Sexually Transmitted Infection Screening in Key Populations of Persons Living with HIV

Abstract

Annual screening for bacterial sexually transmitted infections (STI), including gonorrhea/chlamydia (GC/CT) and syphilis, is recommended for persons with HIV (PWH). We used the prevention index to quantify the extent to which STI screening was completed at guideline-recommended frequency in African American and Latinx persons, women, persons with alcohol (AUD) and substance (SUD) use disorders. Data from PWH at Kaiser Permanente Northern California were collected from electronic health records. We defined receipt of GC/CT and syphilis screening consistent with recommendations as a prevention index score ≥75%. Among 9655 PWH (17.7% Latinx; 16.2% African American; 9.6% female; 12.4% AUD; 22.1% SUD), prevention index scores for GC/CT and syphilis increased from 2015-2019. African American PWH had lower odds of receiving an annual syphilis screen (aOR 0.87 [95% CI 0.79-0.97]). Female sex was associated with lower odds of GC/CT (aOR 0.30 [95% CI 0.27-0.34]) and syphilis (aOR 0.27 [95% CI 0.24-0.310) screening. AUD and SUD were not associated with differences in annual GC/CT or syphilis screening. Key PWH subgroups experience ongoing challenges to annual STI screening despite comparable healthcare access.

INTRODUCTION

Rates of bacterial sexually transmitted infections (STIs), such as gonorrhea (GC), chlamydia (CT), and syphilis, have increased dramatically over the last several years.[1-3] Between 2015 and 2019, incident GC cases in the United States have increased by over 55%, while CT and syphilis have increased by 19% and 74%, respectively.[1] Persons living with HIV (PWH) have been especially impacted by rising STI incidence [3-5], which have important implications for national HIV prevention goals. Concomitant STIs among PWH can cause local mucosal inflammation and induce HIV viral shedding in genital secretions.[6-8] Although the clinical implications are unclear, syphilis is also known to increase HIV viral load, even among individuals on antiretroviral therapy.[9-11] Delayed diagnosis and treatment of bacterial STIs can lead to increased healthcare costs and morbidity, including infertility, pelvic inflammatory disease, and negative birth outcomes.[7,12] Guidelines from the Centers for Disease Control and Prevention recommend STI screening at least once a year in PWH,[13] as many STIs are often asymptomatic and can be easily missed.[14,15] Increased STI screening to facilitate rapid detection and treatment is important to curtail rising STI rates and a critical component of the STI National Strategic Plan.[16]

Health care systems have been concerned about rising rates of STIs among PWH patients and other vulnerable groups. For example, in 2014, Kaiser Permanente Northern California (KPNC) implemented self-collected GC/CT swabs to improve screening in PWH. Analysis in the four years following implementation saw a notable increase in overall GC/CT screening, from 45.2% to 63.4%.[17] However, disparities in HIV care, particularly among racial and ethnic minorities, [18,19] women, [20] and persons with alcohol and substance use challenges, [21,22] are well-documented but data evaluating quality of care around STI screening in these key populations are limited. In this study, we build on our prior analysis[17] to identify gaps in STI screening and use the prevention index as a novel quality of care indicator to quantify the extent to which STI screening was completed among PWH, specifically in African American and Latinx individuals, women, and persons with alcohol and substance use disorders. We used the prevention index to quantify STI screening because it allows for the measurement of the proportion of person-time that is covered by a recommended service (e.g., syphilis screening) during a defined period.[23-25] This allows for more nuance in measuring care compared to a binary yes/no value. We hypothesized that screening rates would be lower in these key subgroups despite health insurance coverage and access to care.

METHODS

Study Setting and Participants

KPNC is a large integrated healthcare system that provides comprehensive care, including laboratory and pharmacy services, to approximately 36% of all insured individuals in Northern California.[26] This study included all PWH aged 18 years or older with active KPNC health plan membership between January 2014 and December 2019. Study participants were originally enrolled in the Consortium to improve OutcoMes in HIV/Aids, Alcohol, Aging & multi-Substance (COMpAAAS) study. For our analysis, baseline was defined as January 1, 2015, (the first year for which we calculated prevention index scores), or one year after COMpAAAS study enrollment (i.e., for individuals who enrolled in the parent study after 2015). Participants were followed until KPNC health plan disenrollment, defined as a gap in coverage greater than three months, or end of study period (December 2019). PWH were identified using the KPNC HIV Registry.[27] This study was approved by the institutional review boards at KPNC and the University of California, San Francisco, with a waiver of informed consent.

Measures

GC/CT and syphilis data were abstracted from the electronic health record (EHR). At KPNC, GC and CT are tested concurrently and are entered as a single order, thus these tests were combined in our analysis. Exposures of interest were race and ethnicity (White, Latinx, African American, Asian, Other/Unknown), sex, and a history of alcohol use disorder (AUD) and substance use disorder (SUD) before baseline based on International Classification of Diseases (ICD) criteria (Appendix 1). Baseline demographic and clinical covariates extracted from the EHR and administrative databases included age; neighborhood deprivation index, a marker of neighborhood-level socioeconomic status [28]; smoking status (former, current, never); and number of ambulatory outpatient clinic visits within the internal medicine, infectious disease or family practice departments during the year prior to baseline. Acute GC/CT infections during the study period were identified using laboratory data and incident syphilis infections were identified based on an algorithm previously used by our group (i.e., positive treponemal antibody test or a reactive rapid plasma reagin test and concomitant receipt of benzathine penicillin by intramuscular injection).[29]

Analysis

Differences in KPNC health plan discontinuation by exposures of interest and receipt of STI screening in the year prior to baseline by demographic characteristics were evaluated using chi-square tests. To calculate the prevention index, the time covered by a recommended service was divided by the total amount of time during that period, and then multiplied by 100 to estimate the percent of time covered.[23,25] For example, a patient who is five months overdue for GC/CT screening is shown as covered for seven of the 12 months or 58.3% of the time. If an individual was screened multiple times for GC/CT during the year, the next service is due one year following the last screen. Separate prevention index scores for GC/CT and syphilis were calculated for each participant beginning in 2015 or one year after COMpAAAS study enrollment to allow for a one year “look back” period (i.e., one year prior to baseline) to determine when initial screening was completed and when the next service is due (i.e., 12 months after). For participants who did not complete GC/CT or syphilis screening in the year prior to baseline, screening was marked as due on the first day of the following year (e.g., January 1, 2015). Mean prevention index scores for each calendar year were calculated and tests of trend were estimated using orthogonal polynomial contrasts, a postestimation operator that tests for statistically significant linear, quadratic, cubic, or quartic trend.[30] For this analysis, we tested for linear trend as we were most interested in evaluating any linear changes any STI screening versus fluctuations in screening patterns over time.

We defined prevention index scores of 75% or greater as indicating compliance with screening recommendations. We selected this cut-off to allow a 3-month grace period in each calendar year to complete screening. We estimated odds ratios (ORs) and 95% confidence intervals (CI) to evaluate the association between our exposures of interest (race and ethnicity [reference: White], sex [reference: Male], AUD history [reference: No], SUD history [reference: No]) and STI screening compliance (i.e., time-updated prevention index scores ≥ 75%) using generalized estimating equations with a robust variance estimator, accounting for repeated observations within participants. Exposures of interest were evaluated individually in unadjusted analyses. We then ran adjusted models that included all exposures of interest and sociodemographic and clinical covariates identified a priori based on clinical judgment, including baseline age, neighborhood deprivation index, STI positivity in the year prior to baseline, number of outpatient visits prior to baseline, and smoking status. We also ran the same models using recently diagnosed AUD and SUD (i.e., diagnosed in the year prior to baseline) as sensitivity analyses. Analyses were completed using Stata 14 (College Station, TX).

RESULTS

The total analytical sample included 9655 PWH, of whom 52.9% were White, 17.7% Latinx, 16.2% African American; 7.0% Asian; 70.0% men who have sex with men (MSM); 9.6% female; 12.4% with a history of AUD; and 22.1% with a history of SUD (Table 1). Median age was 49 years (interquartile range [IQR] 41-57). Of the 9655 participants, 41.7% discontinued KPNC health plan membership during the study. Of those who discontinued, most were White (49.4%), followed by Latinx (20.1%), and African American (16.2%; P<.001); and had no SUD history (79.2%; P=.01). There were no significant differences in health plan discontinuation by sex (P=.55) or AUD history (P=.13). In total, 11.1% were diagnosed with either GC/CT or syphilis in the year prior to baseline. Within racial and ethnic groups, STI positivity in the year prior to baseline was highest among Latinx (16.4%) and African American (10.1%) PWH, and lowest among White PWH (9.6%). Prevalence of STI positivity in the year prior to baseline was 3.3% among women; 9.3% among persons with AUD; and 12.2% among persons with SUD. Differences in key demographic characteristics between participants who received and did not receive STI screening in the year prior to baseline are summarized in Table 2.

Table 1.

Demographic and clinical characteristics of persons living with HIV at Kaiser Permanente Northern California (N=9655)

	N	(%)
Age
18-25	429	(4.4)
26-35	1177	(12.2)
36-45	1960	(20.3)
46-55	3380	(35.0)
>55	2709	(28.1)
Race and ethnicity
White	5111	(52.9)
Latinx	1704	(17.7)
African American	1559	(16.2)
Asian	671	(7.0)
Other/Unknown	610	(6.3)
Female	924	(9.6)
Men who have sex with men	6762	(70.0)
Alcohol use disorder	1198	(12.4)
Substance use disorder	2130	(22.1)
Smoking status
Never	3991	(44.3)
Former	3555	(39.4)
Current	1468	(16.3)
STI in year prior to baseline	1070	(11.1)
Neighborhood deprivation index, quartile
1st (lowest)	2723	(28.2)
2nd	2036	(21.1)
3rd	2412	(25.0)
4th (highest)	2484	(25.7)
Number of outpatient visits in year prior to baseline, median (interquartile range)⊥	4	(2-7)

STI = sexually transmitted infection (gonorrhea/chlamydia, syphilis).

^{^}Includes participants who reported >1 race/ethnicity and those with missing race/ethnicity data.

^⊥Outpatient visits include ambulatory visits in internal medicine, infectious diseases, or family practice.

Table 2.

Demographic characteristics of participants who received STI screening in the year prior to baseline compared to participants who did not receive screening.

	Screening in Year Prior to Baseline
	Screened (N =7,618)	Not Screened (N = 2,037)	P-value
Age			<.001
18-25	375 (87.41)	54 (12.59)
26-35	1,031 (87.60)	146 (12.40)
36-45	1,589 (81.07)	371 (18.93)
46-55	2,647 (78.31)	733 (21.69)
>55	1,976 (72.94)	733 (27.06)
Sex			<.001
Male	7,079 (81.08)	1,652 (18.92)
Female	539 (58.33)	385 (41.67)
Race and ethnicity			<.001
Black	1,153 (73.96)	406 (26.04)
Hispanic	1,395 (81.87)	309 (18.13)
White	4,064 (79.51)	1,047 (20.49)
Other/unknown	1,006 (78.53)	275 (21.47)
Neighborhood deprivation index, quartile			<.001
1st (lowest)	2,235 (82.08)	488 (17.92)
2nd	1,582 (77.70)	454 (22.30)
3rd	1,867 (77.40)	545 (22.60)
4th (highest)	1,934 (77.86)	550 (22.14)

Overall, mean prevention index scores for GC/CT screening increased between 2015 and 2019, from 49.4% (95% CI 48.4-50.4) to 61.5% (95% CI 60.5-62.6) (test of linear trend P<.001). Mean prevention index scores for syphilis were notably higher, and we observed a modest but statistically significant increase over the same period, from 70.6% (95% CI 69.7-71.5) to 72.9% (95% CI 72.0-73.8) (test of linear trend P<.001). Mean prevention index scores for GC/CT and syphilis increased over time among Latinx, African American, and Asian PWH (test of linear trend P-values <.001) (Figure 1). However, scores were consistently lower among African American PWH. Among females, mean prevention index scores for GC/CT increased from 27.5% (95% CI 24.6-30.3) in 2015 to 34.3% (95% CI 31.1-37.4) in 2019 (test of linear trend P=.002), while mean scores for syphilis remained largely stable (test of linear trend P=.58) (Figure 2). In persons with AUD, mean prevention index scores for GC/CT increased between 2015 and 2019, from 50.4% (95% CI 47.7-53.1) to 59.9% (95% CI 56.9-62.9%) (test of linear trend P<.001). However, mean scores for syphilis were stable between 72.9% (95% CI 70.5-75.3) and 74.3% (95% 71.7-76.9) among PWH in this group (test of linear trend P=.28) (Figure 3). GC/CT screening increased in persons with SUD, from 51.7% (95% CI 49.7-53.82) in 2015 to 60.9% (95% CI 58.7-63.1) in 2019 (test of linear trend P<.001), but syphilis screening remained largely unchanged (test of linear trend P=.19) (Figure 4).

Figure 1.

Prevention index scores for gonorrhea/chlamydia (GC/CT) and syphilis by race and ethnicity, 2015 to 2019.

Figure 2.

Prevention index scores for gonorrhea/chlamydia (GC/CT) and syphilis by sex, 2015 to 2019.

Figure 3.

Prevention index scores for gonorrhea/chlamydia (GC/CT) and syphilis by alcohol use disorder (AUD) diagnosis, 2015 to 2019.

Figure 4.

Prevention index scores for gonorrhea/chlamydia (GC/CT) and syphilis by substance use disorder (SUD) diagnosis, 2015 to 2019.

In adjusted analyses (Table 3), Latinx PWH had greater odds of receiving annual syphilis screening (aOR 1.16 [95% CI 1.04-1.28]) compared to their White counterparts, while African American PWH had lower odds of receiving an annual syphilis screen (aOR 0.87 [95% CI 0.79-0.97]). Female sex was independently associated with lower odds of receiving annual screening for GC/CT (aOR 0.30 [95% CI 0.27-0.34]) and syphilis (aOR 0.27 [95% CI 0.24-0.30]). AUD and SUD were not significantly associated with receipt of annual GC/CT or syphilis screening. In sensitivity analyses, persons with recently diagnosed AUD had lower odds of receiving GC/CT (aOR 0.84 [95% CI 0.66-1.08]) and syphilis (aOR 0.84 [95% CI 0.64-1.09]) screening, but these estimates were not statistically significant. Persons recently diagnosed with SUD had higher odds of receiving GC/CT (aOR 1.17 [95% CI 0.99-1.39]) and syphilis (aOR 1.31 [95% CI 1.09-1.58]).

Table 3.

Associations between race and ethnicity, sex, alcohol use disorder, substance use disorder, and STI screening among persons living with HIV (N=9655)

	GC/CT		Syphilis
	aOR	(95% CI)	aOR	(95% CI)
Race and ethnicity
White	Ref		Ref
Latinx	1.10	(1.00-1.22)	1.16	(1.04-1.28)
African American	0.98	(0.88-1.08)	0.87	(0.79-0.97)
Asian	1.00	(0.87-1.16)	1.04	(0.90-1.20)
Other/Unknown^	0.89	(0.77-1.04)	0.90	(0.77-1.04)
Sex
Male	Ref		Ref
Female	0.30	(0.27-0.34)	0.27	(0.24-0.30)
Alcohol use disorder
No	Ref		Ref
Yes	0.98	(0.85-1.14)	1.05	(0.90-1.22)
Substance use disorder
No	Ref		Ref
Yes	0.97	(0.87-1.09)	1.05	(0.93-1.18)
Age (each 10-year increase)	0.71	(0.69-0.74)	0.81	(0.79-0.84)
Neighborhood deprivation index, quartile
1st (lowest)	Ref		Ref
2nd	0.86	(0.78-0.95)	0.76	(0.69-0.84)
3rd	0.93	(0.85-1.03)	0.76	(0.69-0.84)
4th (highest)	0.93	(0.84-1.02)	0.72	(0.65-0.79)
Positive STI, diagnosed year prior to baseline
No	Ref		Ref
Yes	3.41	(2.98-3.91)	0.34	(2.88-3.88)
Number of outpatient visits, year prior to baseline	1.01	(1.00-1.02)	1.01	(1.00-1.02)
Smoking status
Never	Ref		Ref
Current	0.83	(0.75-0.92)	0.86	(0.78-0.96)
Former	0.96	(0.89-1.04)	1.01	(0.93-1.09)

Compliance with annual screening recommendations for gonorrhea/chlamydia (GC/CT) and syphilis was defined as prevention index scores ≥ 75%.

^{^}Includes participants who reported >1 race/ethnicity and those with missing race/ethnicity data.

DISCUSSION

In this study, we sought to evaluate STI screening rates among key PWH subgroups known to experience disparities in care. We found that between 2015 and 2019, STI screening increased overall and across all racial and ethnic groups. These results corroborate findings from our previous analysis following implementation of patient self-collected GC/CT swabs across KPNC and reflect an encouraging trend.[17] We also found comparable rates of GC/CT and syphilis screening between PWH with and without a history AUD and SUD. When we limited our analysis to individuals with a recent SUD diagnosis (i.e., one year prior to baseline), those with SUD had greater odds of receiving annual GC/CT and syphilis screening. However, we noted ongoing challenges to annual STI screening among key demographic subgroups despite comparable healthcare access. Although GC/CT and syphilis screening increased over time among African American PWH, prevention index scores were consistently lower among individuals in this group, particularly for syphilis. Similarly, GC/CT and syphilis screening were strikingly lower among women living with HIV.

African American and Latinx individuals have been disproportionately impacted by the STI epidemic. In 2019, incident STI diagnoses were 5 to 8 times higher among African American individuals compared to Whites, and rates of STIs among Latinx persons were up to twice as high as that of White individuals.[31] In our cohort, prevalence of any STI was highest among Latinx and African American PWH, consistent with national trends. While we observed high rates of annual GC/CT and syphilis screening among Latinx PWH, the persistent low rates of screening among African American PWH suggest inequitable distribution of preventive services. We did not have data to understand the reasons for differences in screening, but stigma, medical mistrust, discomfort in discussing sexual behaviors, low perceived risk, and healthcare discrimination have been identified as some of the reasons that drive healthcare inequities among African American individuals.[32-36] Lower odds of screening among individuals with lower neighborhood-level socioeconomic status also suggest broader structural barriers, such as cost or issues around access due to transportation challenges or availability to see a provider outside work hours. While implementation of self-collected GC/CT screening at KPNC has been associated with increased testing overall,[17] other strategies to reduce barriers are likely needed and future research should examine the extent to which social and structural factors may contribute to disparities in STI screening among PWH.

The health consequences of untreated STIs on sexual and reproductive health are especially pronounced among women.[37] Recent estimates from the National Health and Nutrition Examination Survey suggest that incidence of GC/CT is higher in women aged 15 to 39 years compared to men in the same age group.[38] Data from the same survey showed that while incident syphilis infections were higher among men, women comprised nearly one-in-four new infections.[2] In our analysis, 3.3% of women living with HIV had an STI in the year prior to baseline, higher than estimates observed in other studies.[39,40] For example, in a cohort of women receiving HIV care at the University of Alabama at Birmingham, positivity for GC/CT or syphilis ranged from 2.0% to 2.6% between 2013 and 2015. The proportion of women who received GC/CT screening during the same period ranged from 67.2% to 71.5%, while syphilis screening ranged from 64.8% to 87.3%.[39] Similarly, among women enrolled in the CFAR Network of Integrated Clinical Systems (CNICS) cohort, syphilis screening ranged from 55.0% to 69.0% across sites.[41] Although we used a different measure to quantify STI screening in our analysis, our findings suggest that current STI screening practices are likely inadequate and that more effort is needed to improve STI screening uptake among women living with HIV. Concerns around being stigmatized, lack of confidentiality and privacy, and reluctance of healthcare providers in discussing sexual health have been documented as important contributors to low STI screening in this population.[42,43]

AUD and SUD are highly prevalent among PWH [44,45] but, to our knowledge, there are limited data examining STI quality of care in PWH with alcohol and substance use challenges. Alcohol and drug use can lead to impaired sexual decision-making and risk disinhibition, which have been associated with elevated STI risk.[46-49] For example, in experimental alcohol studies, persons who consumed alcohol felt more inclined to engage in condomless sex and were more likely to perform poorly on objectively measured sexual communication and negotiation skills compared to those who did not receive alcohol.[50,51] Further, heavy drinking and substance use have been associated with poor HIV treatment outcomes, including care retention, medication adherence, and viral suppression,[21,22,52,53] which magnify the public health consequences of untreated STIs. Although unexpected, our findings that saw comparable STI screening rates among individuals in these groups (or better screening among those with a recent SUD diagnosis) are reassuring. One potential reason may be sufficient primary care engagement among PWH with AUD and SUD at KPNC. In contrast to other studies,[21,54] our prior analysis found that PWH with high levels of alcohol use had comparable retention in care and viral suppression outcomes as those without problematic patterns of drinking.[55] Clinicians may also be more accustomed to offering routine STI screening among individuals with AUD and SUD, particularly those active problematic use, given the well-established associations between alcohol and substance use and sexual risk. Relatedly, as individuals in this group had a higher prevalence of STI positivity in the year prior to baseline, they may be more aware of the need for regular STI screening.

A key strength of this current study is the use of robust longitudinal EHR data to characterize STI screening among key PWH populations in an integrated healthcare system. Additionally, we used the prevention index as a novel measure to assess compliance with STI screening guidelines, which provided our analysis with greater precision. Our study also has some limitations. We did not have data on sexual activity and, thus, individuals who were not sexually active and for whom annual STI screening is not appropriate may have been included. However, prior data found that approximately 80% of MSM living with HIV and receiving care in a large KPNC medical center were sexually active.[56] We did not differentiate between diagnostic testing (i.e., testing due to symptoms) and routine screening. Differential testing due to symptoms across PWH subgroups may have occurred, especially among groups experiencing stigma, and may account for some of the variability in prevention index scores that we observed. Participants who lost KPNC health plan coverage during follow-up were censored and any differential loss to follow-up, particularly by race and ethnicity and SUD history, may have led to biased estimates. We were unable to evaluate extragenital GC/CT screening and future studies will benefit from more robust assessments of extragenital STI screening. It is also possible that individuals received STI screenings in clinics outside KPNC and were therefore not captured in our analysis. Lastly, this is an insured cohort of PWH, and findings may not generalize to other settings such as public health systems or to PWH with limited access to care.

In conclusion, these findings suggest that overall GC/CT and syphilis screening among PWH has improved between 2015 and 2019. These results demonstrate a promising step towards achieving the goals of the STI National Strategic Plan.[16] However, ongoing assessments of STI screening quality of care, especially in key populations, are critical to monitor disparities and ensure progress. The inequities we observed in STI screening in key demographic subgroups with a known disproportionate burden of STIs highlight important areas of unaddressed need. Broader access to STI screening, particularly among African American PWH and women living with HIV, can help improve health equity, reduce secondary transmission, and better target resources to mitigate STI-related morbidity.