.5	8	47.1	10	19.2	52	29.9
2010–2013	20	28.2	10	29.4	2	11.8	10	19.2	42	24.1
2014–2017	15	21.1	12	35.3	2	11.8	15	28.9	44	25.3

Prior HIV testing

Overall, 75/176 (44%) of participants reported having tested HIV negative at least once prior to HIV diagnosis; 5/176 (3%) participants reported not knowing if they had ever tested negative. Among 64 (37%) who were diagnosed with HIV outside the U.S., 23 (36%) reported a previous negative test, and among the 110 participants who were diagnosed with HIV in the U.S., 51 (46%) had ever tested negative prior to diagnosis. In this group diagnosed in the U.S., only 16/107 (15%) reported having had a negative HIV test result after arriving in the U.S. We identified predictors of having at least one negative HIV test before HIV diagnosis in a multivariate model. Of five covariates selected a priori, and in a model adjusted for gender and childhood family income, only education was associated with a previous negative HIV test result (Table 3, aPR for college or post-graduate education compared to less than high school 2.76, 95% CI:1.58–4.80).

Table 3:

Predictors of having a prior negative test among African-born people with a new HIV diagnosis, 2000–2017, King County, WA, New York City, Chicago, and Philadelphi^{^}

			Unadjusted model			Multivariate model
	N	%	Prevalence ratio	(95% confidence interval)		Prevalence ratio	(95% confidence interval)
Gender
Women	42	40.0	0.80	0.57	1.12	0.86	0.63	1.18
Men	33	50.0	1.00	-	-	1.00	-	-
Age at interview, years
18–25	2	25.0	0.62	0.18	2.14
26–35	16	51.6	1.28	0.81	2.02
36–45	27	41.5	1.09	0.73	1.64
46 and older	28	43.1	1.00	-	-
Education
Less than high school	29	33.3	1.00	-	-	1.00	-	-
High school or some college	8	34.8	1.59	0.91	2.77	1.57	0.86	2.89
College and post-graduate	37	68.5	2.79	1.71	4.55	2.76	1.58	4.80
Immigration year
Immigrated before 2010	36	47.4	1.06	0.79	1.42
Immigrated in 2010 or after	38	40.9	1.00	-	-
Childhood family income
Low income	19	38.0	1.00	-	-	1.00	-	-
Middle income	38	42.7	1.12	0.73	1.72	0.94	0.65	1.37
High income	13	59.1	1.56	0.95	2.55	1.10	0.71	1.69
Parenthood
Does not have children	16	41.0	1.00	-	-
Has children	59	44.7	1.09	0.71	1.66

^{^}Statistically significant results in bold font

Reason for HIV testing at diagnosis and potential routes of acquisition

We also asked participants about the reason they tested for HIV at the time of diagnosis; 65/168 (40%) reported they had tested because of symptoms. Participants diagnosed in the U.S. were more likely to report being diagnosed because they had sought care for symptoms than those diagnosed outside the U.S. (43%, and 33%, respectively, p=0.0002). Consistent with low levels of prior HIV testing and seeking care for symptoms at the time of diagnosis, among 103 participants with complete CD4 count at diagnosis, we found relatively low CD4 counts at the time of diagnosis (median CD4: 285, range: 0–1053). Among people diagnosed in the U.S., for whom we collected the most complete clinical data, 46 of 96 (48%) were diagnosed with stage 3 infection (AIDS) within 3 months of HIV diagnosis.

Participants reported a variety of potential types of HIV exposure. Because these data were somewhat more sensitive and thus more likely to be missing, in this paragraph we report both numerators and denominators for frequencies. Among 172 participants with complete data, 168 (98%) reported ever having sexual intercourse. Almost all participants with data on gender of sex partners reported having ever engaged in heterosexual sex (153/157, 97%). Among men, 9/60 (15%) reported having had oral or anal sex with another man. Sexual violence was reported by both genders; 21/106 (20%) women and 7/65 (10%) men reported sexual assault; of these, all but one reported sexual assault before HIV diagnosis. Twenty-nine (17%) participants reported working in a health care setting prior to diagnosis, and of these, 10 (34%) had been exposed to blood or bodily fluids. Among men, 50/62 (81%) had been circumcised in infancy or childhood; among women, 35/103 (34%) reported circumcision. Five participants (of 169, 3%) reported that they had engaged in exchange or transactional sex. Finally, 26 of 166 (15%) reported blood transfusions, and 44/170 (26%) had donated blood, prior to emigrating from Africa.

Acquisition of HIV pre- or post-migration

We categorized 92 (51%) participants as definitely having pre-migration HIV acquisition, and 29 (16.2%) as definitely having post-migration (U.S.) acquisition (Table 4). Of the 92 with definite pre-migration acquisition, 64 (37% of total participants) were diagnosed with HIV prior to immigration, and 25 (14% of total participants) after arrival in the U.S. Our algorithm did not allow us to estimate location of HIV acquisition for 37 participants with inadequate information to make this estimation (21%, Table 3). If we exclude those 37, 20% (29/142) of participants were categorized as having definite, and 9% as having probable, post-migration acquisition. Similarly, excluding those with inadequate information, 101/142 (71%) had definite or probable pre-migration acquisition. Of these, 64 (63%) were diagnosed prior to immigration.

Table 4:

Estimated location of HIV acquisition, African-born people with a new HIV diagnosis, 2000–2017, King County, WA, New York City, Chicago, and Philadelphia^*

	Definitely Africa	Probably Africa	Probably U.S.	Definitely U.S.	Inadequate information	Total
All participants (N, %)	92 (51.4)	9 (5.0)	12 (6.7)	29 (16.2)	37 (20.7)	179 (100)
Percent in each category excluding those with inadequate information	64.8	6.3	8.5	20.4	--
Diagnosed in the U.S. (%)	25 (28.1)	9 (100)	12(100)	28 (100)	36 (100)	110 (63.2)
Median time in U.S. to diagnosis^ (years, range)	0.5 (0.1 –7.6)	1.4 (0.50–10.0)	6.6 (0.75–15.1)	5.0 (0.50–30.6)	3.2 (0.17–22.3)	2.6 (0.7–30.6)
Median CD4 at diagnosis* (range)	135 (13–1053)	89 (26–303)	274 (0–627)	394 (20–656)	321 (4–818)	293 (0–1053)
Concurrent HIV and AIDS diagnosi^ (%)	12 (54.6)	8 (100)	5 (45.5)	6 (25.0)	11 (40.7)	42 (45.6)
Had ≥1 negative HIV test in the U.S.^ (%)	0 (0)	0 (0)	2 (20.0)	13 (44.8)	2 (5.7)	0 (0.0)

^*While participants were newly diagnosed in the U.S. from 2000–2017, in some cases they reported positive tests outside of the U.S. prior to 2000.

^{^}Among those diagnosed in the U.S.

Of note, median CD4 count at diagnosis was highest among those who were categorized as having definite post-migration acquisition, and concurrent HIV and AIDS diagnosis was least frequent in this group (Table 4). Median time spent in the U.S. before diagnosis was relatively short in those participants categorized with definite or probable pre-immigration acquisition (0.5 and 1.4 years, respectively, 7 months for both groups combined), compared to those with post-migration acquisition (6.6 and 5 years among those in the probable and definite post-migration acquisition groups, respectively). Only four of 34 (12%) participants who were categorized with pre-migration acquisition and were diagnosed in the U.S. were diagnosed after being in the U.S. for 2 years or longer; in this group, median time to diagnosis was 6.4 years. Forty-five percent (13/28) in the definite post-migration acquisition group reported ever having a negative HIV test in the U.S.

Predictors of post-migration HIV acquisition

When we examined predictors of post-migration acquisition, arrival in the U.S. prior to 2010 was the strongest predictor of U.S. acquisition (Table 5, aPR:7.40, 95% CI:2.67–20.5) in an adjusted model. Women were less likely than men to have acquired HIV in the U.S. (aPR:0.23 95% CI:0.08–0.68). Those ages 26–35 (aPR:4.39, 95% CI:1.86–10.4) and 36–45 (aPR:2.17, 95% CI:1.09–3.96) vs. participants aged 46 and older were more likely to have post-migration acquisition. Our adjusted multivariate model for predictors of post-migration HIV acquisition included an interaction between sex and parenthood; compared to women without children, women with children were almost 4 times more likely to have acquired HIV in the U.S. (Table 5).

Table 5:

Predictors of HIV acquisition in the U.S. among African-born people with a new HIV diagnosis, 2000–2017, King County, WA, New York City, Chicago, and Philadelphi^{^}

			Unadjusted model			Multivariate model
	N	%	Prevalence ratio	(95% confidence interval)		Prevalence ratio	(95% confidence interval)
Gender
Women	25	22.1	0.91	0.52	1.57	0.23	0.08	0.68
Men	16	24.2	1.00	-	-	1.00	-	-
Age at interview, years
18–25	1	11.1	0.84	0.12	5.87	5.64	0.97	32.64
26–35	8	25.9	1.89	0.80	4.44	4.39	1.86	10.39
36–45	23	33.8	2.56	1.28	5.11	2.17	1.09	3.96
46 and older	9	13.2	1.00	-	-	1.00	1.00	1.00
Education
Less than high school	10	21.1	1.00	-	-	1.00	-	-
High school or some college	18	28.6	1.36	0.72	2.56	0.75	0.40	1.41
College and post-graduate	11	20.4	0.99	0.48	2.04	0.62	0.31	1.25
Immigration year
Immigrated before 2010	6	7.8	4.62	2.05	10.40	7.40	2.67	20.46
Immigrated in 2010 or after	35	36.1	1.00	-	-	1.00	-	-
Childhood family income
Low income	7	13.5	1.00	-	-	1.00	-	-
Middle income	22	23.9	1.74	0.80	3.80	1.27	0.57	2.89
High income	10	41.7	3.04	1.32	7.00	2.69	0.98	7.35
Parenthood
Does not have children	10	23.8	1.00	-	-	1.00	-	-
Has children	31	22.6	0.96	0.52	1.80	0.67	0.32	1.43
Interaction term
Gender and parenthood			5.56*	1.13	27.5	3.89*	1.10	13.79

^*PR for interaction term indicates the additional increase in prevalence of local acquisition among women who have children.

^{^}Statistically significant results in bold font

Discussion

We estimate that among African immigrants living with HIV across four geographically diverse U.S. areas, 44% acquired and were diagnosed with HIV prior to immigration, 25% acquired HIV pre-immigration but were diagnosed in the U.S., and 29% acquired HIV post-migration. Among those who we believe acquired HIV pre-migration and were diagnosed in the U.S., median time in the U.S. before diagnosis was 10 months, indicating most people who acquire HIV outside the U.S. are being diagnosed relatively quickly after arrival, helping to limit onward HIV transmission. However, a small percentage (16%) of participants who likely acquired HIV pre-migration were not diagnosed for 2 years or more after immigration. Among participants who were diagnosed after arrival in the U.S, we found that previous HIV testing was infrequent, both in the U.S. and prior to immigration to the U.S., and that only level of education was associated with a higher likelihood of having had a negative HIV test result prior to diagnosis.

Several groups of investigators in Western Europe have used a variety of methodologies to estimate the prevalence of post-migration HIV acquisition among Sub-Saharan African migrants; estimates have varied substantially, from 12–19% in Sweden to 35–49% in France.[12, 13, 16, 20–22] These differences likely reflect differences in populations, time periods studied, or the methodology used. Studies which examined post-migration acquisition by transmission risk typically found that post-migration acquisition was higher among MSM and people who inject drugs than among heterosexuals, or among men than women.[12–14, 16] To our knowledge, post-migration acquisition among African immigrants has been examined previously in only two studies in the U.S. Both were population-based and conducted using HIV surveillance and partner services data; these studies found that 26 and 34% of African immigrants newly diagnosed with HIV in King County, WA and New York City, respectively, likely acquired HIV post-migration.[2, 10] These results are similar to our estimate of post-migration acquisition of 28%. Similar to heterogeneity found in estimates of post-migration acquisition from Europe, differences in these U.S. estimates may reflect true differences related to local African immigrant population characteristics or HIV prevalence in the study sites, or may be related to the methodology used.

The level of pre- and post-migration acquisition among immigrants in local jurisdictions may have a sizeable impact on calculations of local HIV incidence, as well as the proportion of new cases with delayed diagnoses. Incidence is an important indicator of the success of public health work to prevent HIV, and impacts HIV funding levels for local jurisdictions. As such, further work is needed to determine how people who arrive in the U.S. and report that they have previously tested HIV positive but have no documentation of test results should be treated for HIV surveillance purposes. Perhaps there should be a third category of new diagnoses – newly reported in the U.S., but diagnosed previously outside the U.S. – to allow jurisdictions to track true incidence and more accurately evaluate HIV prevention efforts. These data would additionally allow a more accurate calculation of delayed HIV diagnoses, a key indicator for measuring the success of local HIV testing efforts. Immigrants from all regions of birth are more likely to have a delayed diagnosis than people born in the U.S.;[4, 6, 8, 23–27] this may reflect to some degree infections acquired years prior to immigration. A critical first step in this process is identifying cases that occur among immigrants. In analyses of national HIV surveillance data, approximately 18% of new HIV diagnoses were reported to CDC missing country of birth, indicating that improvement in simply ascertaining nativity is needed in some local jurisdictions.[7, 8, 25]

There were several limitations to our study. First, we primarily recruited APLWH who were in HIV care. If people who are out of care were more or less likely to acquire HIV in the U.S., our estimate of post-migration acquisition will over- or under-estimate the true proportion of people with post-migration HIV acquisition. While we recruited APLWH from four geographically diverse metropolitan areas, study participants are likely not representative of all African immigrants in the U.S. It is also important to note that our results are not generalizable to other immigrant groups; the level of post-migration acquisition likely varies widely, based on factors such as the HIV prevalence in the sending country, and transmission risk category. We recruited fewer participants than planned,[17] and as a result we are unable to calculate estimates of post-migration acquisition by study site or region/country of birth for participants. This also further limits our ability to generalize our results to all African immigrants in the U.S. In particular, while the gender distribution of our study sample is similar to that of all APLWH diagnosed from 2010–2017 in the U.S., the proportion of men reporting male to male sex in the study population is lower than that among all male APLWH.[25] We did not ask male APLWH about anal and oral sex with other men as separate behaviors, limiting our ability to examine these behaviors, which carry different transmission risks, individually. In some cases, participants were unable to remember dates or provide information about previous negative HIV tests. There were also participants who were simply difficult to categorize into pre- and post-migration acquisition groups because they lacked the “landmark” events that we used in our algorithm, such as a previous negative test in the U.S. However, the clinical values and previous HIV testing history information that we have for this group fall between those of the groups that we believe acquired HIV pre- and post-migration. Thus, we believe that the participants in this group were made up of some with pre-migration acquisition, and some with post-migration acquisition. Finally, while we report time to HIV diagnosis in the U.S for people were diagnosed in the U.S., this estimate is strongly related to the amount of time people have spent in the U.S.

Our results have key implications for the current effort to end the HIV epidemic in the U.S. First, in order to monitor and evaluate progress toward ending the epidemic, accurate incidence data is vital. Tools to determine the proportion of new HIV diagnoses that were acquired outside of the U.S. would allow local jurisdictions to assess the effectiveness of HIV prevention efforts more accurately. Secondly, ending the epidemic will require a laser focus on marginalized populations. While our results indicate that only about a quarter of new HIV diagnoses among African immigrants were acquired in the U.S., a proportion of those who acquired HIV pre-migration were ultimately diagnosed in the U.S., and all study participants require HIV care. Recent policies designed to limit immigration, including those which deter immigrants from using public services such as health care, could make immigrant communities more difficult to reach with HIV prevention and treatment programs.[28, 29] Policies such as these, which lead to the further marginalization of a key population, will also impede our ability to end the epidemic in the U.S.