Assessing the effectiveness of HIV/STI risk communication displays among Melbourne Sexual Health Centre attendees: a cross-sectional, observational and vignette-based study.

INTRODUCTION: Increasing rates of sexually transmitted infections (STIs) over the past decade underscore the need for early testing and treatment. Communicating HIV/STI risk effectively can promote individuals' intention to test, which is critical for the prevention and control of HIV/STIs. We aimed to determine which visual displays of risk would be the most likely to increase testing or use of prevention strategies. METHODS: A vignette-based cross-sectional survey was conducted with 662 clients (a median age of 30 years (IQR: 25-36), 418 male, 203 female, 41 other genders) at a sexual health clinic in Melbourne, Australia, between February and June 2023. Participants viewed five distinct hypothetical formats, presented in a randomised order, designed to display the same level of high risk for HIV/STIs: icon array, colour-coded risk metre, colour-coded risk bar, detailed text report and guideline recommendation. They reported their perceived risk, concern and intent to test for each risk display. Associations between the format of the risk display and the intention to test for HIV/STI were analysed using logistic regression. RESULTS: About 378 (57%) of participants expressed that the risk metre was the easiest to understand. The risk metre (adjusted OR (AOR)=2.44, 95% CI=1.49 to 4.01) and risk bar (AOR=2.08, CI=1.33 to 3.27) showed the greatest likelihood of testing compared with the detailed text format. The icon array was less impactful (AOR=0.73, CI=0.57 to 0.94). The risk metre also elicited the most concern but was the most preferred and understood. High-risk perception and concern levels were strongly associated with their intention to have an HIV/STI test. CONCLUSIONS: Displaying risk differently affects an individual's perceived risk of an HIV/STI and influences their intention to test.

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.