Factors Associated with HIV Seroconversion Among Women Attending an Urban Health Clinic in the South: A Matched Case

Abstract

Compared to knowledge about HIV risk factors among men in the south, less is known about risk factors for women. We conducted an individually matched case–control study to identify factors associated with HIV seroconversion among women. Cases had a clinician-assisted visit (CAV) between 2011 and 2016 at an Atlanta-based public health clinic before HIV diagnosis. Controls were women who visited the clinic but remained HIV negative. Controls were matched to cases in a 2:1 ratio on race, age at first CAV, and date of first CAV. Conditional logistic regression was used to develop a best-fitting model for characterizing HIV risk. Of 18,281 women who were HIV negative at their first visit, 110 (0.6%) seroconverted before 2019. Of these, 80 (73%) had a CAV before HIV diagnosis. Having multiple gonorrhea episodes, a syphilis episode, a greater number of sex partners in the past 2 months, anal sex, history of drug use, history of exchanging drugs or money for sex, and heterosexual sex with >1 sex partner in the last month were individually associated with HIV seroconversion. In multivariate analyses, having a syphilis episode [odds ratio (OR) = 4.7, 95% confidence interval (CI): 1.3–16.3], anal sex (OR = 2.8, 95% CI: 1.0–8.1), and injection drug or crack cocaine use (OR = 33.5, 95% CI: 3.6–313.3) remained associated with HIV. Women having all three risk factors were six times more likely to seroconvert compared to women without these factors. Our results offer insights into which women in a southern HIV “hotspot” may be at greatest risk for HIV.

Introduction

In the early 1990s, limited research existed regarding HIV among women,¹ but over the past 20 years, it has been recognized that the burden of HIV among women has grown considerably. In 2016, women accounted for about a quarter of the new HIV diagnoses in the United States,^1

–5 with an estimated 85% of these infections occurring through heterosexual transmission.⁴

HIV in women is disproportionately concentrated among poor urban women of racial/ethnic minority groups.^2,6,7 Eighty percent of HIV infections in women are among non-Hispanic black and Hispanic women, who make up roughly one-third of the American female population.^8,9 African American women's risk of HIV is ∼20 times that of white women.^7,10 In 2017, new diagnoses of HIV in black heterosexual women was fourfold the new diagnoses in white heterosexual women.² Fortunately, HIV diagnoses among black women have decreased by 25% between 2010 and 2016,² although reductions in HIV diagnoses differ by region and are important to consider and monitor.¹¹

The southern region of the United States has the highest incidence rate of HIV infection in the country, with a rate increasing faster than that of any other region in the country.^12,13 The South contains 37% of the US population, but accounts for 44% of people living with diagnosed HIV.^1,12 Specifically, new HIV diagnoses among black women are increasing in the South.¹⁴

Georgia ranks third highest in lifetime HIV risk in the United States.¹⁰ Fulton County has the highest HIV burden in Georgia and is ranked sixth among all counties in the United States for highest rate of new HIV diagnoses.¹⁵ In 2019, it was named one of 48 “hotspot” counties most necessary to target in the US Department of Health and Human Services' Ending the HIV Epidemic campaign.¹⁶ While Fulton county holds 10% of Georgia's population, it accounts for 25% of its HIV cases.¹⁵ Fifteen percent of new HIV diagnoses in Fulton County are among women.¹⁵ Among these women, 83% of the new HIV cases are non-Hispanic black women and 95% result from heterosexual contact.¹⁵

A number of demographic and behavioral risk factors have been associated with increased risk of acquiring HIV among both men and women. Some of these behavioral risk factors include injection drug use, exchanging sex for drugs and money, inconsistent condom use, and having a high number of sexual partners. Demographic risk factors include poverty, race, unstable housing, and psychological distress/mental health.^{2,4,9,10,17,18} Studies have shown that women have a higher physiological risk of contracting HIV and other sexually transmitted infections (STI) compared to men since the vagina is an ideal environment for bacteria and viruses to thrive.^19
–21 In addition, having a STI increases the risk of acquiring HIV.^19,21,22 Because women are less likely to experience or recognize symptoms for STIs than men, many infections go untreated longer, thus making women more susceptible to HIV.^19,21,22 For example, a study conducted in Louisiana found that women with an STI were diagnosed with HIV at a rate 2.3 times that of women without an STI.²³

While risk factors for HIV among men who have sex with men (MSM) have been well-characterized in various settings and locations across the United States, less is known about the impact these factors have for HIV acquisition among women specifically in urban, low-income Southern US settings.⁸ Despite the high HIV burden among minority women, very few epidemiologic studies have assessed risk factors for HIV among black women in the United States, and more specifically among black women within HIV “hotspots” in the South.¹⁴ Using clinical data obtained on women seeking care at the Fulton County Board of Health (FCBOH) Sexual Health Clinic (SHC) from 2011 until 2016, the goal of this study was to identify factors associated with HIV seroconversion among a primarily underserved African American population in an effort to more effectively inform development of HIV prediction models for women in the South. Such models would then be able to identify women at greatest risk for HIV and guide them toward strategies to prevent HIV acquisition.

Methods

Study design and population

We conducted an individually matched case–control study to identify factors associated with HIV seroconversion among women. Cases and controls for this retrospective matched case–control study were selected from the population of women who sought care at the FCBOH SHC between 2011 and 2016. Women clients were identified as those for whom “female” was selected in the binary gender field available in the FCBOH electronic medical records (EMRs) system; we lacked data on whether women identified as cis- or transgender.

For this study, a case was defined as any woman who seroconverted before 2019, but after having at least one clinician-assisted visit (CAV) on record before her date of HIV diagnosis. A CAV is a clinical encounter occurring for symptom evaluation, laboratory-confirmed sexually transmitted disease treatment, or referral from Partner Services; non-CAV visits are encounters for services such as blood draws for screening tests or retrieval of negative laboratory results. Since detailed sexual histories are only taken at CAVs, non-CAV visits were noted, but not considered due to lack of relevant risk factor information.

All HIV-negative women seeking care at the SHC during this timeframe of interest were examined for potential inclusion as cases. Exclusion criteria included the following: having an HIV diagnosis before 2011, being diagnosed with HIV in the same month and year as the first visit to the SHC, and not having a CAV on record before date of HIV diagnosis. HIV status was confirmed on all women via a cross-check with the Enhanced HIV/AIDS Reporting System (eHARS) database facilitated through a data request to and conducted by the Georgia Department of Public Health.

All women who remained HIV negative (as verified via eHARs) between 2011 and 2016 and had at least one CAV on record were considered for selection as controls. Controls were individually matched to cases in a 2:1 matching ratio on race, age at first CAV (±1 year of age), and date of first CAV (±3 calendar months). Matching on date of first CAV was to control for variations in clinic-level factors such as any changes in personnel, patient triage processes, and/or EMRs documentation that could have affected collection and documentation of patient data.

All analyses were conducted on a de-identified dataset by a study team member (P.H.) lacked access to patient health records. This study protocol was reviewed and approved by both the Institutional Review Boards of the Georgia Department of Public Health and Emory University.

Statistical analysis

Demographic and risk factor data were obtained from electronic SHC medical records associated with the first CAV visit on record for all women who visited the SHC within the timeframe of interest (between 2011 and 2016). Demographic data included age, race, and ethnicity. Risk factor data included number of STI episodes and gynecological infections (e.g., gonorrhea, chlamydia, trichomoniasis, syphilis, and bacterial vaginosis), self-reported sexual behaviors (e.g., sex of sexual partners, number of sex partners, condom use, type of sex, sexual contact with high HIV risk, exchanging drugs/money for sex, and heterosexual sex in the last month), and self-reported drug use (e.g., injecting drug or crack cocaine use). Due to substantial missingness of data on the condom use variables, we created a secondary summary condom variable to describe consistent versus inconsistent condom use. This variable utilized any available data from the three original condom use variables: condom use during last sex (answer options: yes or no), condom use during regular sex (answer options: always, never, or sometimes), and condom use during casual sex (answer options: always, never, or sometimes). If the participant used a condom during their last sexual encounter, always used a condom during regular sex, and always used a condom during casual sex, then they were considered to have consistent condom use. If one or two of these variables was missing but any others reflected consistent condom use based on the previously listed criteria, they were also considered a consistent condom user. If a participant responded to any of the original condom use variables with responses that indicated anything less than routine use (e.g., they did not use a condom during last sexual encounter, never use a condom during regular sex, and sometimes use a condom during casual sex), then they were categorized as an inconsistent condom user.

Comparisons of characteristics between cases and controls were analyzed using Chi-square or Fisher's exact tests for categorical variables to assess factors associated with seroconversion. To establish a best-fitting model that characterizes HIV risk in the studied population, we conducted conditional logistic regression to assess factors independently associated with HIV seroconversion. Variables found to have a significant relationship with seroconversion (p < 0.1) in bivariate analyses were included in an adjusted model; variables were then evaluated for significance using backward elimination (using a p < 0.05 as criteria to remain in the model) to obtain a parsimonious best-fitting model. All statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC).

Results

A total of 18,878 unique women sought care at FCBOH SHC between 2011 and 2016. Five-hundred three women (3%) were excluded due to being HIV positive before their first encounter with the SHC during this timeframe of interest (Fig. 1). Ninety-four more women (0.5%) were excluded for having an HIV diagnosis date with the same month and year as their first encounter at the clinic. Of the remaining 18,281 women who had confirmed HIV-negative status at their initial visit to the SHC, 110 (0.6%) seroconverted between 2011 and 2018. Of these, 30 (27%) were excluded because they did not have CAV on record with the SHC before the date of HIV diagnoses. From the remaining 18,171 women who remained HIV negative throughout the study timeframe, 3753 (21%) were ineligible for control selection because they did not have a CAV on their first visit to the SHC in the timeframe of interest and the remaining 14,418 women comprised the pool of eligible controls. Two controls meeting the matching criteria for each case were then randomly selected from a pool of eligible controls, resulting in a total of 160 controls.

FIG. 1.

Flow diagram for inclusion and exclusion of participants. CAV, clinician-assisted visit; FCBOH, Fulton County Board of Health.

Compared to controls, cases were more likely at baseline to have more gonorrhea episodes (p = 0.09), more syphilis episodes (p = 0.04), a greater number of sex partners in the past 2 months (p = 0.02), anal sex (p = 0.05), history of injection drug or crack cocaine use (p < 0.0001), history of exchanging drugs/money for sex (p = 0.02), and heterosexual sex with more than one sex partner in the last month (p = 0.05; Table 1).

Table 1.

Clinical Characteristics of 80 Incident HIV Cases Among Women and 160 Matched HIV-Negative Controls, Fulton County Board of Health (Atlanta, GA)

Characteristic	Cases (n = 80)	Controls (n = 160)	OR (95% CI)	p
Characteristic	n (%)	n (%)	OR (95% CI)	p
Age^a	32 (20)	31 (19)
Race
Black/African American	74 (93)	148 (93)
White	6 (7)	12 (7)
Ethnicity
Hispanic	1 (1)	6 (4)	—	0.43
Non-Hispanic (ref)	79 (99)	154 (96)	1.0
Number of gonorrhea episodes
0 (ref)	43 (55)	108 (69)	1.0	0.09
1	25 (32)	39 (25)	1.6 (0.9–3.0)
2	8 (10)	8 (5)	2.9 (0.9–9.4)
3	2 (3)	1 (1)	5.2 (0.5–59.5)
Number of chlamydia episodes
0 (ref)	36 (48)	80 (52)	1.0	0.87
1	30 (40)	51 (33)	1.3 (0.7–2.5)
2	6 (8)	15 (10)	1.0 (0.3–3.0)
3	3 (4)	6 (4)	0.9 (0.2–3.8)
4	0 (0)	1 (1)	—
Number of trichomoniasis episodes
0 (ref)	47 (64)	105 (68)	1.0	0.63
1	20 (27)	38 (24)	1.2 (0.6–2.5)
2	4 (6)	10 (6)	1.0 (0.3–3.8)
3	1 (1)	1 (1)	—
4	1 (1)	0 (0)	—
≥5	1 (1)	1 (1)	2.0 (0.1–32.0)
Number of bacterial vaginosis episodes
0 (ref)	51 (67)	98 (66)	1.0	0.88
1	16 (21)	31 (21)	0.9 (0.4–1.9)
2	2 (3)	7 (5)	0.5 (0.1–2.6)
3	2 (3)	7 (5)	0.6 (0.1–2.9)
4	2 (3)	2 (1)	1.9 (0.3–13.9)
≥5	3 (3)	4 (2)	1.2 (0.3–6.0)
Number of syphilis episodes
0 (ref)	70 (89)	152 (96)	1.0	0.04
1	9 (11)	7 (4)	2.8 (1.0–7.9)
Sex with
Both	4 (5)	4 (2)	2.0 (0.5–8.0)	0.58
Females	2 (2)	3 (2)	1.4 (0.2–11.1)
Males (ref)	74 (93)	153 (96)	1.0
Number of sex partners in the past 2 months
0	8 (10)	10 (6)	2.2 (0.8–6.0)	0.02
1 (ref)	46 (58)	111 (70)	1.0
2	12 (15)	23 (14)	1.1 (0.5–2.6)
3	2 (3)	10 (6)	0.5 (0.1–2.5)
4	4 (5)	1 (1)	9.5 (1.0–88.8)
≥5	7 (9)	4 (3)	4.9 (1.2–20.3)
Number of sex partners in the past 12 months
0	0 (0)	2 (1)	—	0.38
1 (ref)	32 (41)	77 (49)	1.0
2	20 (25)	44 (28)	1.1 (0.6–2.0)
3	9 (11)	15 (9)	1.5 (0.6–3.8)
4	5 (6)	6 (4)	2.2 (0.6–7.6)
≥5	13 (17)	14 (9)	3.1 (1.1–8.4)
Consistent condom use
No (ref)	72 (90)	145 (91)	1.0	0.88
Yes	8 (10)	15 (9)	1.0 (0.4–2.7)
Vagina to penis sex
No	1 (1)	4 (2)	0.5 (0.1–4.5)	0.67
Yes (ref)	79 (99)	156 (98)	1.0
Anal sex
No (ref)	68 (86)	148 (94)	1.0	0.05
Yes	11 (14)	10 (6)	2.2 (0.9–5.4)
Sexual contact with bisexual male
No (ref)	79 (99)	159 (99)	1.0	1.00
Yes	1 (1)	1 (1)	2.0 (0.1–32.0)
Sexual contact with injection drug user
No (ref)	80 (100)	160 (100)	1.0	1.00
Yes	0 (0)	0 (0)	—
Sexual contact with known HIV positive person
No (ref)	80 (100)	160 (100)	1.0	1.00
Yes	0 (0)	0 (0)	—
Injecting drug or crack cocaine use
No (ref)	69 (86)	158 (99)	1.0	<0.0001
Yes	11 (14)	2 (1)	20.5 (2.6–159.8)
Exchanged drugs/money for sex
No (ref)	71 (89)	155 (97)	1.0	0.02
Yes	9 (11)	5 (3)	3.6 (1.2–10.7)
Heterosexual sex with >1 sex partner in last month
No (ref)	51 (64)	121 (76)	1.0	0.05
Yes	29 (36)	39 (24)	1.8 (1.0–3.2)

Cases and controls matched by age, race, and date of first clinician-assisted visit. “—” indicate ORs <0.001.

Reported as median (IQR).

CI, confidence interval; IQR, interquartile range; OR, odds ratio.

In bivariate analyses, all of the aforementioned factors were independently associated with HIV seroconversion (Table 1). In a fully adjusted model containing all variables found to be independently associated with HIV, the factors most strongly associated with HIV were injection drug/crack cocaine use [odds ratio (OR) = 23.7, 95% confidence interval (CI): 2.64–230.1], having exchanged drugs or money for sex (OR = 2.2, 95% CI: 0.5–10.8), syphilis episodes (OR = 4.2, 95% CI: 1.1–15.4), and anal sex (OR = 3.0, 95% CI: 0.9–9.4; Table 2). Model selection using backward elimination resulted in a model that included three variables, number of syphilis episodes, anal sex, and injection drug or crack cocaine use, and the following formula:

Table 2.

Multivariate Associations of Risk Factors for HIV Seroconversion Among Women Seeking Care at Fulton County Board of Health Sexual Health Clinic, 2011–2016

Risk factor	Fully adjusted OR (95% CI)	Final fitted model OR (95% CI)
Number of gonorrhea episodes	1.3 (0.8–2.1)	—
Number of syphilis episodes	4.2 (1.1–15.7)	4.7 (1.3–16.3)
Number of sex partners in the past 2 months	1.0 (0.7–1.4)	—
Anal sex	3.0 (0.9–9.4)	2.8 (1.0–8.1)
Injecting drug or crack cocaine use	23.7 (2.4–230.1)	33.5 (3.6–313.3)
Exchanged drugs/money for sex	2.2 (0.5–10.8)	—
Heterosexual sex with >1 sex partner in last month	1.2 (0.5–2.5)	—

CI, confidence interval; OR, odds ratio.

\begin{matrix} l n (o d d s H I V s e r o c o n v e r s i o n) \\ = 1.540 (s y p h i l i s e p i s o d e s) + 1.031 (a n a l s e x) \\ + 3.511 (i n j e c t i o n d r u g o r c r a c k c o c a i n e u s e) . \end{matrix}

Using this model, women who had a history of syphilis, had anal sex, and used injection drugs or crack cocaine were 6.1 times more likely to HIV seroconvert than women who did not have any of these risk factors.

Discussion

The objective of this study was to identify factors associated with HIV seroconversion among women seeking care at the FCBOH SHC from 2011 to 2016 to inform efforts to more accurately identify which women within this population are at greatest risk for HIV. We found that having a history of syphilis, anal sex, and injection drug or cocaine use were the strongest factors for HIV; women having all of those risk factors were six times more likely to seroconvert than similar women without any of those factors.

Our study results align with findings from previous epidemiological research demonstrating the strong role of STI history and injection drug or crack cocaine use in HIV acquisition among persons at risk.^2,19,21,22 In addition, previous studies have found that unprotected anal intercourse is associated with increased risk of HIV; an estimated 40% of HIV cases among 18- to 34-year-old women is due to anal sex.³

Where our study deviates from previous studies is in regard to risk posed by exchanging drugs or money for sex. While this has been a significant risk factor in other studies of HIV acquisition, this variable failed to maintain significance in a model adjusting for other risk factors.⁶ One hypothesis for why this could be is that other studies that have explored behavioral risk factors for HIV or STIs have obtained their study participants from more community-based venues rather than a health clinic. In a community-based study where women were recruited from their residences in New York City, heterosexual women who exchange money or goods for sex were found to be at high risk for HIV acquisition.²⁴ Another study based in Houston, Texas, where they recruited participants from fast-food restaurants, residences, bars, street corners, and post offices, also found that exchanging money or drugs for sex leads to high-risk sexual behaviors, and therefore results in increased risk for HIV infection.²⁵ The Centers for Disease Control and Prevention (CDC) assert that persons who exchange money/drugs for sex are less likely to seek health care because they are unsure where to access services.²⁶ Furthermore, since there is a strong correlation between those who exchange money or drugs for sex and those who are injection drug or crack cocaine users,²⁶ our best-fit model may still encompass women who engage in exchanging money or drugs for sex that way.

A study published in 2018 by Sales et al. established that there was no valid HIV risk assessment tool for identifying women who are at high risk for HIV seroconversion.¹⁰ This study hopefully contributes to creating such risk assessment tools. The CDC's current guidance for identifying candidates for preexposure prophylaxis (PrEP) leaves out important risk factors unique to women, such as substance abuse, gender-based violence, and intimate partner violence.¹⁰ While we lacked data on violence for the women included in this sample, that is an area worth exploring, perhaps by adding questions about violence and abuse into the series of questions that SHC clinicians ask during a CAV. Knowledge of cis- or transgender status would be another important variable to include in future analysis of HIV risk among women as well, since being transgender is a known risk factor for HIV acquisition.²⁷

While PrEP awareness and promotion for MSM has increased rapidly in the Southern United States as a result of the current HIV epidemic, the few studies that have focused on or included women show low PrEP awareness.¹⁰ A 2018 study by Koren et al. showed that a majority of at-risk women in Philadelphia, PA attending a university family planning gynecology clinic had limited knowledge of PrEP, and were most concerned about the cost and side-effects of the prophylaxis medication.²⁸ Koren et al. concluded that women's beliefs and knowledge of PrEP is not congruent with their HIV risk and need for PrEP, and therefore PrEP prevention education must be expanded to women in high risk communities.²⁸ More proximally, a 2018 survey conducted among attendees at two major Atlanta gay pride events in 2018 found that while 90% of gay and bisexual men were aware of PrEP, only 36% of heterosexual women were aware of it (article under review). There are dire consequences to this lack of awareness; other studies are beginning to quantify the number of new HIV diagnoses that could be prevented among women had they only been provided with PrEP at the time key risk factors (such as STI diagnoses) were realized.²³

Efforts to promote PrEP to women are needed, especially in Fulton County, so the results of this study not only get us closer to identifying women at greatest risk of HIV but also quantify that increased risk. This quantification would allow clinicians to tell female patients how much more likely they are to acquire HIV because of certain risk factors. This awareness may then be helpful in affecting change in risk behaviors or accepting new preventive health measures such as PrEP among the most at-risk women. While we lacked data on PrEP counseling or use for the majority of women in this study sample since in-house PrEP provision was not routinized at FCBOH until March 2016, counseling is now incorporated into all HIV screening encounters and CAVs for HIV-negative women who are at increased risk of HIV per the CDC PrEP guidelines.²⁹

A further complication to identifying women at greatest risk for HIV is the overall reluctance of women within this demographic to get tested for HIV. A 2019 study by Cheong et al. that focused on reasons for young African American women living in the South to accept or decline free HIV testing and counseling found that 30% of women declined the offer.³⁰ The main reasons for declination were negative consequences of a positive result and concern for privacy, identifying themselves as low-risk for HIV, and concern of social rejection upon receipt of a positive test result.³⁰ Low perceived HIV risk was the most influential of the three reasons, a finding which again underscores the need for better education on the factors that put women both at risk for HIV and strong candidates for PrEP. Another Atlanta-based study of youth which echoed these findings emphasized the need for community-based HIV testing to be implemented in ways that encourage at-risk individuals to get tested in environments that offer discretion, confidentiality, support, and incentives.^30,31 Knowing these sensitivities around HIV testing are integral to making women feel comfortable both disclosing risk behaviors and being receptive to learning about risk; the HIV counseling now enhanced by the ability to offer free PrEP within the privacy of clinical encounters at FCBOH will hopefully assuage some women's concerns and reluctance toward HIV knowledge generally.

This study has some important limitations. First, as with any case–control study, biases are inherent with the selection of cases. Among our initial cohort of women, there may have been additional seroconversions among women who may have moved out of state. We were only able to verify HIV statuses among women who remained in Georgia during the period of interest. With that said, because we were able to verify the HIV status of all women in the initial cohort through a cross-check with Georgia's eHARs database, we feel confident in the HIV statuses of women selected for our analyses, assuming that they remained Georgia residents for the duration of follow-up.

A related limitation is that in attempting to obtain risk factor data before the HIV diagnosis date, we had to exclude more than one-quarter of seroconversions because they lacked record of a CAV before their diagnosis date. Because no risk factor data are available in SHC medical records for non-CAV visits, we were unable to assess whether and how these women may have differed from those included in the final analytic sample. In addition, we can only draw conclusions on women who seek care at FCBOH or similar safety net health clinics which have high proportions of homeless individuals and individuals of lower socioeconomic statuses compared to other health care venues. While identification of subjects from any health clinic could have biased against women working in the sex trade, the population this clinic serves includes the women at greatest risk for HIV in Atlanta: young, disadvantaged, heterosexual black women.¹⁰

In addition, we only looked at one snapshot in time. We obtained our cases and controls from 2011 until 2016, which means some of the women we excluded for only coming into FCBOH once could have actually been returning clients, having had their first visit before our timeframe of interest. Finally, all variables related to sexual behaviors, history of STIs, and drug use captured during a CAV visit were self-reported by the patients, introducing possible recall and social desirability biases. Because the data collected on these sensitive topics were obtained during a clinical encounter and before the date of HIV diagnosis (for cases), we hope the timing and confidentiality of this setting contributed to patients' willingness to report truthfully on these topics than had the data been collected in other more public venues.

To echo Hodder et al., further research is necessary to find successful interventions that decrease women's HIV risk in the United States, and these interventions need to be feasible to target vulnerable populations, such as black women.⁸ Before interventions can be used, HIV risk needs to be ascertained, and our findings contribute to the growing interest in utilizing clinically accessible risk factor data to identify the best candidates for PrEP.³²

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

There was no funding provided for this article.

References

Schuman

, Sobel

. Women and AIDS. J Obstet Gynaecol, 1993; 33:341–350.

Centers for Disease Control and Prevention. HIV/AIDS: HIV Among Women. 2019. Available at: https://www.cdc.gov/hiv/group/gender/women/index.html (Last accessed January 22, 2020 ).

Evans

, Tao

, Porter

, et al. Low HIV testing rates among U.S. women who report anal sex and other HIV sexual risk behaviors, 2011–2015. Am J Obstet Gynecol, 2018; 219:383.e1–383.e7.

Haley

, Lucas

, Golin

, et al. Retention strategies and factors associated with missed visits among low income women at increased risk of HIV acquisition in the US (HPTN 064). AIDS Patient Care STDS, 2014; 28:206–217.

Breskin

, Adimora

, Westreich

. Women and HIV in the United States. PLoS One, 2017; 12:1–6.

Meyer

, Springer

, Altice

. Substance abuse, violence, and HIV in women: A literature review of the syndemic. J Womens Heal, 2011; 20:991–1006.

Walcott

, Kempf

, Merlin

, et al. Structural community factors and sub-optimal engagement in HIV care among low-income women in the Deep South of the USA. Cult Health Sex, 2016; 18:683–696.

Hodder

, Justman

, Haley

, et al. Challenges of a hidden epidemic: HIV prevention among women in the United States. J Acquir Immune Defic Syndr, 2010; 55:S69–S73.

Davidson

. African Americans and HIV/AIDS-the epidemic continues: An intervention to address the HIV/AIDS pandemic in the black community. J Black Stud, 2011; 42:83–105.

10.

Sales

, Steiner

, Brown

, et al. PrEP eligibility and interest among clinic- and community-recruited young black women in Atlanta, Georgia, USA. Curr HIV Res, 2018; 16:250–255.

11.

McCree

, Chesson

, Bradley

ELP

, et al. US Regional Changes in Racial/Ethnic Dispartities in HIV Diagnoses Among Women in the United States, 2012 and 2017. AIDS Behav, 2019; 1–6.

12.

Darlington

, Hutson

. Understanding HIV-related stigma among women in the Southern United States: A literature review. AIDS Behav, 2017; 21:12–26.

13.

Hixson

, Omer

, Rio

, et al. Spatial clustering of HIV prevalence in Atlanta, Georgia and population characteristics associated with case concentrations. J Urban Heal, 2011; 88:129–141.

14.

Forna

, Fitzpatrick

, Adimora

, et al. A case-control study of factors associated with HIV infection among black women. J Natl Med Assoc, 2006; 98:1798–1804.

15.

Onwubiko U. #StopHIVATL: An epidemiological overview and roadmap to ending the crisis. Oral Presentation at Fulton County Board of Health World AIDS Day Breakfast, Atlanta, GA, 2018.

16.

Fauci

, Redfield

, Sigounas

, et al. Ending the HIV epidemic: A plan for the United States. J Am Med Assoc, 2019; 321:844–845.

17.

Chen

, Clarke

, Ou

, et al. Antiretroviral drug use in a cohort of HIV-uninfected women in the United States: HIV prevention trials network 064. PLoS One, 2015; 10:1–9.

18.

Gordon

, Edelman

, Justice

, et al. HHS public minority men who have sex with men demonstrate increased risk for HIV transmission. AIDS Behav, 2017; 21:1497–1510.

19.

Dias

, Mbofana

, Cassy

, et al. Estimating risk factors for HIV infection among women in Mozambique using population-based survey data. African J AIDS Res, 2018; 17:62–71.

20.

Teitelman

, Campbell

. Voices from the frontlines: The epidemics of HIV/AIDS and violence among women and girls. Heal Care Women Int, 2009; 30:184–194.

21.

Centers for Disease Control and Prevention. 10 Ways STDs Impact Women Differently from Men. 2011. Available at: https://www.cdc.gov/std/health-disparities/stds-women-042011.pdf (Last accessed January 22, 2020 ).

22.

Nacius

, Levison

, Minard

, et al. Serodiscordance and disclosure among HIV-positive pregnant women in the Southwestern United States. AIDS Patient Care STDS, 2013; 27:242–247.

23.

Newman

, Rahman

, Brantley

, et al. Rates of new HIV diagnoses after Reported STI, Women in Louisiana 2000–2015: Implications for HIV Prevention. Clin Infect Dis, 2019; 1–6.

24.

Jenness

, Kobrak

, Wendel

, et al. Patterns of exchange sex and HIV infection in high-risk heterosexual men and women. J Urban Heal, 2011; 88:329–341.

25.

Baseman

, Ross

, Williams

. Sale of sex for drugs and drugs for sex: An economic context of sexual risk behavior for STDs. Sex Transm Dis, 1999; 26:444–449.

26.

Centers for Disease Control and Prevention. HIV/AIDS: HIV Risk Among Persons Who Exchange Sex for Money or Nonmonetary Items. 2018. Available at: https://www.cdc.gov/hiv/group/sexworkers.html (Last accessed January 22, 2020 ).

27.

Centers for Disease Control and Prevention. HIV and Transgender People. 2019. Available at: https://www.cdc.gov/hiv/group/gender/transgender/index.html (Last accessed January 22, 2020 ).

28.

Koren

, Nicols

, Simoncini

. HIV pre-exposure prophylaxis and women: Survey of the knowledge, attitudes, and beliefs in an urban obstetrics/gynecology clinic. AIDS Patient Care STDS, 2018; 32:490–494.

29.

Centers for Disease Control and Prevention. Preexposure prophylaxis for the prevention of HIV infection in the United States—2017 update. 2018. Available at: https://www.cdc.gov/hiv/pdf/risk/prep/cdc-hiv-prep-guidelines-2017.pdf (Last accessed January 22, 2020 ).

30.

Cheong

, Tucker

, Chandler

. Reasons for accepting and declining free HIV testing and counseling among young African American women living in disadvantaged Southern Urban communities. AIDS Patient Care STDS, 2019; 33:25–31.

31.

Murray

, Hussen

, Toledo

, et al. Optimizing community-based HIV testing and linkage to care for young persons in Metropolitan Atlanta. AIDS Patient Care STDS, 2018; 32:234–240.

32.

Gilkey

, Marcus

, Garrell

, et al. Using HIV risk prediction tools to identify candidates for pre-exposure prophylaxis: Perspectives from patients and primary care providers. AIDS Patient Care STDS, 2019; 33:372–378.

Factors Associated with HIV Seroconversion Among Women Attending an Urban Health Clinic in the South: A Matched Case–Control Study