Partnership types and coital frequency as predictors of gonorrhea and chlamydia among young MSM and young transgender women

Abstract

Background

Sexually transmitted infections pose a major public health challenge in the United States and this burden is especially acute in subpopulations like young men who have sex with men (YMSM) and young transgender women (YTW). Yet, the direct behavioral antecedents of these infections are not well understood making it difficult to identify the cause of recent increases in incidence. This study examines how variations in partnership rates and the number of condomless sex acts are associated with STI infections among YMSM-YTW.

Method

This study leveraged 3 years of data from a large longitudinal cohort of YMSM-YTW. A series of generalized linear mixed models examined the association between the number of condomless anal sex acts, number of one-time partners, number of casual partners, and number of main partners and chlamydia, gonorrhea, or any STI.

Results

Results indicated the number of casual partners was associated with gonorrhea [aOR = 1.17 (95% CI: 1.08, 1.26)], chlamydia [aOR = 1.12 (95% CI: 1.05, 1.20)], and any STI [aOR = 1.14 (95% CI: 1.08, 1.21)] while the number of one-time partners was only associated with gonorrhea [aOR = 1.13 (95% CI: 1.02, 1.26)]. The number of condomless anal sex acts was not associated with any outcome.

Conclusion

These findings suggest the number of casual partners is a consistent predictor of STI infection among YMSM-YTW. This may reflect the quick saturation of risk within partnerships making the number of partners, rather than the number of acts, the more relevant factor for STI risk.

Keywords

chlamydia (chlamydia trachomatis)epidemiology gonorrhea (neisseria gonorrhoeae)sexual behaviour

Introduction

Sexually transmitted infections (STIs) continue to pose a major public health challenge in the United States.¹ Rates of these infections are especially elevated, and recent increases more pronounced, in subpopulations such as men who have sex with men (MSM).² In addition, transgender women (TW) are at significant risk for NG and CT and experience suboptimal STI screening and treatment.³ The recent drivers of rising STI rates among MSM and TW are not well understood,⁴ although adoption of smartphone apps⁵ and uptake of biomedical approaches to HIV⁶ have been identified as two potential reasons. However, less emphasis has been placed on identifying the direct behavioral antecedents of STI infection. Identification of these behaviors is essential for isolating the causes for this recent increase, would benefit behavioral surveillance, and could help target resources towards individuals at the highest risk for STIs, especially if low-threshold and cost-effective risk assessment methods could be developed.

Risk of STI acquisition is a function of the total number of partners one has, the probability that each partner has an STI (which may change over the course of a relationship), the number of sex acts per partner, and the per-act transmission probability of a given STI. These factors have a non-linear relationship with infection probability⁷ and the relative contribution of each factor likely differs across subgroups. For example, while the number of partners^8–13 and number of sex acts^10,13,14 have both been associated with increased risk of STIs among MSM, the independent contribution of each of these factors to incident STIs is not well understood. Furthermore, STI risk prediction models in general population studies have found the number of sexual partners is more consistently associated with STI diagnoses than condom use within any partnership.¹⁵

The mechanisms are explained by mathematical modeling studies: highly transmissible STIs, such as Neisseria gonorrhoeae (NG) and Chlamydia trachomatis (CT), are likely to transmit after few sex acts, which leads to infection risk quickly saturating within partnerships after a small number of acts.⁷ Accordingly, the number of partners may be more predictive of STI acquisition as higher partnership rates increase the likelihood of coming into contact with an STI. In addition, all partnerships are not equal and these predictive factors may vary substantially across partnership types. For example, MSM are less likely to use condoms with main partners and report more sex acts with these partners compared to casual partners.^16,17 In contrast, MSM may be more likely to use condoms with one-time or other casual partners¹⁷ and the characteristics of these partners may also differ compared to main partners (e.g., age of partners).¹⁸ Accordingly, partnership types may provide additional information regarding the relationship between partnership rates, sex acts, and STI risk.

The current study seeks to quantify these behaviors (i.e., partnership rates and number of condomless sex acts) and their associations with NG and CT infections among a large and diverse cohort of young MSM and young TW (YMSM-YTW). While initial results indicating the high prevalence of NG and CT in this cohort have already been published,¹⁹ the current analyses extends this work by further disentangling behavioral risk factors for STIs, focusing on the independent contribution of these two behaviors, separately examining multiple partnership types (i.e., one-time, casual, and serious), using data from this large longitudinal cohort which enables within-person estimates of these associations.

Method

Participants and procedures

Data for this analysis comes from the RADAR cohort, a longitudinal cohort study of YMSM-YTW in Chicago.²⁰ Enrollment criteria for this cohort included being 16–29 years old at enrollment, being assigned male at birth (TW and other non-cisgender identified participants were eligible for enrollment), speaking English, reporting sex with a man in the previous year or identifying as gay or bisexual, and being able to attend in-person research visits in Chicago. Data for the current analysis were collected from February 2015 through March 2020. Participants in the current study included all active cohort members with a baseline visit by March 2017 who were therefore eligible for a 7^th visit (i.e., 3-years follow-up) by March 2020. A total of 868 participants were included in this analysis, of whom 741 (77.8%) had completed a 7^th visit by March 2020. Baseline observations and every yearly (i.e., 12 months, 24 months, 36 months) visits were used in the current analysis for all eligible participants, as these visits included assays for prevalent STIs.

Measures

Data on sexual partnerships were collected as part of an interviewer-assisted sexual network survey. Data were collected on all sexual partners in the past 6 months. Following elicitation, detailed information on each dyadic partnership was collected, including number of anal sex acts, number of condomless anal sex acts, relationship type, and dates of first and last sex.

For the outcome variables, participants were tested for urethral and rectal STIs on a yearly basis. Rectal infections were screened via nucleic acid amplification (NAAT) testing via self-administered rectal swabs and urogenital infections were screened via urine testing.¹⁹ Participants testing positive for any STI were referred to treatment.

For the exposure variables, anal sex partners were classified as one of three types at each visit: main, casual, and one-time. Main partners were defined as any partner that the participant indicated they were “currently in a serious relationship.” Casual partners were non-main partners where sex occurred on more than 1 day, while one-time partners were non-main partners who had the same first and last date of sex, with no prior reports of sex with this partner. One-time partnerships were considered separately from casual partners as they could accrue more rapidly and may have had different rates of condom use than consistent casual partners. For all partnership types, partnership counts reflected the cumulative number of each type in the 6 months before the study visit. Number of condomless sex acts reflected the average monthly number of condomless sex acts across all partners during that same 6-month period (i.e., dividing the total number of condomless sex acts across all partners by 6).

Finally, as sexual behavior may also vary as a function of HIV and PrEP use status, we include these as control variables in our analysis. Participant HIV status was based on screening at their previous study visit (i.e., 6 months prior to each interview) or self-reported HIV positive test result since the last visit, thereby reflecting the participant’s perceived HIV status of during the period of reported behavior. HIV infection was measured using Alere Determine 4^th-Generation HIV-1/HIV-2 Ab/Ag Combo test with confirmatory testing conducted according to CDC guidance.²¹ For each visit, participants were categorized as HIV-negative not on PrEP, HIV-negative on PrEP (i.e., if the participant was HIV negative and indicated taking any PrEP in the last 6 months), and HIV-positive.

Analysis

Our analytic objective was to estimate the association between prevalent STIs at each study period and two types of exposure variables (partnership rates and condomless sex act counts) in the 6 months preceding the infection. We used a series of generalized linear models with binomial distributions and random intercepts to account for multiple observations for each participant. We first examined the bivariate associations between each exposure variable and NG, CT, or any NG/CT STI. We then used multivariable modeling that included all predictor (exposure) variables for the purpose of controlling for potential confounding. In our main analysis, we use the combined participants of all gender identities, but also provide stratified results for YMSM and YTW, except for participants of gender identities other than TW and cisgender men due to small sample size. Missing data were present in 14.3% of observations. Most observations with missing data (90.3%) were missing both outcome and exposure variables, reflecting a missed study visit. Examining demographic variables (i.e., age, gender, race/ethnicity, and gender identity) and visit number as correlates of missingness, the only significant association was lower levels of missing data at the first visit. For both exposure and outcome variables, missing data were handled using multiple imputation by chained equations using predictive mean matching for count variables, logistic regression for dichotomous variables, and polytomous regression imputation for unordered categorical variables.^22,23 Unless specified, all results reflect the pooled estimates using Rubin’s rule²⁴ to aggregate estimates across models from multiply imputed data.

Results

Table 1 presents demographic and descriptive statistics on exposure and outcome variables for all participants included in the current analysis (n = 868). CT was more prevalent than NG, with positive results in 313 (10.1%) and 223 (7.4%) of observations, respectively. Nearly all NG (95.7%) and CT (93.2%) infections were rectal, rather than urethral. NG infections were concentrated among 179 participants of whom 37 experienced more than one observation with NG; comparable numbers for CT infections were 250 and 52. Figure 1 displays all participants with repeat infections of NG (1a) or CT (1b) to further illustrate patterns of infection across visits.

Table 1.

Descriptive statistics on demographics, behavior, and STIs.

	Mean (SD)	n (%)
Participants^a
Age	21.3 (2.9)	—
Race/Ethnicity
Black/African American	—	309 (35.6)
White	—	198 (22.8)
Hispanic/Latino	—	265 (30.5)
Other	—	96 (11.1)
Gender identity
Cisgender men	—	800 (92.4)
Transgender women	—	48 (5.5)
Other	—	20 (2.3)
HIV Status
HIV Negative no PrEP	—	711 (81.9)
HIV Negative on PrEP	—	49 (5.6)
HIV Positive	—	108 (12.4)
Visits
Number of CAS acts	2.39 (5.8)	—
One-time partners	0.57 (1.2)	—
Casual partners	0.95 (1.7)	—
Main partners	0.35 (0.5)	—
Gonorrhea	—	223 (7.3)
Urethral	—	10 (1.0)
Rectal	—	221 (7.4)
Chlamydia	—	313 (10.3)
Urethral	—	22 (2.0)
Rectal	—	302 (10.1)
Coinfection	—	72 (2.3)
Any STI	—	464 (15.3)

^aAt baseline.

For brevity, we focus our discussion on results on the combined analysis but note again stratified analyses by gender identity can be found in Table 2. For NG bivariate analyses, no evidence of an association was observed for overall number of CAS acts [OR = 1.00 (95% CI: 0.98, 1.03)] or main partner counts [OR = 0.92 (95% CI: 0.68, 1.23)]. However, the number of one-time partners [OR = 1.12 (95% CI: 1.01, 1.24)] and casual partners [OR = 1.12 (95% CI: 1.04, 1.21)] was positively associated with NG. For CT bivariate analyses, no evidence of an association was observed for number of CAS acts [OR = 1.01 (95% CI: 0.98, 1.04)], one-time partners [OR = 0.99 (95% CI: 0.89, 1.11)], or main partners [OR = 0.83 (95% CI: 0.64, 1.08)]. Here too, however, the number of casual partners was positively associated with CT [OR = 1.08 (95% CI: 1.01, 1.16)]. For any STI bivariate analyses, no evidence of an association was observed for number of CAS acts [OR = 1.00 (95% CI: 0.99, 1.02)], one-time partners [OR = 1.06 (95% CI: 0.97, 1.16)], or main partners [OR = 0.87 (95% CI: 0.69, 1.11)]. Once again, the number of casual partners was positively associated with any STI [OR = 1.11 (95% CI: 1.05, 1.18)].

Table 2.

Mixed model results for association between exposure variables and sexually transmitted infections.

	Gonorrhea		Chlamydia		Any STI
	Bivariate OR [95% CI]	Multivariable aOR [95% CI]	Bivariate OR [95% CI]	Multivariable aOR [95% CI]	Bivariate OR [95% CI]	Multivariable aOR [95% CI]
All participants
Number of CAS acts	1.00 [0.98, 1.03]	1.01 [0.98, 1.04]	1.00 [0.98, 1.02]	1.00 [0.98, 1.03]	1.00 [0.99, 1.02]	1.01 [0.99, 1.03]
One-time partners	1.12 [1.01, 1.24]	1.16 [1.02, 1.32]	0.99 [0.89, 1.11]	0.93 [0.82, 1.05]	1.06 [0.97, 1.16]	1.04 [0.95, 1.15]
Casual partners	1.12 [1.04, 1.21]	1.15 [1.06, 1.25]	1.08 [1.01, 1.16]	1.13 [1.05, 1.22]	1.11 [1.05, 1.18]	1.15 [1.07, 1.23]
Main partners	0.94 [0.68, 1.29]	1.02 [0.72, 1.44]	0.83 [0.64, 1.08]	0.85 [0.64, 1.14]	0.87 [0.69, 1.11]	0.92 [0.71, 1.18]
Young men who have sex with men
Number of CAS acts	1.01 [0.98, 1.04]	1.01 [0.98, 1.04]	1.00 [0.97, 1.02]	1.01 [0.98, 1.03]	1.00 [0.98, 1.02]	1.01 [0.99, 1.03]
One-time partners	1.16 [1.03, 1.31]	1.20 [1.06, 1.36]	0.96 [0.85, 1.08]	0.95 [0.84, 1.08]	1.07 [0.97, 1.18]	1.08 [0.98, 1.20]
Casual partners	1.13 [1.04, 1.23]	1.11 [1.02, 1.21]	1.09 [1.02, 1.17]	1.12 [1.04, 1.21]	1.12 [1.05, 1.19]	1.12 [1.05, 1.20]
Main partners	0.94 [0.66, 1.33]	1.00 [0.69, 1.45]	0.85 [0.63, 1.13]	0.88 [0.64, 1.20]	0.87 [0.67, 1.12]	0.92 [0.70, 1.21]
Young transgender women
Number of CAS acts	1.01 [0.93, 1.09]	0.99 [0.92, 1.07]	1.00 [0.94, 1.05]	0.99 [0.93, 1.05]	1.02 [0.97, 1.07]	1.01 [0.93, 1.09]
One-time partners	0.52 [0.13, 2.09]	0.42 [0.09, 2.04]	0.62 [0.29, 1.35]	0.55 [0.24, 1.25]	0.54 [0.26, 1.14]	0.52 [0.13, 2.09]
Casual partners	1.74 [1.05, 2.87]	1.54 [1.00, 2.35]	1.11 [0.84, 1.45]	1.04 [0.78, 1.39]	1.37 [1.03, 1.82]	1.74 [1.05, 2.87]
Main partners	0.88 [0.22, 3.60]	1.11 [0.30, 4.08]	0.74 [0.31, 1.74]	0.81 [0.32, 2.04]	1.09 [0.50, 2.40]	0.88 [0.22, 3.60]

Note: All multivariable models control for age, race/ethnicity, gender identity, and HIV status/PrEP use.

All multivariate models closely mirrored the bivariate results. For NG, the number of one-time [aOR = 1.16 (95% CI: 1.02, 1.32)] and casual partners [aOR = 1.15 (95% CI: 1.06, 1.25)] were both associated with a positive test result while there was no evidence of associations with the number of CAS acts [aOR = 1.01 (95% CI: 0.98, 1.04)] or number of main partners [aOR = 1.02(95% CI: 0.72, 1.44). For CT, number of casual partners was again associated with a positive test result [aOR = 1.13 (95% CI: 1.05, 1.22)] but there was no evidence of an association with number of CAS acts [aOR = 1.00 (95% CI: 0.98, 1.03)], one-time partners [aOR = 0.93 (95% CI: 0.82, 1.05)], or main partners [aOR = 0.85 (95% CI: 0.64, 1.14)]. Similarly, number of casual partners [aOR = 1.15 (95% CI: 1.07, 1.23)] was positively associated with any STI while number of CAS acts [aOR = 1.01 (95% CI: 0.99, 1.03)], one-time partners [aOR = 1.04 (95% CI: 0.95, 1.15)], or main partners [aOR = 0.92 (95% CI: 0.71, 1.18)] had no evidence of association.

Results of the stratified analysis (Table 2) largely mirrored the results of the combined analysis with a notable exception. One-time partners were not associated with any STIs among YTW with point estimates indicated an inverse association, although with extremely wide confidence intervals.

Figure 2 presents the predicted probability for any STI across observed ranges of each predictor. Predicted probabilities of any STIs varied from 7% (95% CI: 8%, 10%) at 0 casual partners to 28% (95% CI: 16%, 43%) at 11 casual partners. In contrast, minimal differences in the probability any STI were predicted by one-time partnership counts, from 9% (95% CI: 8%, 11%) at 0 one-time partners to 14% (95% CI: 5%, 32%) at 11 one-time partners. Similar minimal differences for any STI were predicted across the number of monthly CAS acts from 9% (95% CI: 8%, 11%) at 0 acts to 15% (95% CI: 6%, 30%) at 50 acts.

Discussion

In this study, we found that the number of casual and one-time partners, but not CAS acts or main partners, were associated with NG and CT infections among YMSM-YTW. This finding supports the notion that the number of one-time and casual partners is closely linked to the acquisition of STIs, as has been suggested from modeling studies^7,25 and risk prediction studies in the general population.¹⁵ The significant association between these partnerships and STIs may be attributable to the high per-act transmission probability of NG and CT. Given that transmission is likely to occur even with a small number of condomless sex acts within any partnership, acquisition of these infections may therefore be more closely associated with the likelihood of encountering a new partner with an STI (i.e., higher partnership rates), rather than frequency of exposure within partnership (i.e., number of condomless sex acts).

Interestingly, not all partnership types showed a significant correlation with STIs. We found an association between one-time partners and NG but no evidence of an association with either CT or any STI. Contrastingly, the number of casual sex partners was consistently associated with both STIs, while main partners were not consistently predictive. As noted in the introduction, the likelihood of acquisition of an STI is based on the probability a partner has an STI, the transmission probability per act, the number of sex acts, and the number of partners. Given the specific details of this population, it appears that the combination of these factors makes the number of casual partners most consistently associated with STIs. For example, the association between casual partners and STIs could be driven by casual partners differing in their likelihood to have an STI compared to other partner types. The difference in predictors across NG and CT could also be related to the estimated higher per-act transmission probability for NG compared to CT.²⁶ Of note, there is also considerable uncertainty in our estimates, particularly at the highest level of partners and sex acts, and the failure to find significant associations between these behaviors and STIs is not sufficient evidence of their absence.

Nonetheless, given the substantial and increasing burden of NG and CT in the US,¹ it is essential for research to further clarify the main pathways to STI acquisition¹⁵ and improve targeting of prevention resources. For example, few studies have rigorously validated risk prediction to optimize targeting screening^15,27 and many studies have struggled to replicate findings on external datasets.^27,28 In addition, evidence is especially lacking for asymptomatic STIs²⁷ because many studies rely on individuals seeking care at sexual health clinics that are less likely to capture these infections, which are common for rectal STIs.²⁹ Nonetheless, our findings concur with prior reviews indicating that number of sexual partners is a consistent predictor of STIs.¹⁵ Accordingly, these findings suggest future studies should continue to explore characteristics that are most likely to measure exposure to partners with STIs rather than the level of exposure within partnerships. For example, casual partners with other concurrent partnerships may be most likely to facilitate the spread of STIs.³⁰ Documenting such features would help with the identification of individuals at highest risk for STI and also inform targeting screening efforts among MSM and TW. We do note that traditional covariates, such as HIV status – although not PrEP status, were also predictive of each STI and will continue to be important for identifying individuals at highest risk for STIs.

In addition, most sexual risk reduction interventions measure CAS acts to initially demonstrate their efficacy³¹ due to lower statistical power to detect incident STIs and ease of measuring self-reported behavior change. The current findings, combined with other similar results, suggest a potential disconnect in the behavioral versus biological outcomes of such interventions for NG and CT if the number of sex acts is used as the primary behavioral outcome, as condomless sex acts may not be predictive of STI infections. Accordingly, these results suggest casual partnerships may be a more useful proxy for STI risk when measuring STIs directly is infeasible.

The current study had several limitations. First, our analysis included urethral and rectal STIs but could not distinguish between insertive and receptive anal sex acts, because this was only collected on the four most recent partners. Accordingly, mismatches between these variables could attenuate the association between number of condomless sex acts and infections. We thus conducted an additional analysis using data on the number of receptive anal sex acts and rectal infections with the four most recent partners. Multivariate models confirmed the findings of our primary analysis with casual partners being consistently associated with all rectal STI outcomes while condomless rectal sex acts had no association. This suggests these results were not substantially impacted by measurement error in CAS acts due to sexual positioning.

Second, data were collected in the context of an ongoing longitudinal cohort. Yet, these data may be most beneficial in other contexts, such as in health care clinics. However, it is unclear if patients would be willing to provide this information in these settings. For example, sexual history taking in places like primary care offices occurs at low rates,³² particularly among MSM.³³ Accordingly, to the extent that similar behavioral data would be useful to improve STI testing, the identification of optimal methods to collect these data in those settings is urgently needed.

In summary, we found that the number of casual partners was consistently associated with both NG and CT, while the number of one-time partners was associated with NG. These findings provide additional evidence of the association between partnership rates and STIs among YMSM-YTW, raise important questions about the best way to capture risk for STIs in this population, and emphasize the importance of careful measurement of behavior change for STI prevention interventions. Future studies should continue efforts to disentangle the mechanisms of NG and CT spread to best inform efforts to target resources identify distal causes of recent increases in STIs, and reduce the burden of STIs.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants from the National Institute of Allergy and Infectious Diseases (R01AI138783) and the National Institutes on Drug Abuse (U01DA036939) at the National Institutes of Health.

ORCID iD

Patrick Janulis

References

Scott-Sheldon

LAJ

Chan

. Increasing sexually transmitted infections in the U.S.: a call for action for research, clinical, and public health practice. Arch Sex Behav 2020; 49: 13–17.

Dewart

Bernstein

DeGroote

, et al. Prevalence of rectal chlamydial and gonococcal infections: a systematic review. Sex Transm Dis 2018; 45: 287–293.

Van Gerwen

Tamhane

Westfall

, et al. Prevalence of and factors associated with genital and extragenital chlamydia and gonorrhea among transgender women in HIV care in the United States, 2005 to 2016. Sex Transm Dis 2021; 48: 410–416.

Chow

EPF

Grulich

Fairley

. Epidemiology and prevention of sexually transmitted infections in men who have sex with men at risk of HIV. Lancet HIV 2019; 6: e396–e405.

DeVost

Beymer

Weiss

, et al. App-based sexual partner seeking and sexually transmitted infection outcomes: a cross-sectional study of HIV-negative men who have sex with men attending a sexually transmitted infection clinic in Los Angeles, California. Sex Transm Dis 2018; 45: 394–399.

Ramchandani

Golden

. Confronting rising STIs in the Era of PrEP and treatment as prevention. Curr HIV/AIDS Rep 2019; 16: 244–256.

Garnett

White

Ward

. Fewer partners or more condoms? Modelling the effectiveness of STI prevention interventions. Sex Transm Infect 2008; 84(Suppl 2): 4–11.

Callander

Guy

Fairley

, et al. Gonorrhoea gone wild: rising incidence of gonorrhoea and associated risk factors among gay and bisexual men attending Australian sexual health clinics. Sex Health 2019; 16: 457–463.

Chow

EPF

Tomnay

Fehler

, et al. Substantial increases in chlamydia and gonorrhea positivity unexplained by changes in individual-level sexual behaviors among men who have sex with men in an Australian sexual health service from 2007 to 2013. Sex Transm Dis 2015; 42: 81–87.

10.

Grov

Cain

Rendina

, et al. Characteristics associated with urethral and rectal gonorrhea and chlamydia diagnoses in a US National sample of gay and bisexual men: results from the one thousand strong panel. Sex Transm Dis 2016; 43: 165–171.

11.

Golub

Peña

Fikslin

, et al. Partners, not condom use, drive STI rates among PrEP users in community health center. Top Antivir Med 2018; 26: 468.

12.

Hassan

Blumenthal

Dube

, et al. Effect of rectal hygiene on sexually transmitted infections among HIV-negative men who have sex with men (MSM). Open Forum Infectious Diseases 2017; 4: S103.

13.

Goddard

Poynten

Petoumenous

, et al. Prevalence, incidence and predictors of anal Chlamydia trachomatis, anal Neisseria gonorrhoeae and syphilis among older gay and bisexual men in the longitudinal study for the prevention of anal cancer (SPANC). Sex Transm Infect 2019; 95: 477–483.

14.

Jin

Prestage

Mao

, et al. Incidence and risk factors for urethral and anal gonorrhoea and chlamydia in a cohort of HIV-negative homosexual men: the health in men study. Sex Transm Infect 2007; 83: 113–119.

15.

Falasinnu

Gilbert

Hottes

, et al. Predictors identifying those at increased risk for STDs: a theory-guided review of empirical literature and clinical guidelines. Int J Std Aids 2015; 26: 839–851.

16.

Sullivan

Salazar

Buchbinder

, et al. Estimating the proportion of HIV transmissions from main sex partners among men who have sex with men in five US cities. Aids 2009; 23: 1153–1162.

17.

Boyer

Greenberg

Korelitz

, et al. Sexual partner characteristics, relationship type, and HIV risk among a community venue-based sample of urban adolescent and young adult men who have sex with men. Youth Soc 2019; 51: 219–246.

18.

Janulis

Feinstein

Phillips

II , et al. Sexual partner typologies and the association between drug use and sexual risk behavior among young men who have sex with men. Arch Sex Behav 2018; 47: 259–271.

19.

Morgan

Dyar

Newcomb

, et al. PrEP use and sexually transmitted infections are not associated longitudinally in a cohort study of young men who have sex with men and transgender women in Chicago. Aids and behavior. 2020; 24(5): 1334–1341. Epub ahead of print 2019. DOI: 10.1007/s10461-019-02664-9

20.

Mustanski

Morgan

DʼAquila

, et al. Individual and network factors associated with racial disparities in HIV among young men who have sex with men: Results from the RADAR cohort study. Jaids Journal of Acquired Immune Deficiency Syndromes 2019; 80: 24–30, https://journals.lww.com/jaids/Fulltext/2019/01010/Individual_and_Network_Factors_Associated_With.4.aspx

21.

Branson

Owen

Wesolowski

, et al. Laboratory testing for the diagnosis of HIV infection: updated recommendations. Atlanta, GA: CDC, 2014. Epub ahead of print 2014. DOI: 10.15620/cdc.23447

22.

Enders

Mistler

Keller

. Multilevel multiple imputation: a review and evaluation of joint modeling and chained equations imputation. Psychol Methods 2016; 21: 222–240.

23.

van Buuren

Groothuis-Oudshoorn

. mice: multivariate imputation by chained equations in R. Journal of Statistical Software, Articles 2011; 45: 1–67.

24.

Rubin

. Multiple imputation for survey nonresponse. Hoboken, NJ: John Wiley & Sons, 2004.

25.

Pinkerton

Chesson

Crosby

, et al. Linearity and nonlinearity in HIV/STI transmission: implications for the evaluation of sexual risk reduction interventions. Eval Rev 2011; 35: 550–565.

26.

Jenness

Weiss

Goodreau

, et al. Incidence of gonorrhea and chlamydia following human immunodeficiency virus preexposure prophylaxis among men who have sex with men: a modeling study. Clin Infect Dis 2017; 65: 712–718.

27.

Falasinnu

Gustafson

Hottes

, et al. A critical appraisal of risk models for predicting sexually transmitted infections. Sex Transm Dis 2014; 41: 321–330.

28.

King

Hughes

Furegato

, et al. Predicting STI diagnoses amongst MSM and young people attending sexual health clinics in England: triage algorithm development and validation using routine clinical data. EClinicalMedicine 2018; 4–5: 43–51.

29.

Lutz

. Screening for asymptomatic extragenital gonorrhea and chlamydia in men who have sex with men: significance, recommendations, and options for overcoming barriers to testing. LGBT Health 2015; 2: 27–34.

30.

Rosenberg

Rothenberg

Kleinbaum

, et al. The implications of respondent concurrency on sex partner risk in a national, web-based study of men who have sex with men in the United States. J Acquir Immune Defic Syndr 2013; 63: 514–521.

31.

Nguyen

Tran

Rocha

LEC

, et al. A systematic review of ehealth interventions addressing HIV/STI prevention among men who have sex with men. Aids Behav 2019; 23: 2253–2272.

32.

St Lawrence

Montaño

Kasprzyk

, et al. STD screening, testing, case reporting, and clinical and partner notification practices: a national survey of US physicians. Am J Public Health 2002; 92: 1784–1788.

33.

Scott

Klausner

. Sexually transmitted infections and pre-exposure prophylaxis: challenges and opportunities among men who have sex with men in the US. AIDS Res Ther 2016; 13: 5.