Abstract
Women who experience HIV seroconversion during pregnancy are missed during early routine pregnancy HIV screening and are at high risk of perinatal HIV transmission. Male partner HIV testing during routine prenatal care may be an effective primary prevention strategy by identifying women at risk of seroconversion and mitigating their risk. Our objective was to assess interest in and uptake of male partner HIV testing services offered during prenatal care. This demonstration project included all pregnant, English-speaking, HIV-negative women receiving publicly funded prenatal care in an urban hospital-based practice located in a high HIV prevalence area. Women were offered free HIV screening for their male sexual partners. From April 2017 to June 2018, enrolled women completed surveys on social demographics, medical access characteristics, and HIV testing history. Women were invited to bring their partners to a prenatal visit where HIV testing was offered to their male partners. Factors associated with women's interest in testing and completion of partner testing were assessed using bivariable and multivariable analyses. Of 392 women approached, 70% (N = 274) completed study surveys. Although the majority (76%, N = 200 of 264 respondents) of women desired their partner undergo HIV testing, testing was underutilized as only 18 (7%) male partners completed testing. While neither maternal characteristics nor male social or attitudinal factors were associated with interest in or completion of partner HIV testing, sensitivity analyses, performed with multiple imputation, demonstrated some association between interest and completion of partner testing and partner medical care access and utilization. In conclusion, although the majority of low-income women in an urban prenatal clinic expressed interest in having their partners undergo HIV testing, uptake of free partner HIV testing services was uncommon. A focused assessment of implementation and uptake barriers is needed to optimize partner testing and eliminate HIV transmission to mothers and their babies.
Introduction
Although the number of children diagnosed with perinatally acquired HIV in the United States continues to decline, ongoing perinatal HIV transmission continues to occur. 1 The Fetal and Infant Mortality Review (FIMR)/HIV Prevention Methodology, a process created to monitor, review, and prevent cases of HIV transmission, reviews such events. 2 Findings from FIMR indicate that cases of maternal seroconversion occur each year and are a leading cause of perinatal HIV transmission. 2 Along with appropriate antenatal care for women with HIV, 3 mitigating the risk of maternal HIV acquisition is essential to preventing perinatal HIV transmission. The existing framework for elimination of perinatal HIV transmission has centered around case finding: identification of pregnant women with HIV and linking them to the requisite medical and social services required to prevent perinatal transmission. 4 Such efforts are essential, yet have proven inadequate to fully eliminate both maternal acquisition of HIV and subsequent transmission to offspring. One adjuvant approach to get closer to the goal of elimination of perinatal transmission of HIV is to expand this framework.
One avenue for expansion is that of primary prevention: preventing maternal acquisition of HIV during pregnancy, which, naturally, avoids risk of perinatal transmission. Compared to nonpregnant women, pregnant women are at increased risk of HIV seroconversion. 5 Furthermore, women who seroconvert during pregnancy are at significantly elevated risk of transmitting HIV to their fetus or newborn. 6,7 Unpublished data from our group demonstrate that in Illinois, 41% of women who seroconverted during pregnancy transmitted HIV to their newborn. Over the past decade, Illinois has continued to have cases (n = 0–4 per year) of perinatal HIV transmission. Thus, identification of pregnant women at high risk of acquiring new HIV infection during pregnancy is critical to the goal of zero babies born positive.
One group of women at high risk of acquiring HIV during pregnancy is women in a serodiscordant relationship. However, men who have sex with women, particularly in racial or ethnic minority populations, are suboptimally tested and thus often lack knowledge of their HIV serostatus. 8 These data were corroborated by our team, demonstrating that fewer than a quarter of pregnant women in a high HIV prevalence (defined as incidence greater than 1 per 1,000 individuals) community knew their partner's HIV serostatus. 9,10 Such lack of knowledge is a key gap in the prevention cascade, as identification of positive sexual partners affords a multitude of prevention opportunities, including treatment as prevention, 11 pre-exposure prophylaxis, 12,13 or barrier protection. 14
The potential impact of these prevention strategies is dramatic, but positive partners must be identified to deploy these effective strategies. With use of the 4th generation HIV test, male partners undergoing seroconversion themselves (those at highest risk of transmitting HIV to partners), as well as those living with chronic infection, will be identified. While testing services are available and can be obtained free of charge in the state of Illinois through Departments of Public Health, uptake of testing by key populations such as men who have sex with women is limited. A needs assessment conducted at our center identified that although the majority of women desired for their partners to undergo HIV testing, few of their partners had access to medical care or were aware of community resources to facilitate HIV testing. 9 Based on these findings, we implemented a quality improvement program of free partner HIV testing embedded within prenatal care. Similar programs have been employed in international settings 15,16 ; however, due to fundamental differences in health care access, HIV prevalence, and gender roles, these data are limited in their ability to inform the successful conduct of similar programs in the United States. Furthermore, scant data exist regarding partner testing during maternal prenatal care in the United States.
Thus, our primary objective was to measure the uptake of free HIV testing for male partners of pregnant women receiving publicly funded prenatal care in a high HIV prevalence urban center in the United States. Secondary objectives included an assessment of women's interest in partner testing as well as identification of characteristics associated with maternal interest in testing and partner uptake of testing services.
Materials and Methods
Study participants: female partners
In this prospective quality improvement project, pregnant women receiving publicly funded prenatal care at Northwestern Medicine in Chicago, IL, were approached for participation between April 2017 and June 2018. This institution is a large academic medical center located in a geographic region noted to have high HIV prevalence. 17 A diverse population of women, including women from low-income and racial/ethnic minority communities in Chicago, receive obstetric care at this center. Most patients who receive care in this clinic are English speakers, with less than 5% self-reporting as monolingual Spanish speakers. This project took place within a single hospital-based specialized practice that predominantly cares for low-income women and serves a particularly vulnerable population. Of the ∼12,000 deliveries that occur annually at this tertiary-care hospital, 300–400 are patients from this practice. Women were eligible for participation if they were pregnant, HIV negative, and English speaking, and had publicly funded prenatal care. Notably, women living with HIV are seen in a separate multidisciplinary clinic with subspecialists in perinatal HIV care.
This study was approved by the Northwestern University Institutional Review Board. Participants did not receive any compensation for their participation. Women were approached for participation at their first prenatal visit or shortly thereafter by a trained study coordinator. The coordinator self-identifies as a Black woman. While she is not bilingual, she had 3 years of prior experience with patient engagement within this specific obstetric practice. The quality improvement team consisted of leaders in maternal-fetal medicine and infectious diseases, who also have expertise in quality improvement, implementation science, and health services research methodology.
Measures: female partners
After undergoing written, informed consent, women completed a one-time brief survey. Participants were given the option to complete the survey at the time of enrollment either online using Research Electronic Data Capture (REDCap) 18 software or on paper, depending on patient preference and computer availability in clinic. Paper data entry to REDCap was confirmed for accuracy by a second team member. Survey topics included relationship status, whether her partner had health insurance, whether her partner had a primary care provider (PCP), and whether she wanted her partner tested for HIV. These items were drawn from established literature; additional items were generated by the research team based on our prior needs assessment. 9 Medical records were reviewed by trained research assistants to obtain sociodemographic information along with medical and pregnancy history. All survey and medical record data were stored securely in REDCap.
Study participants: male partners
After completion of enrollment surveys, participating women underwent HIV counseling with specific emphasis on the risk of HIV acquisition in pregnancy and the importance of knowledge of the serostatus of sexual partners. Women also underwent education on intimate partner violence and the potential of identification of HIV serodiscordance to potentiate the risks. After this counseling, women were queried regarding their interest in having their male sexual partners undergo free HIV testing at the site of prenatal care. Any male partner or other guest present in the room was asked to leave before approaching women to ensure patient safety. Women who expressed interest in having their partner tested were invited to bring their partners with them to a prenatal visit; the onus on male partner recruitment thus resided primarily with the woman. If a woman expressed interest in partner testing and her male partner was not present at the subsequent prenatal visit, a research assistant followed up with the female participant in the clinic to establish ongoing interest. This consent construct was designed to allow women to invite the male partner of their choosing for testing and to promote safety of the female participants.
Men were eligible for inclusion if they were at least 18 years of age or over and English speaking. Men were eligible for participation regardless of the recency of prior HIV testing, insurance status, or any other health care access factor. Male partners were approached by the study coordinator and underwent written, informed consent for study participation. Pre-test HIV counseling occurred and educational written materials were provided. This information included the prevalence of HIV in Illinois, the importance of HIV testing and treatment for their own health, and the role of partner testing for prevention of perinatal transmission. For men who consented to HIV testing, a blood draw occurred in the laboratory site affiliated with the prenatal office site following enrollment in the health system medical record and coordination of a post-test counseling plan. Male partners were given two reminder calls from the research coordinator regarding their opportunity to complete HIV testing. Post-test communication of results occurred largely by phone. Testing was recommended to occur during the visit wherein informed consent was obtained to optimize the chance of completion of testing. However, men were given the opportunity to return to the laboratory for testing any time before their pregnant partners' delivery. We did not collect the timing of when tests were completed relative to the timing of consent.
Disclosure planning
Before program initiation, an algorithm was created with input from the Northwestern Medicine Division of Infectious Disease, the Ethics Committee, and Department of Social Work that encouraged male partners to communicate test results to their pregnant partners. Consent for participation stipulated that if disclosure of a confirmed positive HIV serostatus did not occur within a clinically appropriate timeline, the clinical investigators would directly inform the pregnant partners of the results. Informing female participants of positive test results was deemed ethically necessary due to both the risk posed to her health and the high risk of perinatal HIV transmission. Both male and female participants were informed of this protocol.
Statistical design and power
Preliminary data demonstrated that only 20% of pregnant patients were aware of partners' HIV test within the preceding 6 months, within the timeframe of their current pregnancy. 9 Using an alpha of 0.05, we anticipated enrolling 120 male partners to have 80% power to detect a 20% absolute increase in HIV testing (to a final 40% partner testing rate). Assuming 70% of partners of interested women would express interest in free testing services and 45% of partners of interested women would ultimately receive testing, we planned to enroll 270 pregnant women and 190 partners.
The primary outcome was a quality improvement metric—uptake of free HIV testing for the male partners of enrolled pregnant women. For analysis of secondary outcomes, a woman was considered to have been interested in partner testing if she responded affirmatively to the query regarding desire for partner testing, regardless of whether her partner was contacted, presented for participation, or ultimately underwent testing. Female sociodemographic and attitudinal (survey responses) factors associated with (1) maternal desire for testing and (2) male uptake of testing were assessed using chi-squared, Fisher's exact, and Wilcoxon rank sum tests, as appropriate. For the evaluation of maternal desire for testing, the sample was restricted to respondents of the survey item regarding maternal desire. For the evaluation of male uptake of testing, all consented women were included in analyses. The a priori study plan was for a complete case analysis, and thus, missing data were omitted as individual data points from the analyses. Variables with more than 5% of data points missing were evaluated in a sensitivity evaluation to determine whether data were missing at random. For data not missing at random, multiple imputation was performed post-hoc using chained equations (m = 5), using auxiliary variables that were associated (p < .05) with the missing variable of interest in bivariable analyses. Sensitivity analyses were performed using these imputed data entries.
Multivariable models were not generated to assess demographic or access/testing history factors independently associated with either outcome because of the lack of statistically significant differences on bivariable models. We did not a priori identify any covariate we wished to control for as we considered this a health service intervention that was targeted to reach the entire clinic population systematically. Note that this analysis is of female responses only; male attitudinal responses were not included in this analysis due to the substantial bias associated with poor male participation (described below). All analyses were performed with Stata v. 15 (StataCorp, College Station, TX). This report abides by SQUIRE 2.0 reporting guidelines.
Results
During the study period, 392 women were approached for participation, of whom 274 (70%) consented for participation (Fig. 1). Of the 274 eligible women, 264 responded to the survey item pertaining to their interest in partner testing. Of these women, 172 (65%) reported affirmatively when asked if they had previously discussed HIV testing with their partners. The median age of women who participated was 27 (Table 1). Most women were non-Hispanic black and were married or partnered in an unmarried couple.
Study flow diagram.
Sociodemographic, HIV Testing, and Medical Access Characteristics Stratified by Endorsed Desire for Partner Testing
Data presented as median (interquartile range) or n (%).
Includes all women (n = 264) who responded to the survey item regarding desire for partner testing.
PCP, primary care provider.
Of the 264 women who responded to the survey item regarding desire for partner testing, 200 women (76%) reported they desired for their sexual partners to obtain the free HIV testing being offered within the prenatal clinic. Of these 200 women, only 65 had partners who presented to the prenatal clinic for discussion of HIV testing. Of these 65 male partners, 50 (77% of those who presented) consented to participate and 40 (80% of those who consented) reported a desire to undergo HIV testing. However, ultimately, only 18 (7%) of the partners of women underwent HIV testing through our program. There were no sociodemographic differences between women who desired and did not desire partner testing (Table 1). Similarly, there were no differences in women's desire for testing based on HIV testing history of themselves, their reported testing history of their partners, or their partners' access to health care.
There were no differences in female sociodemographic characteristics based on whether male partners completed testing versus did not complete testing. Although men who had been tested in the past were less likely to undergo HIV testing, there were no other HIV testing or medical access characteristics associated with completing partner testing (Table 2).
Sociodemographic, HIV Testing, and Medical Access Characteristics Stratified by Whether Partner Completed HIV Testing
Data presented as median (interquartile range) or n (%).
Includes all women who consented to participate.
Review of missing data demonstrated four variables with a high (>5%) degree of missing-ness: nulliparity (8%), awareness that the female participant had an HIV test during pregnancy (6%), male partner insurance status (20%), and male partner having a PCP (41%). Bivariable analyses were performed to analyze the association between each of these variables and sociodemographic and clinical characteristics of participants (Table 3). These identified associations were used to perform multiple imputation. Bivariable analyses were conducted with these imputed data, with results depicted in Table 4. These data suggest that, compared to women who did not express interest in having their partner undergo HIV testing, women who wanted their partner tested were less likely to be nulliparous. In addition, they were less likely to report their partner had a PCP and more likely to report that he was uninsured. Compared to women whose male partner did not complete HIV testing, those whose male partner did complete HIV testing were less likely to know they had an HIV test during their pregnancy. In addition, they were less likely to report their partner was uninsured and more likely to report that he had a PCP.
Bivariable Associations Between Variables with a High Degree of Missing-Ness
Emboldened text represents statistical significance at p < .05.
N/A, not applicable.
Sensitivity Analyses with Multiple Imputation
Discussion
In this demonstration project, we attempted to eliminate some of the financial, educational, and logistical barriers to obtaining HIV testing by offering free partner testing embedded within an urban obstetric clinic serving predominantly racial/ethnic minority women with publicly funded prenatal care. However, implementation of this free partner testing program did not result in robust uptake of these services. Although the majority of women who consented to participate desired partner testing for HIV, very few of their partners utilized the free testing services, suggesting that additional barriers to testing services were present.
Our findings underscore the need for further optimization of acceptability of a partner testing program before clinical implementation. One hypothesized reason for the lack of uptake of free HIV testing services in our study could be that the testing was performed individually as opposed to couple-based testing, which has been effective in international settings. 19 –21 A negative HIV test by the pregnant partner may be interpreted as an absence of a need for male partner testing. Future programmatic efforts should explore whether couple-based testing improves uptake of testing services. In addition, as this project was conducted as part of a research program, the requirement of informed consent, which included the potential for disclosure of results to their partner, may have presented a barrier to participation. However, of the 65 men approached, 50 (77%) consented and yet, of these men, only 18 (36%) underwent testing, suggesting that provision of consent was not the only barrier to the uptake of free HIV testing. Future programmatic efforts should also address the target population's motivators for testing, in addition to barriers to receiving testing. Finally, there was no specific engagement approach utilized to support male partner testing outside of phone calls regarding the opportunity to have HIV testing performed. Future research should engage male partners and utilize implementation strategies that focus on partners.
Illinois has been consistently cited by the Centers for Disease Control and Prevention (CDC) as one of the highest HIV seroprevalence states in the United States, 22 which places pregnant women at a high risk of seroconversion, given the potential risk of exposure if having sex with partners of unknown serostatus. Despite public health guidelines recommending that all adults be tested for HIV, only half of pregnant women reported knowing that their partners had ever had an HIV test. This is consistent with our prior work in this same population, demonstrating not only that male partners had low rates of testing in accordance with CDC guidelines but also that pregnant women lacked knowledge of their partners' testing results. 9 In addition, CDC data from 2016 indicate that in geographic regions of high HIV prevalence, only 39.2% of males who had had sex with a member of the opposite sex in the last 12 months, never injected drugs, and met low-income or low education guidelines had completed HIV testing in the last 12 months. 23 Thus, there is an urgent need to improve testing in male partners of pregnant women, particularly in high seroprevalence areas, as well as to improve couples' communication about the outcomes of such testing when it does occur.
Male partner testing affords the opportunity to promote health for the family unit by reinforcing HIV prevention as well as identifying men who need HIV care for themselves. HIV testing for men who have sex with women, and in particular pregnant women accessing prenatal care, represents a potentially underexplored avenue to expand knowledge of HIV status, improve the health of the family unit, and prevent perinatal transmission of HIV. Partner testing has been found to be a high-yield approach to improving the health of the family in international settings 19 –21 and has been identified as a cost-effective perinatal transmission prevention strategy. 24
In addition to its epidemiologic risks, seroconversion during pregnancy contributes to the observed and widening racial and ethnic disparities in HIV transmission in the United States. Non-Hispanic black individuals have the highest burden of incident HIV diagnoses in the United States (44%); moreover, non-Hispanic black women experience the highest number of new HIV diagnoses among women. 25 Furthermore, the majority (82%) of women who seroconverted during pregnancy are non-Hispanic black, contributing to the widening racial disparity in perinatal transmissions. We propose that further efforts to reduce the burden of HIV among women in disadvantaged communities and promote HIV-related health equity must continue to focus expanded education and testing interventions on communities with the greatest need, and additionally seek input from community members to optimize such targeted interventions.
Strengths of this study include its prospective design with the availability of counseling and testing throughout clinic hours. The research coordinator had a designated cellular phone with texting services to facilitate coordination of the partners' consent and testing at his convenience. Moreover, this study was performed in a diverse population of women seeking prenatal care, and we provided comprehensive HIV testing education as a component of the intervention. Limitations of this study include implementation within a single prenatal clinic, limiting generalizability of our findings to other settings, and limiting testing availability to those with geographic or schedule-related constraints. In addition, we were unable to explore male attitudinal or demographic factors that may have precluded testing uptake due to the overall limited male involvement. Small sample sizes also limited the power in some comparisons. Furthermore, for several covariates, including maternal age, awareness of female partner test performed during pregnancy, partner insurance, and whether the partner has a PCP, we encountered significant missing data. The identification of associations between desiring and completing partner testing and couples' characteristics such as engagement in health care in sensitivity analyses with imputation suggest that unmeasured bias may have thwarted observation of associations. However, we believe it is important to interpret this sensitivity analysis with caution due to the instability of findings in this limited sample. Also, although reliance was purposefully placed on women for initiating a conversation about partner testing for her safety, this may have resulted in lower uptake and selection bias. Finally, with a substantial number of male partners indicating interest in completing an HIV test and never doing so, social desirability bias may have played a role in low testing uptake.
Expansion of the perinatal HIV framework to include primary prevention of maternal HIV acquisition is essential to reaching the goals of improving maternal health and eliminating perinatal transmission of HIV. Improving the proportion of partners of pregnant women who have had HIV testing has the potential to minimize the risk of seroconversion during pregnancy. Unfortunately, free partner HIV testing embedded within a prenatal clinic did not result in significant utilization of partner HIV testing services. In an effort to explore contextual elements that influenced the success of the demonstration project, recent work by our group has identified barriers and facilitators to partner testing and reasons for low male uptake. 26 Future work should focus on designing partner testing programs embedded within prenatal care that are convenient and affordable, and align with patients' priorities to encourage HIV testing within this particularly vulnerable population of low-income pregnant women and their partners. Future work must also address issues such as cost of any HIV testing program within the context of the HIV epidemiology of the practice or region.
Conclusions
Elimination of perinatal HIV transmission remains a critical public health goal. Male partner HIV testing has been successful in international settings; however, similar approaches have not been attempted in a U.S. context. We thus initiated a quality improvement project, which offered no-cost HIV tests to the male partners of women receiving publicly funded prenatal care at hospital-based clinic located in a high prevalence area. Despite high female partner interest in obtaining HIV tests for male partners, uptake of testing by males was low. We did not identify any sociodemographic, HIV testing, or medical access characteristic that accounted for differences in desire for testing in both males and females. Future work and attempts at such programs should target the motivations of this population to receive testing as well as work to eliminate any potential barrier to engagement in testing.
Footnotes
Authors' Contributions
All authors contributed equally to this work. E.S.M. and L.M.Y. co-conceived of and supervised the project, co-authored the grant that funded the project, and wrote/edited the article. P.M.G. and S.H.S. contributed scientific and clinical guidance to the study design and implementation, co-authored the grant that funded the project, and provided critical scientific contributions to the article. J.J. collected and analyzed data and edited the article. K.L. collected and analyzed data and wrote/edited the article.
Author Disclosure Statement
No competing financial interests exist.
Funding Statement
This work was supported by the Northwestern Memorial Foundation/Friends of Prentice FY2017 Women's Health Grants Initiative and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (K12 HD050121-11 to L.M.Y. and K12 HD050121-09 to E.S.M.) at the time of study. The funding sources had no involvement in the study design, collection, analysis and interpretation of data, or drafting of the report.