A critical appraisal of risk models for predicting sexually transmitted infections

Clinical features-based machine learning models to separate sexually transmitted infections from other skin diagnoses

INTRODUCTION

Many sexual health services are overwhelmed and cannot cater for all the individuals who present with sexually transmitted infections (STIs). Digital health software that separates STIs from non-STIs could improve the efficiency of clinical services. We developed and evaluated a machine learning model that predicts whether patients have an STI based on their clinical features.

METHODS

We manually extracted 25 demographic features and clinical features from 1315 clinical records in the electronic health record system at Melbourne Sexual Health Center. We examined 16 machine learning models to predict a binary outcome of an STI or a non-STI diagnosis. We evaluated the models' performance with the area under the ROC curve (AUC), accuracy and F1-scores.

RESULTS

Our study included 1315 consultations, of which 36.8% (484/1315) were diagnosed with STIs and 63.2% (831/1315) had non-STI conditions. The study population predominantly consisted of heterosexual men (49.5%, 651/1315), followed by gay, bisexual and other men who have sex with men (GBMSM) (25.7%), women (21.6%) and unknown gender (3.2%). The median age was 31 years (intra-quartile range (IQR) 26-39). The top 5 performing models were CatBoost (AUC 0.912), Random Forest (AUC 0.917), LightGBM (AUC 0.907), Gradient Boosting (AUC 0.905) and XGBoost (AUC 0.900). The best model, CatBoost, achieved an accuracy of 0.837, sensitivity of 0.776, specificity of 0.831, precision of 0.782 and F1-score of 0.778. The key important features were lesion duration, type of skin lesions, age, gender, history of skin disorders, number of lesions, dysuria duration, anorectal pain and itchiness.

CONCLUSIONS

Our best model demonstrates a reasonable performance in distinguishing STIs from non-STIs. However, to be clinically useful, more detailed information such as clinical images, may be required to reach sufficient accuracy.

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

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.

Risk Prediction Score for Screening Asymptomatic Sexually Transmitted Infections in HIV-positive French Men Who Have Sex with Men (ANRS 9520 DRIVER)

OBJECTIVES

Asymptomatic sexually transmitted infections (STI) are frequent among men who have sex with men (MSM). Identifying asymptomatic STIs is a crucial issue, not only for secondary but also for primary prevention, as early treatment can reduce transmission risk. We aimed to develop a self-reported predictive score for early identification of asymptomatic STIs.

METHODS

Participants provided clinical data and completed a self-administered questionnaire including sociodemographic variables and behaviors during the 6 previous months. We used multivariable logistic regression to identify factors associated with asymptomatic STIs. We calculated the accuracy of the model by the non-parametric area (AUC) under the receiver-operating-characteristic (ROC) curve to find the optimal discriminant threshold for screening.

RESULTS

A total of 781 HIV-positive MSM were included with a mean age of 46.8 years. Asymptomatic STI prevalence was 13.2%. Detectable plasma HIV RNA (adjusted odds ratio (aOR [95% CI): 2.54 [1.23;5.25]), inconsistent condom use during anal sex (2.20 [1.36;3.56]), group sex (2.00 [1.15;3.45]), during or-genital practices (1.83 [1.12;3.01]), not being in stable relationship (1.70 [1.01;2.66] and an item from a sensation-seeking behavioral scale "I don't like watching porn videos" (1.61 [1.01;2.59] were associated with asymptomatic STI. AUC was 0.7 and with optimal threshold of 0.1082 for this model; sensitivity was 80.4%. Self-reported asymptomatic STI predictive score was built with this threshold according to the 6 factors in the final model.

CONCLUSIONS

As this predictive score is not designed to be diagnostic, but to provide indications for diagnostic tests, its ease of administration and sensitivity remain the most important features. Its use in clinical practice for early detection of asymptomatic STIs potentially can reinforce STI primary and secondary prevention.

Identifying youth at high risk for sexually transmitted infections in community-based settings using a risk prediction tool: a validation study

BACKGROUND : Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) are the most common bacterial sexually transmitted infections (STIs) worldwide. In the absence of affordable point-of-care STI tests, WHO recommends STI testing based on risk factors. This study aimed to develop a prediction tool with a sensitivity of > 90% and efficiency (defined as the percentage of individuals that are eligible for diagnostic testing) of < 60%.

METHODS

This study offered CT/NG testing as part of a cluster-randomised trial of community-based delivery of sexual and reproductive health services to youth aged 16-24 years in Zimbabwe. All individuals accepting STI testing completed an STI risk factor questionnaire. The outcome was positivity for either CT or NG. Backwards-stepwise logistic regression was performed with p ≥ 0.05 as criteria for exclusion. Coefficients of variables included in the final multivariable model were multiplied by 10 to generate weights for a STI risk prediction tool. A maximum likelihood Receiver Operating Characteristics (ROC) model was fitted, with the continuous variable score divided into 15 categories of equal size. Sensitivity, efficiency and number needed to screen were calculated for different cut-points.

RESULTS

From 3 December 2019 to 5 February 2020, 1007 individuals opted for STI testing, of whom 1003 (99.6%) completed the questionnaire. CT/NG prevalence was 17.5% (95% CI 15.1, 19.8) (n = 175). CT/NG positivity was independently associated with being female, number of lifetime sexual partners, relationship status, HIV status, self-assessed STI risk and past or current pregnancy. The STI risk prediction score including those variables ranged from 2 to 46 with an area under the ROC curve of 0.72 (95% CI 0.68, 0.76). Two cut-points were chosen: (i) 23 for optimised sensitivity (75.9%) and specificity (59.3%) and (ii) 19 to maximise sensitivity (82.4%) while keeping efficiency at < 60% (59.4%).

CONCLUSIONS

The high prevalence of STIs among youth, even in those with no or one reported risk factor, may preclude the use of risk prediction tools for selective STI testing. At a cut-point of 19 one in six young people with STIs would be missed.

Prediction model for gonorrhea, chlamydia, and trichomoniasis in the emergency department

OBJECTIVE

History and physical examination findings can be unreliable for prediction of genitourinary tract infections and differentiation of urinary tract infections from sexually transmitted infections (STIs). The study objective was to develop a prediction tool to more accurately identify patients with STIs.

METHODS

A retrospective review of 64,490 emergency department (ED) encounters between April 18, 2014, and March 7, 2017, where patients age 18 years or older had urinalysis and urine culture or testing for gonorrhea, chlamydia, or trichomonas, was used to develop a prediction model for men and women with Neisseria gonorrhoeae or Chlamydia trachomatis, or both, and for women with Trichomonas vaginalis. The data set was randomly divided into two-thirds discovery and one-third validation. Groups were assigned through a random number generator. Backward step regression modeling was used to identify the best model for each outcome.

RESULTS

With use of age, race, marital status, and findings from vaginal wet preparation (white blood cells [WBCs], clue cells, and yeast) and urinalysis (squamous epithelial cells, protein, leukocyte esterase, and WBCs), the models had areas under the receiver operating characteristic curve of 0.80 for men with N gonorrhoeae or C trachomatis, or both; 0.75 for women with N gonorrhoeae or C trachomatis, or both; and 0.73 for women with T vaginalis.

CONCLUSIONS

The model estimated likelihood of ED patients having STIs was reasonably accurate with a limited number of demographic and laboratory variables. In the absence of point-of-care STI testing, use of a prediction tool for STIs may improve antimicrobial stewardship.

Validation of a Clinical Prediction Rule to Predict Asymptomatic Chlamydia and Gonorrhea Infections Among Internet-Based Testers

BACKGROUND

Clinical prediction rules (CPRs) can be used in sexually transmitted infection (STI) testing environments to prioritize individuals at the highest risk of infection and optimize resource allocation. We previously derived a CPR to predict asymptomatic chlamydia and/or gonorrhea (CT/NG) infection among women and heterosexual men at in-person STI clinics based on 5 predictors. Population differences between clinic-based and Internet-based testers may limit the tool's application across settings. The primary objective of this study was to assess the validity, sensitivity, and overall performance of this CPR within an Internet-based testing environment (GetCheckedOnline.com).

METHODS

We analyzed GetCheckedOnline online risk assessment and laboratory data from October 2015 to June 2019. We compared the STI clinic population used for CPR derivation (data previously published) and the GetCheckedOnline validation population using χ2 tests. Calibration and discrimination were assessed using the Hosmer-Lemeshow goodness-of-fit test and the area under the receiver operating curve, respectively. Sensitivity and the fraction of total screening tests offered were quantified for CPR-predicted risk scores.

RESULTS

Asymptomatic CT/NG infection prevalence in the GetCheckedOnline population (n = 5478) was higher than in the STI clinic population (n = 10,437; 2.4% vs. 1.8%, P = 0.007). When applied to GetCheckedOnline, the CPR had reasonable calibration (Hosmer-Lemeshow, P = 0.90) and discrimination (area under the receiver operating characteristic, 0.64). By screening only individuals with total risk scores ≥4, we would detect 97% of infections and reduce screening by 14%.

CONCLUSIONS

The application of an existing CPR to detect asymptomatic CT/NG infection is valid within an Internet-based STI testing environment. Clinical prediction rules applied online can reduce unnecessary STI testing and optimize resource allocation within publicly funded health systems.

Performance of a tool to identify different types of self-reported sexual risk among women attending a contraception and sexual health clinic: results of a cross-sectional survey

INTRODUCTION

A clinical prediction rule (CPR) using psychosocial questions was previously derived to target sexual healthcare in general practice by identifying women at risk of unintended pregnancy (UIP) and sexually transmitted infections (STIs). This psychosocial CPR may help target resources within contraception and sexual health (CASH) services. This study investigated how well it predicted recent self-reported risk of UIP and STI acquisition among women attending a CASH clinic.

METHODS

Female patients aged 16-44 years attending a CASH clinic in South-East England were offered a questionnaire on arrival. This comprised psychosocial questions, and others addressing three sexual risks: (1) two or more male sexual partners in the last year (2+P), (2) risk of STI acquisition through most recent partner and (3) risk of UIP in the last 6 months. A CPR score was calculated for each participant and cross-tabulated against self-report of each sexual risk to estimate CPR sensitivity and specificity.

RESULTS

The psychosocial questions predicting 2+P had sensitivity 83.2% (95% CI 79.3% to 86.5%) and specificity 56.1% (95% CI 51.3%-60.6%). Those predicting combined 2+P and/or risk of STI acquisition through most recent partner had a sensitivity of 89.1% (95% CI 85.7%-91.8%) and specificity of 43.7% (95% CI 39.0%-48.5%). Questions predicting risk of UIP in the last 6 months had a sensitivity of 82.5% (95% CI 78.6%-86.0%) and specificity of 48.3% (95% CI 43.4%-53.1%).

CONCLUSIONS

The CPR demonstrated good sensitivity but low specificity, so may be suited to triaging or stratifying which interventions to offer CASH patients and by which mode (eg, online vs face-to-face). Further investigation of causal links between psychosocial factors and sexual risk is warranted to support development of psychosocial interventions for this patient group.

"Mini Dial-A-Nurses" and "Good Brands": What Are the Desirable Features of Online HIV/STI Risk Calculators?

A wide variety of risk calculators estimate individuals' risk for HIV/sexually transmitted infections (STI) online. These tools can help target HIV/STI screening and optimize clinical decision-making. Yet, little evidence exists on suitable features for these tools to be acceptable to end-users. We investigated the desirable characteristics of risk calculators among STI clinic clients and testing service providers. Participants interacted with online HIV/STI risk calculators featuring varied target audiences, completion lengths, and message outputs. Thematic analysis of focus groups identified six qualities that would make risk calculators more appealing for online client use: providing personalized risk assessments based on users' specific sexual behaviors and HIV/STI-related concerns; incorporating nuanced risk assessment and tailored educational information; supplying quantifiable risk estimates; using non-stigmatizing and inclusive framing; including explanations and next steps; and developing effective and appropriate branding. Incorporating these features in the design of online HIV/STI risk calculators may improve their acceptability among end-users.

Descriptive Evaluation of Male Emergency Department Patients in the United States With Gonorrhea and Chlamydia

Introduction

Sexually transmitted infections are commonly tested for in the emergency department (ED), but diagnostic test results are often unavailable during the clinical encounter.

Methods

We retrospectively reviewed health records of 3,132 men ≥18 years that had an emergency department visit in northeast Ohio between April 18, 2014 and March 7, 2017. All subjects underwent testing for Neisseria gonorrhoeae and Chlamydia trachomatis. Independent t-tests and chi-square analyses were performed as well as multivariable regression analysis.

Results

On univariable analysis, men with N gonorrhoeae and/or C trachomatis, compared with uninfected men, were younger (25.9 vs 32.4 years), more likely to be of Black race (91.7% vs 85.6%), less likely to be married (3.7% vs 10.2%), less likely to arrive to the ED by ambulance or police (1.7% vs 4.1%), and more likely to be diagnosed with a urinary tract infection (8.3% vs 3.7%), to be treated for gonorrhea and chlamydia in the ED (84.6% vs 54.9%), and to have higher emergency severity index (ESI) scores (3.8 vs 3.6) (P ≤ .03 for all). On urinalysis, men infected with N gonorrhoeae and/or C trachomatis had significantly more white blood cells (55.1 vs 20.9); more mucus (1.3 vs 1.2); higher leukocyte esterase (1.5 vs .4); fewer squamous epithelial cells (.6 vs 1.4); higher urobilinogen (1.1 vs .8); higher bilirubin (.09 vs .05); and more protein (.4 vs .3) (P ≤ .04).

Conclusions

Demographic and urinalysis findings can be associated with an increased odds of men being infected with N gonorrhoeae and/or C trachomatis.

Protecting our future generation: study protocol for a randomized controlled trial evaluating a sexual health self-care intervention with Native American youth and young adults

Background

Disparities in sexually transmitted infections (STI) are an urgent problem among Native American youth and young adults which are not fully explained by different sexual or related behaviors. These sexual health disparities are more likely attributed to social environments and structural determinants such as a shortage of sexual healthcare providers, lower socioeconomic status, and access barriers to STI screening and treatment, including geographic isolation and confidentiality concerns. Innovative, non-clinic based alternatives to promote STI screening and treatment are essential for alleviating these disparities. Self-care, or the care taken by individuals towards their own health and well-being may be such a strategy. This study will assess the efficacy of a self-care intervention, called Protecting Our Future Generation, for increasing uptake of STI screening and impacting sexual risk and protective behaviors among Native American youth and young adults living in a reservation-based community in the Southwestern United States.

Methods

The proposed study is a randomized controlled trial to test the efficacy of a self-care intervention compared to a control condition. Participants will be Native Americans ages 14–26 years old who have had vaginal or anal sex at least once in their lifetime. Participants will be randomized to the intervention which includes: 1) a sexual health self-assessment with embedded clinical prediction tool predicting STI positivity, and 2) personalized messaging with key steps to lower risk for STIs, or the control condition which includes: 1) a self-assessment about water, soda and sugar sweetened beverage consumption, and 2) personalized messaging to meet recommended daily intake. All participants will be offered a self-administered STI test. Participants will complete assessments at baseline, 3- and 6-months follow-up. The primary outcome measure is completion of STI screening.

Discussion

Protecting Our Future Generation is among the first self-care interventions uniquely focused on sexual health among a Native American population, who endure significant sexual health disparities and are under-represented in research. If efficacious, the intervention will be a model of sexual health self-care for Native American youth and young adults adaptable for use in healthcare and community-based settings.

Trial registration

Clinical Trials: http://clinicaltrials.gov; NCT03895320; Registered 03/28/2019.

The development of an online risk calculator for the prediction of future syphilis among a high-risk cohort of men who have sex with men and transgender women in Lima, Peru

Background Syphilis incidence worldwide has rebounded since 2000, particularly among men who have sex with men (MSM). A predictive model for syphilis infection may inform prevention counselling and use of chemoprophylaxis.

METHODS

Data from a longitudinal cohort study of MSM and transgender women meeting high-risk criteria for syphilis who were followed quarterly for 2 years were analysed. Incidence was defined as a four-fold increase in rapid plasma reagin (RPR) titres or new RPR reactivity if two prior titres were non-reactive. Generalised estimating equations were used to calculate rate ratios (RR) and develop a predictive model for 70% of the dataset, which was then validated in the remaining 30%. An online risk calculator for the prediction of future syphilis was also developed.

RESULTS

Among 361 participants, 22.0% were transgender women and 34.6% were HIV-infected at baseline. Syphilis incidence was 19.9 cases per 100-person years (95% confidence interval (CI) 16.3-24.3). HIV infection (RR 2.22; 95% CI 1.54-3.21) and history of syphilis infection (RR 2.23; 95% 1.62-3.64) were significantly associated with incident infection. The final predictive model for syphilis incidence in the next 3 months included HIV infection, history of syphilis, number of male sex partners and sex role for anal sex in the past 3 months, and had an area under the curve of 69%. The online syphilis risk calculator based on those results is available at: www.syphrisk.net.

CONCLUSIONS

Using data from a longitudinal cohort study among a population at high risk for syphilis infection in Peru, we developed a predictive model and online risk calculator for future syphilis infection. The predictive model for future syphilis developed in this study has a moderate predictive accuracy and may serve as the foundation for future studies.

Sexually Transmitted Infection Prevalence in Women With HIV: Is There a Role for Targeted Screening?

BACKGROUND

Rates of sexually transmitted infections (STIs) and HIV are highest in the southern United States but vary widely by sex, age, and risk behavior. Current guidelines recommend annual screening for chlamydia, gonorrhea, syphilis, and trichomoniasis in all sexually active women with HIV.

METHODS

Screening rates and test positivity for chlamydia, gonorrhea, syphilis, and trichomoniasis were determined per calendar year in this retrospective cohort study of women in care at an urban HIV clinic in Birmingham, Alabama, from 2013 to 2015. Chlamydia, gonorrhea, and trichomonas infections were detected by molecular diagnostics and syphilis by serology. A combined end point for chlamydia/gonorrhea/syphilis (STI-3) was created based on similar test positivity and predictors. Predictors of STI-3 were identified using logistic regression and generalized estimating equations.

RESULTS

Among 745 women with HIV, median age was 46.8 years, 78.8% were black, and 61% were sexually active. In 2015, 83.7% of women were tested for STI. Test positivity was 1.0% for chlamydia, 0.5% for gonorrhea, 1.6% for syphilis, and 13.3% for trichomoniasis. Independent predictors of STI-3 were recent chlamydia or gonorrhea (odds ratio [OR], 3.7; 95% confidence interval [CI], 1-13.4; P = 0.047), public insurance compared with private (OR, 3.5; CI, 1-11.8; P = 0.048), and sex after drugs/alcohol (OR, 3.0; CI, 1.2-8.0; P = 0.025). Women 50 years or older were less likely to have STI (OR, 0.3; CI, 0.1-1; P = 0.040).

CONCLUSIONS

In a cohort of women engaged in HIV care in the southern United States, detection of chlamydia, gonorrhea, and syphilis was infrequent but trichomoniasis was common. Many women screened for STI were low risk and universal testing strategies warrant evaluation.

Deriving a clinical prediction rule to target sexual healthcare to women attending British General Practices

Some women attending General Practices (GPs) are at higher risk of unintended pregnancy (RUIP) and sexually transmitted infections (STI) than others. A clinical prediction rule (CPR) may help target resources using psychosocial questions as an acceptable, effective means of assessment. The aim was to derive a CPR that discriminates women who would benefit from sexual health discussion and intervention. Participants were recruited to a cross-sectional survey from six GPs in a city in South-East England in 2016. On arrival, female patients aged 16-44 years were invited to complete a questionnaire that addressed psychosocial factors, and the following self-reported outcomes: 2+ sexual partners in the last year (2PP) and RUIP. For each sexual risk, psychosocial questions were retained from logistic regression modelling which best discriminated women at risk using the C-statistic. Sensitivity and specificity were established in consultation with GP staff. The final sample comprised N = 1238 women. 2PP was predicted by 11 questions including age, binge-drinking weekly, ever having a partner who insulted you often, current smoking, and not cohabiting (C-statistic = 0.83, sensitivity = 73% and specificity = 77%). RUIP was predicted by 5 questions including sexual debut <16 years, and emergency contraception use in the last 6 months (C-statistic = 0.70, sensitivity = 69% and specificity = 57%). 2PP was better discriminated than RUIP but neither to a clinically-useful degree. The finding that different psychosocial factors predicted each outcome has implications for prevention strategies. Further research should investigate causal links between psychosocial factors and sexual risk.

The patient rationale for seeking HIV PEP: Qualitative results from a nurse-led program

BACKGROUND

HIV PEP is an intervention to prevent HIV seroconversion among HIV-negative persons after they have come into contact with the virus. The literature suggests PEP is greater than 80% effective at reducing HIV seroconversion. To expand access to this prevention strategy, we established a nurse-led PEP program in two STI clinics.

AIMS

We sought to explore why people seek PEP, and their experiences using it.

MATERIALS & METHODS

We invited participants who used our PEP program to engage in semi-structured qualitative interviews. These were audio-recorded, transcribed verbatim, and analyzed using the techniques of thematic analysis.

RESULTS

Twelve men who initiated PEP through our program participated in an interview. Thematic analysis of the data identified three themes: (1) assessing risk, (2) euphoria and distress, and (3) reducing distress.

DISCUSSION

Based on our findings, it appears as though instances of condomless sex were atypical and unplanned, they occurred due to contextual factors (e.g., consent, desire to explore, sexual chemistry/intensity), and they resulted in a state of temporary euphoria, which, afterward, transformed into distress (i.e., "feel[ing] stupid"). The participants used PEP not only for HIV prevention, but also (and perhaps more so) to address their emotional state. PEP used seemed to primarily occur to address situations of atypical sexual contact with partners they did not "know" or "trust".

CONCLUSION

We feel these results help contextualize how potential PEP users view HIV risk assessments, and how they determine their need for this intervention.

Assessing association between IWantTheKit risk quiz tool and sexually transmitted infection positivity in male users for sexually transmitted infection screening

Our previous pilot study during 2010-2013, based on the IWantTheKit (IWTK) home self-collection program for sexually transmitted infections (STIs), showed that voluntary risk score tool predicted STIs well in female volunteers compared to their male counterparts. Risk score became a required part of the IWTK program in August 2013. We investigated association of IWTK risk score and presence of STI in 592 male participants living in Maryland and Washington DC from August 2013 to April 2015. The risk score quiz includes questions on demographic and sexual risk behavior. Data were analyzed using the Cochran-Armitage test for trend to determine if prevalence of STIs (Chlamydia, gonorrhea, or trichomoniasis) increased with the higher risk score category. Overall, 57% of participants were aged < 30 years (mean: 30.1 ± 9.3 years); 42% white, 42% black, and 16% other races. The majority (67%) of participants had medium risk scores of 3-6, followed by high scores of 7-10 (22%), and 0-2 (11%). The overall prevalence of STIs was 10.5% (62/592). The prevalence of STIs was 3.1% for users with risk scores of 0-2, 10.4% for those with scores 3-6, and 14.3% for those with scores 7-10 (trend test: p = 0.019). Medium and high IWTK risk scores successfully predicted the probability of STIs in male participants after elimination of potential selection biases.

REACH: a mixed-methods study to investigate the measurement, prediction and improvement of retention and engagement in outpatient HIV care

Background

Antiretroviral therapy (ART) benefits individuals living with human immunodeficiency virus (HIV) through reduced morbidity and mortality, and brings public health gains through a reduction in HIV transmission. People living with human immunodeficiency virus (PLWH) need to know their HIV status and engage in HIV care in order for these individual and public health benefits to be realised.

Objective

To explore, describe and understand HIV outpatient attendance in PLWH, in order to develop cost-effective interventions to optimise engagement in care.

Design

A mixed-methods study incorporating secondary analysis of data from the UK Collaborative HIV Cohort (UK CHIC) study and primary data collection.

Methods

Phase 1 – an engagement-in-care (EIC) algorithm was developed to categorise patients as in care or out of care for each month of follow-up. The algorithm was used in group-based trajectory analysis to examine patterns of attendance over time and of the association between the proportion of months in care before ART initiation and post-ART mortality and laboratory test costs. Phase 2 – a cross-sectional survey was conducted among patients attending seven London HIV clinics. Regular attenders (all appointments attended in past year), irregular attenders (one or more appointments missed in past year) and non-attenders (recent absence of ≥ 1 year) were recruited. A ‘retention risk tool’ was developed to identify those at risk of disengaging from care. Individual in-depth interviews and focus groups were conducted with PLWH. Phase 3 – key informant interviews were conducted with HIV service providers. Interventions were developed from the findings of phases 2 and 3.

Results

Plots from group-based trajectory analysis indicated that four trajectories best fitted the data. Higher EIC is associated with reduced mortality but the association between EIC before starting ART, and post-ART mortality [relative hazard (RH) per 10% increase in EIC 0.29, 95% confidence interval (CI) 0.18 to 0.47] was attenuated after adjustment for fixed covariates and post-ART cluster of differentiation 4 counts and viral loads (RH 0.74, 95% CI 0.42 to 1.30). Small differences were found in pre-ART EIC and the costs of post-ART lab tests. The final model for the retention risk tool included age at diagnosis, having children, recreational drug use, drug/alcohol dependency, insufficient money for basic needs and use of public transport to get to the clinic. Quantitative and qualitative data showed that a range of psychological, social and economic issues were associated with disengagement from care. The negative impact of stigma on attendance was highlighted. Interventions were proposed that support a holistic approach to care including peer support, address stigma by holding clinics in alternative locations and involve training staff to encourage attendance.

Conclusions

The study shows the adverse health impacts of disengaging from HIV care and demonstrates the importance of the wider health and social context in managing HIV effectively. Although phase 1 analysis was based on UK data, phases 2 and 3 were limited to London. The interventions proposed are supported by the data but their cost-effectiveness requires testing. Future research is needed to evaluate the interventions, to validate our retention risk tool across populations and settings, and to fully analyse the economic costs of disengaging from HIV care.

Funding

The National Institute for Health Research Health Services and Delivery Research programme. The UK CHIC study is funded by the Medical Research Council UK (grant numbers G0000199, G0600337, G0900274 and M004236).

Can psychosocial and socio-demographic questions help identify sexual risk among heterosexually-active women of reproductive age? Evidence from Britain's third National Survey of Sexual Attitudes and Lifestyles (Natsal-3)

Background

Contraceptive advice and supply (CAS) and sexually transmitted infection (STI) testing are increasingly provided in primary care. Most risk assessment tools are based on sexual risk behaviours and socio-demographics, for use online or in specialist services. Combining socio-demographic and psychosocial questions (e.g. religious belief and formative experience) may generate an acceptable tool for targeting women in primary care who would benefit from intervention. We aimed to identify psychosocial and socio-demographic factors associated with reporting key sexual risk behaviours among women in the British general population.

Methods

We undertook complex survey analysis of data from 4911 hetero-sexually active women aged 16–44 years, who participated in Britain’s third National Survey of Sexual Attitudes and Lifestyles (Natsal-3), a national probability sample survey undertaken 2010–2012. We used multivariable regression to examine associations between the available psychosocial and socio-demographic variables in Natsal-3 and reports of three key sexual behaviours: a) 2+ partners in the last year (2PP); b) non-use of condoms with 2+ partners in the last year (2PPNC); c) non-use of condoms at first sex with most recent sexual partner (FSNC). We adjusted for key socio-demographic factors: age, ethnicity and socio-economic status (measured by housing tenure).

Results

Weekly binge drinking (6+ units on one occasion), and first sex before age 16 were each positively associated with all three sexual behaviours after adjustment. Current relationship status, reporting drug use (ever), younger age and living in rented accommodation were also associated with 2+ partners and 2 + partners without condoms after adjustment. Currently being a smoker, older age and respondent ethnicity were associated with FSNC after adjustment for all other variables. Current smoking status, treatment for depression (last year), and living at home with both parents until the age of 14 were each associated with one or more of the behaviours.

Conclusions

Reported weekly binge drinking, early sexual debut, and age group may help target STI testing and/or CAS among women. Further research is needed to examine the proportion of sexual risk explained by these factors, the acceptability of these questions to women in primary care and the need to customise them for community and other settings.

Electronic supplementary material

The online version of this article (doi:10.1186/s12889-016-3918-8) contains supplementary material, which is available to authorized users.

Get Checked… Where? The Development of a Comprehensive, Integrated Internet-Based Testing Program for Sexually Transmitted and Blood-Borne Infections in British Columbia, Canada

Background

Testing for sexually transmitted and blood-borne infections (STBBI) is an effective public health strategy that can promote personal control of one’s health and prevent the spread of these infections. Multiple barriers deter access to testing including fear of stigmatization, inaccurate health care provider perceptions of risk, and reduced availability of clinic services and infrastructure. Concurrent increases in sexually transmitted infection (STI) rates and demands on existing clinical services make this an even more pressing concern. Web-based testing offers several advantages that may alleviate existing clinical pressures and facilitate appropriate testing access.

Objective

This paper describes the planning, development, and usability testing of a novel Web-based testing service, GetCheckedOnline (GCO), as a complementary testing option integrated within existing sexual health services within British Columbia (BC).

Methods

From 2009 to 2014, we engaged a multidisciplinary team in the design and development of GCO. We conducted 3 initial research studies to ascertain the opinions of youth, men who have sex with men (MSM), and STI clinic clients regarding Web-based testing and elicited perspectives of sexual health care providers through focus groups. We developed an informed consent process, risk assessment questions, and test recommendations based on provincial and national guidelines and evaluated these through consultations with clinical and community stakeholders. We also conducted a preliminary health equity impact assessment whose findings also informed the GCO program mode. Finally, from April 2011 to December 2012 we gathered qualitative data from 25 participants on the functionality and usability of a GCO prototype and incorporated their recommendations into a final model.

Results

GCO launched in the fall of 2014 across 6 pilot sites in Vancouver, BC. The service involves 3 main steps: (1) create an account, complete an assessment, and print a laboratory requisition, (2) provide blood and urine specimens at participating laboratory locations, and (3) receive test results on the Internet or by phone. During this pilot phase, we promoted GCO to existing STI clinic clients and MSM in the Greater Vancouver region. A rigorous mixed-method evaluation of GCO’s uptake, acceptability, and health system impacts is currently underway.

Conclusions

GCO is the first comprehensive Web-based STBBI testing program in Canada that is integrated with existing sexual health services, with the potential to reduce pressures on existing clinical services and reach populations facing the greatest barriers to testing. Our experience highlights the facilitators and challenges of developing and implementing novel complex eHealth interventions within the health care system, and underscores the importance of considering broader implementation contexts.

Use of a risk quiz to predict infection for sexually transmitted infections: a retrospective analysis of acceptability and positivity

BACKGROUND

Individuals who are sexually active may want to make a decision as to whether they are at risk for having a sexually transmitted infection (STI) such as Chlamydia trachomatis, Neisseria gonorrhoeae and Trichomonas vaginalis. Our goal was to develop and evaluate a simple self-taken sexual risk quiz for participants, ordering an online STI self-collection test kit to determine whether the score predicted infection status.

METHODS

As part of the IWantTheKit programme for home sample self-collection for STIs, 2010-2013, the programme asked male and female users to voluntarily take a risk quiz. The six-question quiz was about risk behaviour and included an age question. Data analyses were stratified by gender as determined a priori. Scores 0-10 were stratified into risk groups for each gender based on similar risk score-specific STI prevalence. Retrospective analyses were performed to assess whether risk group predicted aggregate STI positivity. Urogenital/rectal mailed samples were tested by nucleic acid amplification tests.

RESULTS

More females (N=836) than males (N=558) provided voluntary risk scores. The percentage of eligible participants who submitted scores was 43.9% for both females and males. There was a higher STI infection rate in females (14.0%) than in males (7.0%) for having any STI (p<0.001). Multivariate logistic analysis for females, which controlled for age and race, demonstrated that a higher risk score group independently predicted risk for having an STI (OR of 2.2 for risk scores 5-7 and 4.2 OR for scores of 8-10). For males, the multivariate model, which controlled for race, indicated that no risk score group was associated having an STI.

CONCLUSIONS

Results of a participant's own sexual risk quiz score independently predicted STI positivity for women, but not for men. Further study of this simple risk quiz is required.

Deriving and validating a risk estimation tool for screening asymptomatic chlamydia and gonorrhea

BACKGROUND

There has been considerable interest in the development of innovative service delivery modules for prioritizing resources in sexual health delivery in response to dwindling fiscal resources and rising infection rates.

METHODS

This study aims to derive and validate a risk scoring algorithm to accurately identify asymptomatic patients at increased risk for chlamydia and/or gonorrhea infection. We examined the electronic records of patient visits at sexual health clinics in Vancouver, Canada. We derived risk scores from regression coefficients of multivariable logistic regression model using visits between 2000 and 2006. We evaluated the model's discrimination, calibration, and screening performance. Temporal validation was assessed in visits from 2007 to 2012.

RESULTS

The prevalence of infection was 1.8% (n = 10,437) and 2.1% (n = 14,956) in the derivation and validation data sets, respectively. The final model included younger age, nonwhite ethnicity, multiple sexual partners, and previous infection and showed reasonable performance in the derivation (area under the receiver operating characteristic curve = 0.74; Hosmer-Lemeshow P = 0.91) and validation (area under the receiver operating characteristic curve = 0.64; Hosmer-Lemeshow P = 0.36) data sets. A risk score cutoff point of at least 6 detected 91% and 83% of cases by screening 68% and 68% of the derivation and validation populations, respectively.

CONCLUSIONS

These findings support the use of the algorithm for individualized risk assessment and have important implications for reducing unnecessary screening and saving costs. Specifically, we anticipate that the algorithm has potential uses in alternative settings such as Internet-based testing contexts by facilitating personalized test recommendations, stimulating health care-seeking behavior, and aiding risk communication by increasing sexually transmitted infection risk perception through the creation of tailored risk messages to different groups.

Predictors identifying those at increased risk for STDs: a theory-guided review of empirical literature and clinical guidelines

SummarySexually transmitted diseases (STDs) are leading causes of substantial morbidity worldwide. Identification of risk factors for estimating STD risk provides opportunities for optimising service delivery in clinical settings, including improving case finding accuracy and increasing cost-efficiency by limiting the testing of low-risk individuals. The current study was undertaken to synthesise the evidence supporting commonly cited chlamydia and gonorrhoea risk factors. The level of empirical support for the following predictors was strong/moderate: age, race/ethnicity, multiple lifetime sexual partners, sex with symptomatic partners and concurrent STD diagnosis. The following predictors had weak evidence: socio-economic status, transactional sex, drug/alcohol use, condom use and history of STD diagnosis. The most frequently listed predictors among nine clinical guidelines were younger age and multiple sexual partners; the least consistently listed predictor was inconsistent condom use. We found reasonably good concordance between risk factors consistently listed in the recommendations and predictors found to have strong empirical support in the literature. There is a need to continue building the evidence base to explicate the mechanisms and pathways of STD acquisition. We recommend periodic reviews of the level of support of predictors included in clinical guidelines to ensure that they are in accordance with empirical evidence.