Abstract
Background:
Migrants in sub-Saharan Africa are at increased risk of HIV acquisition following migration, but little is known about their sexual partners at place of destination.
Setting:
Rakai Community Cohort Study (RCCS) in Uganda.
Methods:
From 1999 through 2016, persons aged 15–49 were surveyed in the RCCS and reported on their four most recent sexual partners in the last year. We compared the characteristics of sexual partners reported by migrants moving into RCCS communities in the last 2 years (i.e., in-migrants) to those of long-term residents with no recent migration history. Among a subset of participants in cohabitating epidemiologically-linked couples of known HIV serostatus, we also assessed prevalence of having ≥1 untreated HIV-positive partner among in-migrants and long-term residents.
Results:
116,744 sexual partners were reported by 29,423 participants. The sexual partnerships of in-migrants were significantly less likely to be marital, more likely to span community boundaries, and be shorter in duration than those of long-term residents. In-migrants also reported more sexual partners and were less likely to know their partner’s HIV status or to have told their partner their HIV status. Among 7,558 epidemiologically-linked couples, HIV-negative in-migrants were more likely to partner with untreated HIV-positive persons compared to HIV-negative long-term residents (women: 6.3 vs. 4.1%; prevalence risk ratio [PRR]=1.77, 95% CI: 1.49–2.11; men: 6.9 vs. 3.9%; PRR=1.72, 95% CI: 1.38–2.14).
Conclusion:
There is a higher frequency of risky sexual behaviors among the partnerships of in-migrants compared to long-term residents. Among cohabitating couples, in-migrants are more likely to partner with untreated HIV-positive individuals.
Keywords: HIV, epidemiology, migration, mobility, networks, Uganda, antiretroviral therapy
INTRODUCTION
Migration has been associated with the spread of the human immunodeficiency virus (HIV) since the onset of the HIV pandemic more than four decades ago. 1–3 It continues to pose serious challenges to disease control programs globally,4–6 including within sub-Saharan Africa, where the global burden of HIV is concentrated.7 Sub-Saharan African migrating populations are more likely to be HIV-positive,8–10 are less likely to be taking antiretroviral therapy (ART), 4,5 and experience greater AIDS-related mortality compared to non-migrants.11–13 They also tend to have increased propensity for HIV-related risk behaviors and are at higher risk of HIV acquisition compared to their non-migrating sexually-active peers.14–16 Despite recent declines in HIV incidence and substantial increases in ART coverage across the African continent,17,18 disparities in HIV risk and viral load suppression persist by migration status.19,20
One hypothesized mechanism for increased HIV risk among migrant populations is the destabilization of personal sexual networks as a result of the migration process. 21 While sexual network disruption, characterized by partnership turnover resulting in the formation of new relationships, has been postulated to be causally linked to HIV risk among migrants, there is little empirical data on migrant sexual partnerships or the prevalence of untreated HIV within migrants’ sexual partner pools in comparison to non-migrating persons. A better understanding of the casual mechanisms that drive higher HIV incidence among migrants may result in more effective, targeted HIV services to these populations, facilitating HIV control within them and the general population more broadly.
Here, we used multiple data sources from the Rakai Community Cohort Study (RCCS), an open population-based cohort in Uganda, to assess the sexual partner networks of recently migrating persons. Specifically, we compared the self-reported sexual partnership characteristics of individuals who migrated into RCCS study communities within the prior two years (i.e., recent in-migrants) to those of long-term residents in the same communities with no recent history of migration. Using data from epidemiologically-linked cohabitating couples of known HIV serostatus, we also measured the proportion of migrating and long-term resident participants with an HIV-positive partner, including HIV-positive partners with no self-reported ART use (i.e., untreated HIV). We hypothesized that there would be higher prevalence of HIV-related risk factors and untreated HIV infection within the sexual partnerships of recent in-migrants as compared to those of long-term residents.
METHODS
Study population and data collection procedures
The Rakai Community Cohort Study (RCCS) is an open population-based cohort of persons aged 15–49 years living in 40 communities in and near the Rakai district in southcentral Uganda. The RCCS was initiated in 1994 by the Rakai Health Sciences Program (RHSP), and is currently ongoing.22 At approximately 18-month intervals, a household census enumerates all residents in study communities and collects data on in- and out-migrations, births, and deaths that have occurred since the previous census. Recent in-migrants, herein simply referred to as in-migrants, are classified as persons who migrate into a study community between survey rounds. In-migrants carry this designation until the next survey round, at which time the participants become known as long-term residents. Long-term residents also include individuals who have resided in study communities since birth.
Following the census, eligible consenting participants are interviewed regarding their sociodemographic characteristics, sexual behaviors, health status, and health-seeking behaviors, including male circumcision status and current ART use. Self-reported ART use has been previously validated in the RCCS using high performance liquid chromatography with 77% sensitivity and 99% specificity.23 Participants are also interviewed about their four most recent sexual partners in the last year (i.e., egocentric partner block data). Detailed information is collected on the nature of these partnerships, such as relationship type (current husband/wife, current cohabitating non-marital partner, boyfriend/girlfriend, occasional/casual partner), partnership duration, condom and alcohol use within partnerships, as well as individual-level characteristics of the partner, including their occupation and age. Additionally, the RCCS collects the names of cohabitating sexual partners during the census and survey. When a named partner also participates in the RCCS, the couple’s data are epidemiologically-linked irrespective of whether one or both partners name each other. All consenting participants provide venous blood samples, which are tested for HIV (see laboratory methods).
For this study, we used census and survey interview data collected from participants in 30 continuously surveyed RCCS communities between April 6, 1999 and September 2, 2016. This period included 12 survey rounds (see Table S1, Supplemental Digital Content). Participation rates among eligible individuals ranged from 59% to 66% between surveys.22 Participation was lower among men, younger residents, and those living in semi-urban trading communities (as compared to participants living in rural agrarian communities).22
Ethical Considerations
The RCCS was approved by the Uganda National Council for Science and Technology, the Uganda Virus Research Institute’s Research and Ethics Committee, and the Western Institutional Review Board (Olympia, WA) and was conducted in accordance with the Helsinki Declaration of 1975 (as revised in 2000).
Scale-up of combination HIV prevention
The RCCS’s combination HIV prevention (CHP) strategy consists of HIV testing and counseling, ART, and voluntary medical male circumcision (VMMC) services. ART was introduced to Rakai in 2004 with funding from the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR). From 2003 through 2006, RHSP conducted a large randomized controlled trial to assess the efficacy of VMMC.24 Following the conclusion of the trial, RHSP provided free VMMC services to trial controls and since 2007, with PEPFAR support, RHSP has provided free VMMC services to males ages 13 and older throughout the district.
For this study, we stratified time in relation to the scale-up of PEPFAR VMMC and ART programs. Survey rounds 1 through 5 (1999–2004) were denoted as the pre-CHP period. Survey rounds 6 through 9 (2005–2011) were denoted as early CHP and survey rounds 10 through 12 (2011–2016) as late CHP (see Table S1, Supplemental Digital Content).
Laboratory methods
Participants provide venous blood samples for HIV testing. Prior to October 2011, enzyme immunoassays (EIAs) were used to test for HIV, with confirmation by Western blot or polymerase chain reaction (PCR).25 In October 2011, a parallel three test rapid testing algorithm was introduced with confirmation by two EIAs (Vironostika HIV-1, BioMerieux, and Recombigen, Cambridge Biotech).26 If EIA results were discordant, results were confirmed with Western blot (GS HIV-1 Western Blot, Bio-Rad Laboratories, Redmond, WA, USA, BioMerieux-Vitek),27 or by PCR as required by the Ugandan HIV Testing Services Policy.28
Statistical analysis
Guided by a theoretical framework (see Figure S1, Supplemental Digital Content), we used RCCS egocentric partner block data to compare the self-reported characteristics of the sexual partners of in-migrants to those of long-term residents. Individuals may have been counted as partners more than once if multiple participants partnered with the same individual or if a participant reported partnership with the same individual at multiple survey rounds. The characteristics of index RCCS participants and their partners included partner age, age difference, partner primary occupation, relationship type, relationship length, whether the relationship was ongoing at time of survey, partner residence in index household/community, male partner circumcision status, condom use, index and partner alcohol use before sex, and knowledge of index’s and partner’s HIV status (see Tables S2–S6, Supplemental Digital Content). Analyses of partnership data were stratified by the index participant’s residence status (i.e., in-migrant vs. long-term resident), sex, and HIV serostatus because prior studies in this same population have suggested effect modification by these factors.20,27 Unadjusted log-binomial regression and Mann-Whitney U tests were used to compare categorical and continuous variables by residence status, respectively. Comparisons for categorical data were reported as prevalence risk ratios (PRR) with 95% confidence intervals (CI).
Among epidemiologically-linked cohabitating couples where the HIV serostatus of both partners was confirmed through laboratory testing in the RCCS, we estimated HIV burden among the sexual partners of in-migrants and long-term residents. Here, the primary outcome was sexual partnership with ≥1 HIV-positive partner. Log-binomial regression with generalized estimating equations was used to estimate the relative risk of the primary outcome by residence status. Multiple within-participant correlation structures, including independent, exchangeable, and first-order autoregressive, were evaluated. The final working correlation structure (independent) was selected using quasi-likelihood information criterion.29 Three separate analyses were conducted. First, data were analyzed irrespective of the index participant’s HIV status, and second restricting to HIV-negative participants only. Third, we assessed risk of partnership with ≥1 untreated HIV-positive partners among HIV-negative participants only, where untreated HIV-positive partners were defined as those who were HIV-positive and did not self-report current ART use. All analyses were stratified by residence status and sex of the index participant as well as calendar period (i.e., pre-CHP, early CHP, and late CHP). Multivariate analyses were adjusted for the index participant’s age and marital status.
RESULTS
In total, 116,744 sexual partners were reported by 29,423 RCCS participants (Table 1). Of these partners, 56,192 (48%) were reported by female participants (n=16,911), and 60,552 (52%) were reported by male participants (n=12,512). HIV-negative participants (n=25,575) reported 100,132 (86%) partners, and HIV-positive participants (n=4,956) reported the remaining 16,612 (14%). In-migrants (n=11,385) reported 15,570 (13%) partners, and long-term residents (n=23,809) reported 101,174 (87%) partners. At baseline, both male and female in-migrants were more likely to be married, HIV-positive and untreated, and report more sexual partners on the egocentric partner blocks compared to long-term residents (see Table S7, Supplemental Digital Content).
Table 1.
Relationship characteristics of the self-reported sexual partnerships of RCCS participants stratified by participants’ residence status and sex, 1999–2016
| Self-reported male partners (n=56,192) | Self-reported female partners (n=60,552) | |||||
|---|---|---|---|---|---|---|
| Long-term resident women (n=46,870) | In-migrant women (n=9,322) | p-value or PRR (95% CI) | Long-term resident men (n=54,304) | In-migrant men (n=6,248) | p-value or PRR (95% CI) | |
| Relationship type | n=46870 | n=9322 | n=54304 | n=6248 | ||
| Current husband/wife | 19807 (42%) | 2166 (23%) | 0.55 (0.53, 0.57) | 15902 (29%) | 1118 (18%) | 0.61 (0.58, 0.65) |
| Current cohabitating partner (non-marital) | 11596 (25%) | 3593 (39%) | 1.56 (1.51, 1.61) | 10361 (19%) | 1374 (22%) | 1.15 (1.10, 1.21) |
| Boyfriend/girlfriend | 14602 (31%) | 3279 (35%) | 1.13 (1.09, 1.16) | 23287 (43%) | 2993 (48%) | 1.12 (1.09, 1.15) |
| Occasional/casual partner | 398 (0.85%) | 168 (1.8%) | 2.12 (1.77, 2.53) | 3588 (6.6%) | 620 (9.9%) | 1.50 (1.38, 1.63) |
| Other | 422 (0.90%) | 99 (1.1%) | 1.18 (0.94, 1.46) | 1069 (2.0%) | 132 (2.1%) | 1.07 (0.89, 1.28) |
| Length of relationship in years | n=46512 | n=9188 | n=51602 | n=5919 | ||
| Median (IQR) | 6.0 (3.0, 13) | 1.0 (0.82, 4.0) | <0.0001 | 2.0 (0.66, 8.0) | 1.0 (0.49, 4.0) | <0.0001 |
| Relationship ongoing | n=46870 | n=9322 | n=54304 | n=6248 | ||
| Yes | 39461 (84%) | 7433 (80%) | 0.95 (0.94, 0.96) | 38338 (71%) | 3903 (62%) | 0.88 (0.87, 0.90) |
| No | 6715 (14%) | 1738 (19%) | 1.30 (1.24, 1.36) | 13198 (24%) | 2020 (32%) | 1.33 (1.28, 1.38) |
| Don’t know | 495 (1.1%) | 32 (0.34%) | 0.33 (0.22, 0.46) | 161 (0.30%) | 10 (0.16%) | 0.54 (0.27, 0.97) |
Data are n (%) unless otherwise noted. Denominators vary for characteristics due to different survey rounds in which data were obtained (see Table S2, Supplemental Digital Content) and include sexual partners for whom data is missing. Some column percentages do not add to 100 due to rounding and missing data. RCCS = Rakai Community Cohort Study. PRR = Prevalence risk ratio with the reference group defined as long-term resident RCCS participants. CI=Confidence interval.
The sexual partnership characteristics of in-migrants and long-term residents
The sexual partners of in-migrant men and women were more likely to be younger compared to partners reported by long-term residents, although age difference between partners did not significantly differ by residence status (see Table S8, Supplemental Digital Content). Partners of in-migrants were also less likely to live in the same household and more likely to be from outside their community of current residence (23% vs. 19% for male partners reported by female participants; 29% vs. 25% for female partners reported by male participants). Additionally, in-migrants’ partners were significantly less likely to be marital partners compared to long-term residents’ partners (23% vs. 42% for male partners; 18% vs. 29% for female partners) (Table 1). Median duration of sexual partnerships reported by in-migrants was significantly shorter (women: 1.0 vs. 6.0 years; men: 1.0 vs. 2.0 years; both genders, p<0.0001), and relationships were less likely to be ongoing compared to those reported by long-term residents (80% vs. 84% for male partners; 62% vs. 71% for female partners) (Table 1).
Table 2 shows risk factors and sexual behaviors related to HIV among the reported sexual partnerships of in-migrant and long-term resident participants. The male partners of in-migrant women were more likely to be circumcised compared to the male partners of long-term resident women (43% vs. 36%); however, inconsistent condom use was significantly more common among partnerships reported by female in-migrants compared to those of female long-term residents (26% vs. 20%). Irrespective of sex, in-migrants were less likely to know their partner’s HIV status and to have disclosed their HIV status to their partners as compared to long-term residents (Table 2).
Table 2.
Sexual risk factors and behaviors among the self-reported sexual partnerships of RCCS participants stratified by participants’ residence status and sex, 1999–2016
| Self-reported male partners (n=56,192) | Self-reported female partners (n=60,552) | |||||
|---|---|---|---|---|---|---|
| Long-term resident women (n=46,870) | In-migrant women (n=9,322) | PRR (95% CI) | Long-term resident men (n=54,304) | In-migrant men (n=6,248) | PRR (95% CI) | |
| Partner's circumcision status | n=27702 | n=4991 | ||||
| Circumcised | 10101 (36%) | 2143 (43%) | 1.18 (1.14, 1.22) | |||
| Uncircumcised | 17362 (63%) | 2773 (56%) | 0.89 (0.86, 0.91) | |||
| Don’t know | 207 (0.75%) | 61 (1.2%) | 1.64 (1.22, 2.16) | |||
| Condom usage with partner | n=37010 | n=7559 | n=42824 | n=4956 | ||
| Never | 25157 (68%) | 4768 (63%) | 0.93 (0.91, 0.95) | 23100 (54%) | 2794 (56%) | 1.05 (1.02, 1.07) |
| Inconsistent | 7427 (20%) | 1929 (26%) | 1.27 (1.22, 1.33) | 8851 (21%) | 1006 (20%) | 0.98 (0.93, 1.04) |
| Always | 4404 (12%) | 846 (11%) | 0.94 (0.88, 1.01) | 10760 (25%) | 1140 (23%) | 0.92 (0.87, 0.97) |
| Don’t know | 2 (<0.01%) | 0 (0.0%) | -- | 4 (<0.01%) | 1 (0.020%) | -- |
| RCCS participant uses alcohol before sex with partner | n=9308 | n=2568 | n=11369 | n=1792 | ||
| Yes | 1983 (21%) | 585 (23%) | 1.07 (0.98, 1.16) | 3837 (34%) | 551 (31%) | 0.91 (0.84, 0.98) |
| No | 7325 (79%) | 1971 (77%) | 0.98 (0.95, 1.00) | 7524 (66%) | 1230 (69%) | 1.04 (1.00, 1.07) |
| Partner uses alcohol before sex with RCCS participant | n=9308 | n=2568 | n=11369 | n=1792 | ||
| Yes | 3758 (40%) | 973 (38%) | 0.94 (0.89, 0.99) | 2130 (19%) | 359 (20%) | 1.06 (0.97, 1.18) |
| No | 5549 (60%) | 1583 (62%) | 1.03 (1.00, 1.07) | 9231 (81%) | 1422 (79%) | 0.98 (0.95, 1.00) |
| Partner's HIV status was ever known by RCCS participant | n=17658 | n=4382 | n=21032 | n=3035 | ||
| Yes | 11094 (63%) | 2590 (59%) | 0.94 (0.92, 0.97) | 11539 (55%) | 1495 (49%) | 0.90 (0.86, 0.93) |
| No | 6557 (37%) | 1773 (40%) | 1.09 (1.05, 1.13) | 9478 (45%) | 1529 (50%) | 1.12 (1.07, 1.16) |
| RCCS participant informed partner of his/her HIV status | n=17658 | n=4382 | n=21032 | n=3035 | ||
| Yes | 8600 (49%) | 2032 (46%) | 0.95 (0.92, 0.99) | 8283 (39%) | 1014 (33%) | 0.85 (0.80, 0.89) |
| No | 2772 (16%) | 733 (17%) | 1.07 (0.99, 1.15) | 5750 (27%) | 772 (25%) | 0.93 (0.87, 0.99) |
| Never tested/got HIV results | 1011 (5.7%) | 521 (12%) | 2.08 (1.88, 2.29) | 2478 (12%) | 765 (25%) | 2.14 (1.99, 2.30) |
| Received couple counseling | 5270 (30%) | 1080 (25%) | 0.83 (0.78, 0.87) | 4507 (21%) | 472 (16%) | 0.73 (0.66, 0.79) |
| Don’t remember | 1 (<0.01%) | 0 (0.0%) | -- | 1 (<0.01%) | 0 (0.0%) | -- |
Data are n (%) unless otherwise noted. Denominators vary for characteristics due to different survey rounds in which data were obtained (see Table S2, Supplemental Digital Content) and include sexual partners for whom data is missing. Some column percentages do not add to 100 due to rounding and missing data. RCCS = Rakai Community Cohort Study. PRR = Prevalence risk ratio with the reference group defined as long-term resident RCCS participants. CI=Confidence interval.
Results were similar when stratified by the HIV serostatus of RCCS participants with few exceptions (see Tables S9–S12, Supplemental Digital Content). Of note, consistent condom use was less common within sexual partnerships of HIV-positive in-migrant men as compared to HIV-positive long-term resident men (19% vs. 26%). Additionally, alcohol use before sex was more common within the partnerships of HIV-positive in-migrants compared to HIV-positive long-term residents, regardless of sex.
HIV prevalence among epidemiologically-linked cohabitating couples by residence status, sex, and calendar time
There were 13,708 study participants (n=7,368 women; 6,340 men) in an epidemiologically-linked partnership with at least one other RCCS participant. These epidemiologically-linked couples (n=7,558) contributed 21,137 couple-visits (i.e., a single survey visit at which two participants were in an epidemiologically-linked couple) and comprised 47% (n=13,708/29,423) of all study participants. Of these couple-visits, 82.9% (n=17,527) were between two long-term residents, 5.5% (n=1,162) between two in-migrants, 10.1% (n=2,149) between female in-migrants and male long-term residents, and 1.4% (1,162) between male in-migrants and female long-term residents. Couple-visits including recent in-migrants (one or both partners) were more likely to be HIV serodiscordant and concordant HIV seropositive than were couple-visits among long-term residents exclusively. For example, 17.7% of couple-visits between long-term resident women and in-migrant men were classified as HIV serodiscordant compared to only 8.4% of couple-visits between long-term residents (see Table S13, Supplemental Digital Content).
Overall, in-migrant status was associated with a 1.5-fold increased risk of sexual partnership with ≥1 HIV-positive persons among both sexes (Table 3). Among HIV-negative participants only, risk of partnership with an HIV-positive person among in-migrants was nearly doubled that relative to long-term residents (women: adjPRR=1.92, 95% CI: 1.64–2.23; men: adjPRR=1.87, 95% CI: 1.55–2.25). Over the entire study period, HIV-negative in-migrant women were significantly more likely to have an untreated HIV positive male partner compared to HIV-negative long-term resident women (adjPRR=1.77, 95% CI: 1.49–2.11). Similarly, HIV-negative in-migrant men also were more likely have an untreated HIV-positive partner compared to HIV-negative long-term resident men (adjPRR=1.72, 95% CI: 1.38–2.14).
Table 3.
Prevalence of ≥1 sexual partnership with an HIV-positive person stratified by RCCS participants’ residence status, sex, and CHP calendar period
| Long-term resident women | In-migrant women | Female adjPRR (95% CI) | Long-term resident men | In-migrant men | Male adjPRR (95% CI) | |
|---|---|---|---|---|---|---|
| Partnership with ≥1 HIV-positive person among both HIV-negative and HIV-positive RCCS participants | ||||||
| All survey rounds | 12.5% (2229/17786) | 15.3% (505/3299) | 1.53 (1.39, 1.68) | 12.6% (2363/18779) | 17.3% (249/1437) | 1.48 (1.32, 1.67) |
| Pre-CHP (1999–2004) | 11.8% (712/6045) | 14.6% (135/922) | 1.46 (1.23, 1.75) | 11.1% (700/6286) | 15.0% (57/380) | 1.44 (1.13, 1.85) |
| Early CHP (2005–2011) | 12.7% (795/6245) | 16.1% (177/1101) | 1.57 (1.34, 1.83) | 12.9% (842/6546) | 17.9% (85/475) | 1.47 (1.20, 1.80) |
| Late CHP (2011–2016) | 13.1% (722/5496) | 15.1% (193/1276) | 1.61 (1.38, 1.88) | 13.8% (821/5947) | 18.4% (107/582) | 1.47 (1.22, 1.77) |
| Partnership with ≥1 HIV-positive person among HIV-negative RCCS participants | ||||||
| All survey rounds | 5.2% (805/15640) | 7.9% (220/2783) | 1.92 (1.64, 2.23) | 5.1% (836/16371) | 9.3% (115/1240) | 1.87 (1.55, 2.25) |
| Pre-CHP (1999–2004) | 5.3% (285/5412) | 7.7% (61/792) | 1.75 (1.33, 2.32) | 4.3% (239/5522) | 7.1% (23/326) | 1.78 (1.19, 2.67) |
| Early CHP (2005–2011) | 5.0% (276/5481) | 7.5% (69/918) | 1.97 (1.50, 2.58) | 4.9% (280/5687) | 9.3% (38/409) | 1.88 (1.35, 2.61) |
| Late CHP (2011–2016) | 5.1% (244/4747) | 8.4% (90/1073) | 2.17 (1.67, 2.81) | 6.1% (317/5162) | 10.7% (54/505) | 1.72 (1.30, 2.27) |
| Partnership with ≥1 untreated HIV-positive persons among HIV-negative RCCS participants | ||||||
| All survey rounds | 4.1% (635/15640) | 6.3% (175/2783) | 1.77 (1.49, 2.11) | 3.9% (644/16371) | 6.9% (85/1240) | 1.72 (1.38, 2.14) |
| Pre-CHP (1999–2004) | 5.3% (285/5412) | 7.7% (61/792) | 1.75 (1.33, 2.32) | 4.3% (239/5522) | 7.1% (23/326) | 1.78 (1.19, 2.67) |
| Early CHP (2005–2011) | 4.1% (225/5481) | 6.4% (59/918) | 2.01 (1.49, 2.70) | 4.3% (247/5687) | 8.3% (34/409) | 1.89 (1.33, 2.68) |
| Late CHP (2011–2016) | 2.6% (125/4747) | 5.3% (55/1073) | 1.97 (1.39, 2.79) | 3.1% (158/5162) | 5.5% (28/505) | 1.54 (1.03, 2.30) |
Data are % (n/N) unless otherwise noted, with n representing the number of person-visits. RCCS = Rakai Community Cohort Study. adjPRR =Prevalence risk ratio adjusted for age and marital status with the reference group defined as long-term resident RCCS participants. CI=Confidence interval. CHP = Combination HIV prevention.
Figures 1A and B show the proportion of couple-visits contributed by HIV-negative participants at each of the 12 survey visits in which they were epidemiologically-linked to an untreated HIV-positive partner, stratified by their residence status and sex. The proportion of untreated serodiscordant partnerships among HIV-negative in-migrant women peaked at 10.3% (95% CI: 6.0–16.1%) in 2000 and declined to 3.4% (95% CI: 1.8–5.8%) by 2016. Among HIV-negative long-term resident women, this proportion was generally lower than that of HIV-negative in-migrant women, ranging from 5.9% (95% CI: 4.4–7.6%) in 1999 to 2.1% (95% CI: 1.5–3.0%) in 2014. Among HIV-negative in-migrant men, the proportion of untreated serodiscordant partnerships declined from 9.5% (95% CI 4.4–17.2%) in 2002 to 1.9% (95% CI: 0.4–5.4%) by 2016, and among HIV-negative long-term resident men from 5.2% (95% CI: 3.9–6.7%) in 1999 to 2.2% (95% CI: 1.6–2.9%) by 2016. Figures S2–S3 show the prevalence of sexual partnership with an untreated HIV-positive partner among untreated and treated HIV-positive RCCS participants, respectively. Stratification by calendar period (pre-CHP, early CHP, and late CHP) showed in-migrants of both sexes had a consistently higher likelihood of having at least one untreated HIV-positive partner irrespective of rising ART and male circumcision coverage (Table 3).
Figure 1.
Prevalence of sexual partnership with an untreated HIV-positive partner among HIV-negative women (A) and HIV-negative men (B) in the Rakai Community Cohort Study (RCCS), 1999–2016. Here, prevalence is defined as the proportion of couple-visits contributed by HIV-negative participants in which they partner with an untreated HIV-positive partner. Prevalence is stratified by migration status of the index HIV-negative RCCS participant.
Lastly, we stratified data by both the residence status of the index participant and their epidemiologically-linked partners (Table 4; see Tables S14–16, Supplemental Digital Content, for further stratification by CHP calendar period). The proportion of HIV-negative participants with ≥1 untreated HIV-positive partner was highest among male in-migrants with only long-term resident female partners (11%), while lowest among male residents with no-migrant partners (3.4%; adjPRR=2.86, 95% CI:1.96–4.17). HIV-negative female in-migrants with only long-term resident partners were twice as likely to have an untreated HIV-positive male partner compared to HIV-negative female long-term residents with no migrant partners (7 vs 4%; adjPRR=2.03, 95% CI: 1.66–2.48). In contrast, HIV-negative female in-migrants with a male in-migrant partner were no more likely to have an untreated HIV-positive partner relative to HIV-negative long-term resident women with only long-term resident male partners.
Table 4.
Proportion of RCCS participants with ≥1 HIV-positive sexual partner by residence status of index participant and their sexual partners, 1999–2016
| Gender and residence status of RCCS participant | Partnership with ≥1 in-migrant | Proportion of participants with ≥1 HIV-positive partner | adjPRR (95% CI) | Proportion of HIV-negative participants with ≥1 HIV-positive partner | adjPRR (95% CI) | Proportion of HIV-negative participants with ≥1 untreated HIV-positive partner | adjPRR (95% CI) |
|---|---|---|---|---|---|---|---|
| Female long-term resident | no | 12.5% (2181/17487) | -- | 5.1% (793/15419) | -- | 4.0% (623/15419) | -- |
| Male long-term resident | no | 12.4% (2003/16641) | -- | 4.6% (670/14580) | -- | 3.4% (498/14580) | -- |
| Female long-term resident | yes | 16.1% (48/299) | 1.27 (0.98, 1.65) | 5.4% (12/221) | 1.03 (0.59, 1.80) | 5.4% (12/221) | 1.36 (0.78, 2.38) |
| Male long-term resident | yes | 16.8% (360/2138) | 1.62 (1.46, 1.80) | 9.3% (166/1791) | 2.24 (1.89, 2.66) | 8.2% (146/1791) | 2.54 (2.11, 3.06) |
| Female in-migrant | no | 16.3% (348/2139) | 1.69 (1.52, 1.89) | 9.0% (161/1794) | 2.30 (1.93, 2.74) | 7.0% (125/1794) | 2.03 (1.66, 2.48) |
| Male in-migrant | no | 27.5% (77/280) | 2.28 (1.87, 2.77) | 17.7% (42/237) | 3.56 (2.66, 4.75) | 11.0% (26/237) | 2.86 (1.96, 4.17) |
| Female in-migrant | yes | 13.5% (157/1160) | 1.25 (1.07, 1.45) | 6.0% (59/989) | 1.34 (1.03, 1.74) | 5.1% (50/989) | 1.35 (1.02, 1.80) |
| Male in-migrant | yes | 14.9% (172/1157) | 1.39 (1.20, 1.61) | 7.3% (73/1003) | 1.73 (1.37, 2.19) | 5.9% (59/1003) | 1.80 (1.38, 2.34) |
Data are % (n/N) unless otherwise noted, with n representing the number of person-visits. RCCS = Rakai Community Cohort Study. adjPRR = Prevalence risk ratio adjusted for age and marital status with the reference group defined as same-sex long-term residents with no in-migrant partners (top two rows).
DISCUSSION
In this study of sexual partner networks in Uganda, we found significant differences between the partner pools of recently migrating persons and long-term residents with no recent history of migration. First, the sexual partnerships of in-migrants were more likely to be non-marital, span community boundaries, and be shorter in duration than those of long-term residents in the same community. In-migrants were also less likely to know their sexual partner’s HIV status and to disclose their own HIV serostatus to those partners. Second, among cohabitating epidemiologically-linked couples of known HIV serostatus, we observed a higher proportion of participants with an HIV-positive partner among in-migrants as compared to long-term residents. We also observed a consistently higher burden of untreated HIV infection among the sexual partners of HIV-negative in-migrants compared to those of HIV-negative long-term residents over calendar time, and that risk was highest among in-migrants partnering with long-term residents at place of destination rather than with other in-migrating persons.
Taken together, our findings suggest that the increased risk of HIV acquisition observed among sub-Saharan African migrating populations, particularly in the period immediately following migration,20 is driven in part by new partnership formation in which migrants draw from a partner pool with a higher prevalence of HIV viremia at place of destination. Indeed, we observed substantially higher prevalence of untreated HIV serodiscordant relationships among HIV-negative in-migrants compared to long-term residents. Risk of HIV transmission is significantly higher in untreated serodiscordant relationships compared to those in which the HIV-positive partner is virally suppressed through ART, and particularly so in the absence of measures providing direct protection to the susceptible partner (e.g., condoms, pre-exposure prophylaxis [PrEP]).30,31 That the highest prevalence of untreated HIV was observed among migrants with long-term resident partners suggests that the most risky relationships may be formed shortly after the migration process rather than during it. Prior research from sub-Saharan Africa and elsewhere suggests that migration results in disruption to personal networks with various negative social impacts, including discrimination, social exclusion due to language and cultural barriers, and poor access to local health services, all of which may ultimately impact HIV risk.4,32,33
Of note, we observed a significantly higher frequency of HIV risk-related behaviors among the partnerships of in-migrants compared to the partnerships of long-term residents. Despite our observation that in-migrants were more likely to be married than long-term residents, they were more likely to report multiple partnerships and be in non-marital partnerships of shorter duration than long-term residents. In-migrants were also more likely to partner with someone from outside their community of residence, a risk factor which has been strongly linked to incident HIV infection,27 and to not disclose their HIV status or know their partner’s HIV status. Overall, our findings are consistent with previous reports of increased frequency of risky sexual behaviors among migrating populations in sub-Saharan Africa.14–16 However, such behavioral differences do not fully explain in-migrants’ increased risk of HIV acquisition. In a prior study in this same population, we found that in-migrant status was associated with increased risk of HIV acquisition even after adjustment for participant demographics, including age, and self-reported sexual behaviors.20 Thus, we hypothesize that it is both the higher prevalence of untreated HIV among in-migrants’ sexual partners, in conjunction with their increased risky behaviors within those partnerships, that ultimately leads to their higher risk of HIV acquisition.
Our findings highlight the need for HIV prevention and treatment programs in sub-Saharan Africa that better facilitate clinic transfers and provide services to in-migrants before and promptly following relocation. PrEP is an effective and feasible method for the prevention of HIV transmission in serodiscordant couples,34 and may reduce HIV risk among migrating persons, particularly when taken around the time of migration.
This study has limitations. Critically, analyses of HIV prevalence within epidemiologically-linked partnerships were restricted to cohabitating relationships, which are typically marital or long-term consensual unions. While nearly 50% of all study participants could be linked to a partner of known HIV status in this study, such cohabitating partnerships may not be fully representative of in-migrants’ relationships, particularly if migrants with cohabiting partners are less vulnerable to social exclusion and exploitation than those with non-cohabitating partners. Our inferences also relied on self-reported sexual behavioral data, which is subject to measurement and social desirability biases. However, this study also has several strengths. While previous studies conducted early in the HIV pandemic assessed HIV prevalence among partners left behind by circular migrants and out-migrants,8,35,36 ours is the first to characterize the partner pool of in-migrants relative to long-term residents following relocation before and after the introduction of ART in eastern Africa. Our household census and survey also enabled linkage of cohabitating sexual partners, irrespective of whether those partners were return migrants, allowing us to estimate HIV burden within a broader set of partnerships by residence status.
In summary, we observed increased frequency of HIV-risk related behaviors and prevalence of HIV among the sexual partnerships of in-migrants relative to long-term residents. We also found that the prevalence of untreated HIV among partners of HIV-negative in-migrants was significantly higher than that of long-term residents. Our findings highlight the need for programs that effectively link in-migrants to HIV treatment and prevention services in the period surrounding relocation.
Supplementary Material
Acknowledgements
We thank the Rakai Health Sciences Program as well as the participants of the Rakai Community Cohort Study who made this study possible. We also thank the personnel at the Office of Cyberinfrastructure and Computational Biology at the National Institute of Allergy and Infectious Diseases for data management support.
Sources of Funding:
Supported by the National Institute of Allergy and Infectious Diseases (grants R01AI110324, U01AI100031, U01AI075115, R01AI102939, R01AI143333, and K01AI125086-01), the National Institute of Mental Health (grants R01MH107275 and R01MH105313), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (grants RO1HD070769 and R01HD050180), the Division of Intramural Research of the National Institute for Allergy and Infectious Diseases, the Johns Hopkins University Center for AIDS Research (grant P30AI094189), the President’s Emergency Plan for AIDS Relief through the Centers for Disease Control and Prevention (grant NU2GGH000817), and a cooperative agreement (W81XWH-07-2-0067) between the Henry M. Jackson Foundation for the Advancement of Military Medicine and the U.S. Department of Defense. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the funding agencies.
Footnotes
Conflicts of Interest: The authors of this manuscript declare no conflicts of interest.
List of Supplemental Digital Content
Supplemental Digital Content. Text file (PDF) containing further study background, figures, and tables.
References
- 1.Quinn TC. Population migration and the spread of types 1 and 2 human immunodeficiency viruses. Proc Natl Acad Sci U S A. 1994;91(7):2407–2414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Faria NR, Rambaut A, Suchard MA, et al. The early spread and epidemic ignition of HIV-1 in human populations. Science. 2014;346(6205):56–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lagarde E, Schim van der Loeff M, Enel C, et al. Mobility and the spread of human immunodeficiency virus into rural areas of West Africa. Int J Epidemiol 2003;32(5):744–752. [DOI] [PubMed] [Google Scholar]
- 4.Tanser F, Barnighausen T, Vandormael A, et al. HIV treatment cascade in migrants and mobile populations. Curr Opin HIV AIDS. 2015;10(6):430–438. [DOI] [PubMed] [Google Scholar]
- 5.Faturiyele I, Karletsos D, Ntene-Sealiete K, et al. Access to HIV care and treatment for migrants between Lesotho and South Africa: a mixed methods study. BMC Public Health. 2018;18(1):668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bygrave H, Kranzer K, Hilderbrand K, et al. Trends in loss to follow-up among migrant workers on antiretroviral therapy in a community cohort in Lesotho. PLoS One. 2010;5(10):e13198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Frank T, Carter A, Jahagirdar D, et al. Global, regional, and national incidence, prevalence, and mortality of HIV, 1980–2017, and forecasts to 2020, for 195 countries and territories: a systemic analysis for the Global Burden of Diseases, Injuries, and Risk Factors Study 2017. Lancet HIV. 2019;6(12):e831–e859. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kishamawe C, Vissers DC, Urassa M, et al. Mobility and HIV in Tanzanian couples: both mobile persons and their partners show increased risk. AIDS. 2006;20(4):601–608. [DOI] [PubMed] [Google Scholar]
- 9.Camlin CS, Akullian A, Neilands TB, et al. Gendered dimensions of population mobility associated with HIV across three epidemics in rural Eastern Africa. Health Place. 2019;57:339–351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Camlin CS, El Ayadi AM, Kwena ZA, et al. High Mobility and HIV Prevalence Among Female Market Traders in East Africa in 2014. J Acquir Immune Defic Syndr 2017;74(5):e121–e128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sartorius B, Kahn K, Collinson M, et al. Dying in their prime: determinants and space-time risk of adult mortality in rural South Africa. Geospat Health. 2013;7(2):237–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Levira F, Todd J, Masanja H. Coming home to die? The association between migration and mortality in rural Tanzania before and after ART scale-up. Glob Health Action. 2014;7:22956. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Determinants of the risk of dying of HIV/AIDS in a rural South African community over the period of the decentralised roll-out of antiretroviral therapy: a longitudinal study. Glob Health Action. 2014;7:24826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dzomba A, Tomita A, Govender K, et al. Effects of Migration on Risky Sexual Behavior and HIV Acquisition in South Africa: A Systematic Review and Meta-analysis, 2000–2017. AIDS Behav 2019;23(6):1396–1430. [DOI] [PubMed] [Google Scholar]
- 15.Schuyler AC, Edelstein ZR, Mathur S, et al. Mobility among youth in Rakai, Uganda: Trends, characteristics, and associations with behavioural risk factors for HIV. Glob Public Health. 2017;12(8):1033–1050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Camlin CS, Akullian A, Neilands TB, et al. Population mobility associated with higher risk sexual behaviour in eastern African communities participating in a Universal Testing and Treatment trial. J Int AIDS Soc 2018;21(S4):e25115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.2020 Global AIDS Update – Seizing the moment – Tackling entrenched inequalities to end epidemics. UNAIDS, July 6, 2020. Available at: https://www.unaids.org/sites/default/files/media_asset/2020_global-aids-report_en.pdf. [Google Scholar]
- 18.Joshi K, Lessler J, Olawore O, et al. Declining HIV incidence in sub-Saharan Africa: a systemic review and meta-analysis of empiric data. MedRxiv (preprint). 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Grabowski MK, Lessler J, Bazaale J, et al. Migration, hotspots, and dispersal of HIV infection in Rakai, Uganda. Nat Commun 2020;11:976. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Olawore O, Tobian AAR, Kagaayi J, et al. Migration and risk of HIV acquisition in Rakai, Uganda: a population-based cohort study. Lancet HIV. 2018;5(4):e181–e189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Cassels S, Jenness S, Khanna A. Conceptual framework and research methods for migration and HIV transmission dynamics. AIDS Behav 2014;18(12):2302–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Grabowski MK, Serwadda DM, Gray RH, et al. HIV Prevention Efforts and Incidence of HIV in Uganda. N Engl J Med 2017;377(22):2154–2166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Grabowski MK, Reynolds SJ, Kagaayi J, et al. The validity of self-reported antiretroviral use in persons living with HIV: a population-based study. AIDS. 2018;32(3):363–369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gray RH, Kigozi G, Serwadda D, et al. Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet. 2007;369(9562):657–666. [DOI] [PubMed] [Google Scholar]
- 25.Gray RH, Makumbi F, Serwadda D, et al. Limitations of rapid HIV-1 tests during screening for trials in Uganda: diagnostic test accuracy study. BMJ. 2007;335(7612):188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Galiwango RM, Musoke R, Lubyayi L, et al. Evaluation of current rapid HIV test algorithms in Rakai, Uganda. J Virol Methods. 2013;192(1–2):25–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Grabowski MK, Lessler J, Redd AD, et al. The role of viral introductions in sustaining community-based HIV epidemics in rural Uganda: evidence from spatial clustering, phylogenetics, and egocentric transmission models. PLoS Med 2014;11(3):e1001610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Republic of Uganda Ministry of Health. National HIV Testing Services Policy and Implementation Guidelines, 4th Edition. Republic of Uganda Ministry of Health, October 2016. Available at: https://health.go.ug/content/national-hiv-testing-services-policy-and-implementation-guidelines. [Google Scholar]
- 29.Cui J. QIC program and model selection in GEE analyses. Stata J 2007;7(2):209–220. [Google Scholar]
- 30.Cohen MS, Chen YQ, McCauley M, et al. Antiretroviral Therapy for the Prevention of HIV-1 Transmission. N Engl J Med 2016;375(9):830–839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Liu H, Su Y, Zhu L, et al. Effectiveness of ART and condom use for prevention of sexual HIV transmission in serodiscordant couples: a systematic review and meta-analysis. PLoS One. 2014;9(11):e111175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.International Organization for Migration. Migration in Uganda: A Rapid Country Profile 2013. International Organization for Migration, 2015. Available at: https://publications.iom.int/books/migration-uganda-rapid-country-profile-2013. [Google Scholar]
- 33.Suphanchaimat R, Sommanustweechai A, Khitdee C, et al. HIV/AIDS healthcare challenges for cross-country migrants in low- and middle-income countries: a scoping review. HIV/AIDS (Auckl). 2014;6:19–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Heffron R, Ngure K, Odoyo J, et al. Pre-exposure prophylaxis for HIV-negative persons with partners living with HIV: uptake, use, and effectiveness in an open-label demonstration project in East Africa. Gates Open Res 2017;1:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lurie MN, Williams BG, Zuma K, et al. The impact of migration on HIV-1 transmission in South Africa: a study of migrant and nonmigrant men and their partners. Sex Transm Dis 2003;30(2):149–156. [DOI] [PubMed] [Google Scholar]
- 36.Lurie MN, Williams BG, Zuma K, et al. Who infects whom? HIV-1 concordance and discordance among migrant and non-migrant couples in South Africa. AIDS. 2003;17(15):2245–2252. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.