Building Gender-Specific Sexually Transmitted Infection Risk Prediction Models Using CatBoost Algorithm and NHANES Data

Building Gender-Specific Sexually Transmitted Infection Risk Prediction Models Using CatBoost Algorithm and NHANES Data | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Building Gender-Specific Sexually Transmitted Infection Risk Prediction Models Using CatBoost Algorithm and NHANES Data Mengjie Hu, Han Peng, Xuan Zhang, Lefeng Wang, Jingjing Ren This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3020338/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 24 Jan, 2024 Read the published version in BMC Medical Informatics and Decision Making → Version 1 posted 3 You are reading this latest preprint version Abstract Aims Sexually transmitted infections (STIs) are a significant global public health challenge due to their high incidence rate and potential for severe consequences when early intervention is neglected. Research shows an upward trend in absolute cases and DALY numbers of STIs, with syphilis, chlamydia, trichomoniasis, and genital herpes exhibiting an increasing trend in age-standardized rate (ASR) from 2010 to 2019. Machine learning (ML) presents significant advantages in disease prediction, with several studies exploring its potential for STI prediction. The objective of this study is to build males-based and females-based STI risk prediction models based on the CatBoost algorithm using data from the National Health and Nutrition Examination Survey (NHANES) for training and validation, with sub-group analysis performed on each STI. The female sub-group also includes human papilloma virus (HPV) infection. Methods The study utilized data from the National Health and Nutrition Examination Survey (NHANES) program to build males-based and females-based STI risk prediction models using the CatBoost algorithm. Data was collected from 12,053 participants aged 18 to 59 years old, with general demographic characteristics and sexual behavior questionnaire responses included as features. The SMOTE algorithm was used to address data imbalance, and 15 machine learning algorithms were evaluated before ultimately selecting the CatBoost algorithm. The SHAP method was employed to enhance interpretability by identifying feature importance in the model's STIs risk prediction. Results The CatBoost classifier achieved AUC values of 0.7891, 0.6558, 0.6607, 0.6118 and 0.6932 for predicting chlamydia, genital herpes, genital warts, gonorrhea, and overall STIs infections among males.The CatBoost classifier achieved AUC values of 0.7082, 0.647, 0.6767, 0.8459, 0.6929 and 0.7005 for predicting chlamydia, genital herpes, genital warts, gonorrhea, HPV and overall STIs infections among females. Sexually transmitted infections CatBoost algorithm NHANES data SHAP algorithm Figures Figure 1 Figure 2 1. Introduction Sexually transmitted infections (STIs) pose a significant global public health challenge due to their high incidence rates, which exert substantial pressure on both family and national healthcare budgets while concurrently impairing individual quality of life [ 1 , 2 ]. Moreover, the widespread issue of delayed STI diagnosis raises the risk of severe consequences such as compromised reproductive and neonatal health when early intervention is neglected [ 3 ]. Research indicates an upward trend in both absolute cases and disability-adjusted life years (DALYs) for STIs between 1990 and 2019 [ 4 ]. Syphilis, chlamydia, trichomoniasis, and genital herpes have demonstrated an increasing trend in age-standardized rates (ASRs) from 2010 to 2019 [ 5 ]. Consequently, STIs remain a persistent global public health concern. Furthermore, since 2010, the age-standardized incidence rate among young people has exhibited an upward trend, particularly regarding syphilis [ 4 ]. As such, early intervention through STI prediction is crucial [ 6 ]. Machine learning (ML) offers significant advantages in disease prediction, with numerous studies already exploring its potential for STI prediction. Bao et al.[ 7 ] aimed to develop and evaluate the performance of machine learning models in predicting the diagnosis of HIV and STIs based on a large retrospective cohort of Australian men who have sex with men (MSM). Fieggen et al.[ 8 ] discussed crucial considerations when selecting variables for model development and evaluating the performance of various machine learning algorithms, as well as the potential role of emerging tools such as Shapley Additive Explanations in understanding and decomposing these models in the context of HIV. Xu et al.[ 9 ] sought to identify determinants and predict chlamydia re-testing and re-infection within one year among heterosexuals with chlamydia to pinpoint potential PDPT candidates. Our study developed male-based and female-based STIs risk prediction models using the CatBoost algorithm, employing data from the National Health and Nutrition Examination Survey (NHANES) for training and validation. Sub-group analyses were conducted for each STI, including genital herpes, genital warts, gonorrhea, and chlamydia infections. The female sub-group also encompassed human papillomavirus (HPV) infection. 2. Methods 2.1 Data source NHANES is a series of studies aimed at evaluating the health and nutritional status of adults and children in the United States [ 10 ]. As a significant initiative of the National Center for Health Statistics (NCHS), NHANES contributes to the Centers for Disease Control and Prevention's (CDC) mission by generating essential health statistics for the nation. Data were collected from the NHANES datasets spanning 2009 to 2016, encompassing 19,998 individuals aged between 18 and 59 years. A total of 7,945 individuals were excluded due to their responses to the Sexual Behavior Questionnaire, specifically those who provided answers other than "yes" or "no" regarding whether a doctor had ever informed them of having HPV, genital herpes, genital warts, gonorrhea, or chlamydia, or those who refused to answer the questions. Consequently, the final sample comprised 12,053 participants, including 6,163 females and 5,890 males. 2.2 Feature selection The study incorporated general demographic characteristics (gender, age, education level, and marital status) along with questions from the Sexual Behavior Questionnaire (codes and corresponding questions are accessible on the NHANES website: https://wwwn.cdc.gov/nchs/nhanes ). For instances of missing data, imputation methods such as mode, median, or mean imputation were employed depending on the data ty 2.3 Algorithm Owing to the considerable imbalance in the datasets, random undersampling could cause substantial data loss, while random oversampling might result in overfitting. To tackle this data imbalance issue, we employed the Synthetic Minority Over-sampling Technique (SMOTE) algorithm. We carried out risk prediction modeling for various STIs cases within the study population using 15 unique machine learning algorithms. After thoroughly evaluating and comparing the performance of these models, we ultimately chose the CatBoost algorithm. The CatBoost algorithm is a robust and highly efficient gradient boosting framework extensively employed in machine learning applications [ 11 ]. It outperforms traditional gradient boosting techniques, especially when managing complex datasets featuring numerous categorical variables. The strength of the CatBoost algorithm lies in its capacity to handle feature interactions accurately while minimizing overfitting, thereby ensuring exceptional predictive power. PyCaret 2.3.1 in Jupyter Notebook was used to train and validate the CatBoost classifier. 2.5 interpretability We utilized the SHAP (SHapley Additive exPlanations) method to identify feature importance in the CatBoost model's STIs risk prediction and enhance its interpretability. 3. Results 3.1 Basic characteristics Table 1 presents the demographic characteristics of the study participants. Among male subjects, the prevalence rates were as follows: Chlamydia infection at 41 (0.70%), genital herpes at 126 (2.14%), genital warts at 159 (2.70%), and gonorrhea at 26 (0.44%). Among female subjects, the prevalence rates were: Chlamydia infection at 92 (1.49%), genital herpes at 341 (5.53%), genital warts at 305 (4.95%), gonorrhea at 20 (0.32%), and HPV infection at 556 (9.02%). Table 1 Demographics of datasets Male (n = 5,890) Female (n = 6,163) Age (years) 39.04 ± 11.37 39.18 ± 11.36 Education (years) Less than 9th grade 338(5.73%) 321(5.21%) 9-11th grade(Include 12th grade with no diploma) 856(14.53%) 734(11.91%) High school graduate/GED or equivalent 1426(24.21%) 1234(20.02%) Some college or AA degree 1759(29.86%) 2168(35.18%) College graduate or above 1511(25.65%) 1706(27.68%) Marital status Married 3022(51.31%) 2959(48.01%) Widowed 41(0.70%) 116(1.88%) Divorced 483(8.20%) 716(11.62%) Separated 164(2.78%) 262(4.25%) Never married 1483(25.18%) 1447(23.48%) Living with partner 697(11.83%) 663(10.76%) Chlamydia 41(0.70%) 92(1.49%) Genital herpes 126(2.14%) 341(5.53%) Genital warts 159(2.70%) 305(4.95%) Gonorrhea 26(0.44%) 20(0.32%) Hpv / 556(9.02%) 3.2 Classification performance The CatBoost classifier was trained and validated using ten-fold cross-validation to estimate out-of-sample performance. Evaluation metrics included AUC, recall, accuracy, F1-score, kappa value, and precision. Tables 2 and 3 display the performance of the CatBoost classifier in predicting STI infection risk among male and female populations, respectively. For males, the CatBoost classifier achieved AUC values of 0.7891, 0.6558, 0.6607, and 0.6118 for predicting chlamydia, genital herpes, genital warts, and gonorrhea infections; it also achieved an AUC value of 0.6932 for overall STIs. For females, the classifier attained AUC values of 0.7082, 0.6470, 0.6767, 0.8459 for chlamydia, genital herpes, genital warts, and gonorrhea infections; it also reached AUC values of 0.6929 for HPV infection and 0.7005 for overall STIs. Table 2 Classification Performance of CatBoost classifier in male populations Male-Label Accuracy AUC Recall Prec. F1 Kappa MCC Chlamydia 0.9925 0.7891 0 0 0 -0.0012 -0.0013 Genital herpes 0.9784 0.6558 0 0 0 -0.0004 -0.0007 Genital warts 0.9738 0.6607 0.0091 0.1 0.0167 0.0158 0.029 Gonorrhea 0.9947 0.6118 0 0 0 -0.0003 -0.0003 STIs 0.95 0.6932 0 0 0 -0.0019 -0.0045 Table 3 Classification Performance of CatBoost classifier in female populations Female-Label Accuracy AUC Recall Prec. F1 Kappa MCC Chlamydia 0.9849 0.7082 0 0 0 -0.0004 -0.0006 Genital herpes 0.9448 0.647 0.0083 0.1 0.0154 0.0129 0.0235 Genital warts 0.9485 0.6767 0.0045 0.1 0.0087 0.0066 0.0176 Gonorrhea 0.9968 0.8459 0 0 0 -0.0002 -0.0002 HPV 0.9075 0.6929 0.0262 0.255 0.0472 0.0327 0.0573 STIs 0.8153 0.7005 0.1031 0.3726 0.1612 0.0917 0.1172 3.3 Model interpretation: Shapley Additive exPlanations (SHAP) Utilizing the SHAP algorithm, the feature ranking interpretation of the CatBoost classifier reveals the top 20 most influential characteristics for predicting outcomes in both male and female populations (Figs. 1 and 2 ). In general, the top three significant predictors of male chlamydia infection risk are identified as sxd510 (female sex partners per year), sxq827 (female vaginal sex partners per year), and sxq251_2 (instances of sex without a condom per year). The top three important predictors for male genital herpes risk include sxq824 (female vaginal sex partners in lifetime), dmdeduc2_5 (education level), and sxq610 (instances of vaginal or anal sex per year). For male genital warts risk, the top three important predictors are dmdeduc2_4 (education level), sxq251_1 (instances of sex without a condom per year), and sxq806_2 (ever having had anal sex with a woman). The top three important predictors for male gonorrhea risk consist of sxq280_1 (circumcision status), sxd510 (female sex partners per year), and sxq645_2 (protection use during oral sex). Lastly, the top three important predictors for total male STI risk include sxq610_3 (instances of vaginal or anal sex per year), sxd806_1 (ever having had anal sex with a woman), and sxq171(female sex partners in lifetime). The top three significant predictors of female chlamydia infection risk are identified as dmdmartl_5 (marital status), sxq294_1 (self-described sexual orientation for females), and ridageye (age per year). The top three important predictors for female genital herpes risk include sxd101 (male sex partners in lifetime), sxq706_1 (ever having had anal sex with a man), and dmdeduc2_4 (education level). For female genital warts risk, the top three important predictors are sxq706_2 (ever having had anal sex with a man), ridageye (age per year), and dmdeduc2_5 (education level). The top three important predictors for female gonorrhea risk consist of sxq727 (male vaginal sex partners per year), sxd648_2 (instances of sex with a new partner per year), and dmdmartl_5 (marital status). The top three significant predictors of female HPV infection risk include ridageye(age per year), sxq706_2(ever having had anal sex with a man) and sxq624(male oral sex partners in lifetime). Lastly, the top three important predictors for total female STI risk include sxd101(male sex partners in lifetime), sxq706_2(ever having had anal sex with a man) and dmdeduc2_4(education level). 4. Discussion We developed risk prediction models for chlamydia, genital herpes, genital warts, and gonorrhea in male populations, as well as for chlamydia, genital herpes, genital warts, gonorrhea, and HPV infection in female populations using the CatBoost algorithm. The AUC values of these models range from 0.6 to 0.85, with overall STI prediction AUC values of 0.6932 and 0.7005 for males and females respectively. Lastly, we conducted an interpretability analysis on the models and obtained feature importance rankings for various prediction models. Previous studies have employed machine learning to predict the risk of STI occurrence. For example, risk prediction tools have been developed to forecast HIV and STIs over the next 12 months [ 12 ], demonstrating acceptable performance for HIV (AUC = 75.0), syphilis (AUC = 73.0), gonorrhea (AUC = 67.12), and chlamydia (AUC = 0.67) infection prediction in test datasets. Xianglong Xu et al.[ 13 ] developed a machine learning-based STI risk prediction tool, MySTIRisk, which exhibits promising performance on the testing dataset (AUC for HIV = 0.78; AUC for syphilis = 0.84; AUC for gonorrhea = 0.78; AUC for chlamydia = 0.70). Furthermore, it demonstrated stable performance on both external validation data from 2019 (AUC for HIV = 0.79; AUC for syphilis = 0.85; AUC for gonorrhea = 0.81; AUC for chlamydia = 0.69) and data from 2020–2021 (AUC for HIV = 0.71; AUC for syphilis = 0.84; AUC for gonorrhea = 0.79; AUC for chlamydia = 0.69). These studies enable individuals to comfortably predict their own risk of HIV and STIs from home. Given that HIV poses higher risks than other STIs, more research has focused on early detection and identification of HIV [ 14 – 16 ]. Our results are comparable to those of the aforementioned studies in terms of predictive performance. We conducted a subgroup analysis based on gender since the likelihood of contracting STIs differs between males and females due to differences in reproductive system structures, aiming to improve our predictive model's accuracy. Additionally, we carried out an interpretability analysis on our models to assist clinical practitioners in better understanding the models and asking more targeted questions (focusing on the top-ranking features) during actual consultations and screening processes. Nonetheless, our study presents several limitations: 1. In the classification models, numerous models exhibit extremely low recall and precision rates, some even as low as 0. This primarily results from the highly imbalanced ratio of positive and negative data in the samples. Although we employed the SMOTE algorithm to address imbalanced data, the issue remains significant; 2. Factors influencing STIs may vary across different races. Furthermore, this study did not conduct external validation of the model on distinct datasets; hence, the model's generalizability has not been tested; 3. The questionnaire data in the database lacks information on HIV and syphilis infection, rendering it impossible to predict associated risks. To mitigate the aforementioned limitations, future research can implement the following improvements: 1. Utilize additional sample augmentation methods, such as enhanced SMOTE algorithms [ 17 ] and Generative Adversarial Networks (GANs) [ 18 , 19 ], to tackle the issue of imbalanced data; 2. Collect more data from diverse races and regions for external validation and generalization testing of the model; 3. In designing sexual behavior questionnaires, incorporate more data collection on various sexually transmitted diseases to enhance the model's overall predictive capacity for related infection risks. In future research, the focus could be directed towards the prevention of STIs in high-risk populations and the intelligent management of STIs-affected individuals. On one hand, developing high-performance early screening models for STIs can expedite the identification of affected populations. On the other hand, for existing diagnosed STIs populations, personalized treatment methods employing artificial intelligence can be adopted to reduce management costs and enhance treatment success rates across different population groups. 5. Conclusion This study found that the CatBoost classifier achieved good classification performance in predicting the risk of different STIs among both male and female populations. The SHAP algorithm identified several important predictors for each STI, with certain demographic characteristics and sexual behaviors being consistently significant across different infections. These findings can inform targeted prevention and intervention efforts to reduce the burden of STIs in the population. Declarations Ethics approval and consent to participate The present study utilized data from the National Health and Nutrition Examination Survey (NHANES), which is conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC). NHANES is a publicly available database that collects health information from a nationally representative sample of the US population. Ethical approval for NHANES was obtained by NCHS, and all participants provided written informed consent prior to their participation in the survey. The consent form explained the purpose of the survey, procedures involved, potential risks and benefits, confidentiality measures, and the right to withdraw from the survey at any time without penalty. Participants were also informed that their data would be kept confidential and used only for research purposes. Consent for publication Not applicable. Availability of data and materials The datasets generated and/or analyzed during the current study are available in the NHANES database (https://www.cdc.gov/nchs/nhanes/index.htm). The data used in this study were accessed through a public access repository and no identifiable information was obtained. Competing interests The authors declare that they have no competing interests. Funding None Authors' contributions Mengjie Hu, Han Peng and Xuan Zhang wrote the main manuscript text and Lefeng Wang prepared figures 1,2. Jingjing Ren provide research ideas. All authors reviewed the manuscript. References Ramchandani MS, Golden MR. Confronting Rising STIs in the Era of PrEP and Treatment as Prevention. Curr HIV/AIDS Rep. 2019;16:244–56. Zhang J, Ma B, Han X, Ding S, Li Y. Global, regional, and national burdens of HIV and other sexually transmitted infections in adolescents and young adults aged 10–24 years from 1990 to 2019: a trend analysis based on the Global Burden of Disease Study 2019. Lancet Child Adolesc Health. 2022 Nov;6(11):763–76. Lemoh C, Guy R, Yohannes K, Lewis J, Street A, Biggs B, Hellard M. Delayed diagnosis of HIV infection in Victoria 1994 to 2006. Sex Health. 2009 Jun;6(2):117–22. Zheng Y, Yu Q, Lin Y, Zhou Y, Lan L, Yang S, Wu J. Global burden and trends of sexually transmitted infections from 1990 to 2019: an observational trend study. Lancet Infect Dis. 2022 Apr;22(4):541–51. Du M, Yan W, Jing W, Qin C, Liu Q, Liu M, Liu J. Increasing incidence rates of sexually transmitted infections from 2010 to 2019: an analysis of temporal trends by geographical regions and age groups from the 2019 Global Burden of Disease Study. BMC Infect Dis. 2022 Jun;26(1):574. Sangani P, Rutherford G, Wilkinson D. Population-based interventions for reducing sexually transmitted infections, including HIV infection. Cochrane Database Syst Rev. 2004;(2):CD001220. Bao Y, Medland NA, Fairley CK, Wu J, Shang X, Chow EPF, Xu X, Ge Z, Zhuang X, Zhang L. Predicting the diagnosis of HIV and sexually transmitted infections among men who have sex with men using machine learning approaches. J Infect. 2021 Jan;82(1):48–59. Fieggen J, Smith E, Arora L, Segal B. The role of machine learning in HIV risk prediction. Front Reprod Health 2022 Dec 22;4:1062387. Xu X, Chow EPF, Fairley CK, Chen M, Aguirre I, Goller J, Hocking J, Carvalho N, Zhang L, Ong JJ. Determinants and prediction of Chlamydia trachomatis re-testing and re-infection within 1 year among heterosexuals with chlamydia attending a sexual health clinic. Front Public Health 2023 Jan 13;10:1031372. Andresen S, Balakrishna S, Mugglin C, Schmidt AJ, Braun DL, Marzel A, Doco Lecompte T, Darling KE, Roth JA, Schmid P, Bernasconi E, Günthard HF, Rauch A, Kouyos RD, Salazar-Vizcaya L, Swiss HIV. Cohort Study. Unsupervised machine learning predicts future sexual behaviour and sexually transmitted infections among HIV-positive men who have sex with men. PLoS Comput Biol 2022 Oct 27;18(10):e1010559. Zipf G, Chiappa M, Porter KS, Ostchega Y, Lewis BG, Dostal J. National health and nutrition examination survey: plan and operations, 1999–2010. Vital Health Stat. 2013;1(56):1–37. Hancock JT, Khoshgoftaar TM. CatBoost for big data: an interdisciplinary review. J Big Data. 2020;7(1):94. Xu X, Ge Z, Chow EPF, Yu Z, Lee D, Wu J, Ong JJ, Fairley CK, Zhang L. A Machine-Learning-Based Risk-Prediction Tool for HIV and Sexually Transmitted Infections Acquisition over the Next 12 Months. J Clin Med. 2022 Mar;25(7):1818. Xu X, Yu Z, Ge Z, Chow EPF, Bao Y, Ong JJ, Li W, Wu J, Fairley CK, Zhang L. Web-Based Risk Prediction Tool for an Individual's Risk of HIV and Sexually Transmitted Infections Using Machine Learning Algorithms: Development and External Validation Study. J Med Internet Res. 2022 Aug 25;24(8):e37850. Bao Y, Medland NA, Fairley CK, Wu J, Shang X, Chow EPF, Xu X, Ge Z, Zhuang X, Zhang L. Predicting the diagnosis of HIV and sexually transmitted infections among men who have sex with men using machine learning approaches. J Infect. 2021 Jan;82(1):48–59. He J, Li J, Jiang S, Cheng W, Jiang J, Xu Y, Yang J, Zhou X, Chai C, Wu C. Application of machine learning algorithms in predicting HIV infection among men who have sex with men: Model development and validation. Front Public Health 2022 Aug 25;10:967681. Kosolwattana T, Liu C, Hu R, Han S, Chen H, Lin Y. A self-inspected adaptive SMOTE algorithm (SASMOTE) for highly imbalanced data classification in healthcare. BioData Min 2023 Apr 25;16(1):15. Kwon C, Park S, Ko S, Ahn J. Increasing prediction accuracy of pathogenic staging by sample augmentation with a GAN. PLoS One 2021 Apr 27;16(4):e0250458. Lan T, Hu Q, Liu X, He K, Yang C. Arrhythmias Classification Using Short-Time Fourier Transform and GAN Based Data Augmentation. Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:308–11. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 24 Jan, 2024 Read the published version in BMC Medical Informatics and Decision Making → Version 1 posted Editor assigned by journal 13 Jun, 2023 Submission checks completed at journal 11 Jun, 2023 First submitted to journal 04 Jun, 2023 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3020338","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":208715194,"identity":"13a56020-08f4-42b3-83f0-5a9a170d81db","order_by":0,"name":"Mengjie Hu","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Mengjie","middleName":"","lastName":"Hu","suffix":""},{"id":208715195,"identity":"e7246a0f-47f1-476f-b1e5-51d913b2c46b","order_by":1,"name":"Han Peng","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Han","middleName":"","lastName":"Peng","suffix":""},{"id":208715198,"identity":"a1c96fe9-b4ea-484b-97cb-40be7bba2a76","order_by":2,"name":"Xuan Zhang","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Xuan","middleName":"","lastName":"Zhang","suffix":""},{"id":208715200,"identity":"beb46a0e-986d-4894-aed4-851d9b83123e","order_by":3,"name":"Lefeng Wang","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Lefeng","middleName":"","lastName":"Wang","suffix":""},{"id":208715202,"identity":"46cffb61-dcee-43b3-9eb6-78c0a9e37e44","order_by":4,"name":"Jingjing Ren","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIie3RMQrCMBSA4VcKcYl0zdRcIaWggx4mLnax0Kk4pgideoCKg2cQLxAJOIXODg56g7oJIto6ONq4CebfHryPPAiAzfaD0V4mGJ+PfQDZjMiABIWSSa2noTmBw5RflrmaiNdkQhwxC7b93I3WmWZQpwq8lfhMeqDDEFcozoRmTlkpIEfZ8UpWDEKc4ngBmrn9XAEjvOMyhYc3jEiEWnI3InvEgzJnHLfEMSFB4UpW64bBPtkVVYTJoYNQem6/8kFpqTanazryvbLrsHdEvj4Tm+43eeKLZZvNZvurnnQTRwqxx9ayAAAAAElFTkSuQmCC","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Jingjing","middleName":"","lastName":"Ren","suffix":""}],"badges":[],"createdAt":"2023-06-04 10:29:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3020338/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3020338/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12911-024-02426-1","type":"published","date":"2024-01-24T15:16:32+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":38480582,"identity":"edc4036b-0a85-4941-95aa-76a4fedbe745","added_by":"auto","created_at":"2023-06-13 15:23:14","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":119606,"visible":true,"origin":"","legend":"\u003cp\u003eThe CatBoost classifiers for predicting chlamydia(A), genital herpes(B), genital warts(C), gonorrhea(D), and overall STIs(E) based on the SHAP algorithm in male populations.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-3020338/v1/14afca1cb5d3d038df02e673.png"},{"id":38480583,"identity":"f1532546-34e6-48f8-acf0-d1ee235afc9d","added_by":"auto","created_at":"2023-06-13 15:23:14","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":148073,"visible":true,"origin":"","legend":"\u003cp\u003eThe CatBoost classifiers for predicting chlamydia(A), genital herpes(B), genital warts(C), gonorrhea(D), HPV(E) and overall STIs(F) based on the SHAP algorithm in female populations.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-3020338/v1/272b5ae306e25465bc6f22d4.png"},{"id":50314044,"identity":"4d0613e3-8841-4437-89cb-d71cc57927cf","added_by":"auto","created_at":"2024-01-29 15:28:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":522893,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3020338/v1/6f2c25b4-7bf9-41b1-a81b-48f30baab8e0.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Building Gender-Specific Sexually Transmitted Infection Risk Prediction Models Using CatBoost Algorithm and NHANES Data","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eSexually transmitted infections (STIs) pose a significant global public health challenge due to their high incidence rates, which exert substantial pressure on both family and national healthcare budgets while concurrently impairing individual quality of life [\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e]. Moreover, the widespread issue of delayed STI diagnosis raises the risk of severe consequences such as compromised reproductive and neonatal health when early intervention is neglected [\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e]. Research indicates an upward trend in both absolute cases and disability-adjusted life years (DALYs) for STIs between 1990 and 2019 [\u003cspan class=\"CitationRef\"\u003e4\u003c/span\u003e]. Syphilis, chlamydia, trichomoniasis, and genital herpes have demonstrated an increasing trend in age-standardized rates (ASRs) from 2010 to 2019 [\u003cspan class=\"CitationRef\"\u003e5\u003c/span\u003e]. Consequently, STIs remain a persistent global public health concern. Furthermore, since 2010, the age-standardized incidence rate among young people has exhibited an upward trend, particularly regarding syphilis [\u003cspan class=\"CitationRef\"\u003e4\u003c/span\u003e]. As such, early intervention through STI prediction is crucial [\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e\n\u003cp\u003eMachine learning (ML) offers significant advantages in disease prediction, with numerous studies already exploring its potential for STI prediction. Bao et al.[\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e] aimed to develop and evaluate the performance of machine learning models in predicting the diagnosis of HIV and STIs based on a large retrospective cohort of Australian men who have sex with men (MSM). Fieggen et al.[\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e] discussed crucial considerations when selecting variables for model development and evaluating the performance of various machine learning algorithms, as well as the potential role of emerging tools such as Shapley Additive Explanations in understanding and decomposing these models in the context of HIV. Xu et al.[\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e] sought to identify determinants and predict chlamydia re-testing and re-infection within one year among heterosexuals with chlamydia to pinpoint potential PDPT candidates.\u003c/p\u003e\n\u003cp\u003eOur study developed male-based and female-based STIs risk prediction models using the CatBoost algorithm, employing data from the National Health and Nutrition Examination Survey (NHANES) for training and validation. Sub-group analyses were conducted for each STI, including genital herpes, genital warts, gonorrhea, and chlamydia infections. The female sub-group also encompassed human papillomavirus (HPV) infection.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003e2.1 Data source\u003c/h2\u003e\n \u003cp\u003eNHANES is a series of studies aimed at evaluating the health and nutritional status of adults and children in the United States [\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e]. As a significant initiative of the National Center for Health Statistics (NCHS), NHANES contributes to the Centers for Disease Control and Prevention\u0026apos;s (CDC) mission by generating essential health statistics for the nation.\u003c/p\u003e\n \u003cp\u003eData were collected from the NHANES datasets spanning 2009 to 2016, encompassing 19,998 individuals aged between 18 and 59 years. A total of 7,945 individuals were excluded due to their responses to the Sexual Behavior Questionnaire, specifically those who provided answers other than \u0026quot;yes\u0026quot; or \u0026quot;no\u0026quot; regarding whether a doctor had ever informed them of having HPV, genital herpes, genital warts, gonorrhea, or chlamydia, or those who refused to answer the questions. Consequently, the final sample comprised 12,053 participants, including 6,163 females and 5,890 males.\u003c/p\u003e\n \u003ch2\u003e2.2 Feature selection\u003c/h2\u003e\u003cspan\u003e\n \u003cp\u003eThe study incorporated general demographic characteristics (gender, age, education level, and marital status) along with questions from the Sexual Behavior Questionnaire (codes and corresponding questions are accessible on the NHANES website: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://wwwn.cdc.gov/nchs/nhanes\u003c/span\u003e\u003c/span\u003e). For instances of missing data, imputation methods such as mode, median, or mean imputation were employed depending on the data ty\u003c/p\u003e\n \u003ch2\u003e2.3 Algorithm\u003c/h2\u003e\n \u003cp\u003eOwing to the considerable imbalance in the datasets, random undersampling could cause substantial data loss, while random oversampling might result in overfitting. To tackle this data imbalance issue, we employed the Synthetic Minority Over-sampling Technique (SMOTE) algorithm.\u003c/p\u003e\n \u003cp\u003eWe carried out risk prediction modeling for various STIs cases within the study population using 15 unique machine learning algorithms. After thoroughly evaluating and comparing the performance of these models, we ultimately chose the CatBoost algorithm.\u003c/p\u003e\n \u003cp\u003eThe CatBoost algorithm is a robust and highly efficient gradient boosting framework extensively employed in machine learning applications [\u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e]. It outperforms traditional gradient boosting techniques, especially when managing complex datasets featuring numerous categorical variables. The strength of the CatBoost algorithm lies in its capacity to handle feature interactions accurately while minimizing overfitting, thereby ensuring exceptional predictive power.\u003c/p\u003e\n \u003cp\u003ePyCaret 2.3.1 in Jupyter Notebook was used to train and validate the CatBoost classifier.\u003c/p\u003e\n \u003c/span\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\n \u003ch2\u003e2.5 interpretability\u003c/h2\u003e\n \u003cp\u003eWe utilized the SHAP (SHapley Additive exPlanations) method to identify feature importance in the CatBoost model\u0026apos;s STIs risk prediction and enhance its interpretability.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\n \u003ch2\u003e3.1 Basic characteristics\u003c/h2\u003e\n \u003cp\u003eTable \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e presents the demographic characteristics of the study participants. Among male subjects, the prevalence rates were as follows: Chlamydia infection at 41 (0.70%), genital herpes at 126 (2.14%), genital warts at 159 (2.70%), and gonorrhea at 26 (0.44%). Among female subjects, the prevalence rates were: Chlamydia infection at 92 (1.49%), genital herpes at 341 (5.53%), genital warts at 305 (4.95%), gonorrhea at 20 (0.32%), and HPV infection at 556 (9.02%).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDemographics of datasets\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"3\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMale (n\u0026thinsp;=\u0026thinsp;5,890)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eFemale (n\u0026thinsp;=\u0026thinsp;6,163)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e39.04\u0026thinsp;\u0026plusmn;\u0026thinsp;11.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e39.18\u0026thinsp;\u0026plusmn;\u0026thinsp;11.36\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEducation (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLess than 9th grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e338(5.73%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e321(5.21%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9-11th grade(Include 12th grade with no diploma)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e856(14.53%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e734(11.91%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHigh school graduate/GED or equivalent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1426(24.21%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1234(20.02%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSome college or AA degree\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1759(29.86%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2168(35.18%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCollege graduate or above\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1511(25.65%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1706(27.68%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMarital status\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMarried\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3022(51.31%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2959(48.01%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWidowed\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e41(0.70%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e116(1.88%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDivorced\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e483(8.20%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e716(11.62%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSeparated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e164(2.78%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e262(4.25%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNever married\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1483(25.18%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1447(23.48%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLiving with partner\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e697(11.83%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e663(10.76%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eChlamydia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e41(0.70%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e92(1.49%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGenital herpes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e126(2.14%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e341(5.53%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGenital warts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e159(2.70%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e305(4.95%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGonorrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e26(0.44%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e20(0.32%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHpv\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e/\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e556(9.02%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\n \u003ch2\u003e3.2 Classification performance\u003c/h2\u003e\n \u003cp\u003eThe CatBoost classifier was trained and validated using ten-fold cross-validation to estimate out-of-sample performance. Evaluation metrics included AUC, recall, accuracy, F1-score, kappa value, and precision. Tables \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e display the performance of the CatBoost classifier in predicting STI infection risk among male and female populations, respectively. For males, the CatBoost classifier achieved AUC values of 0.7891, 0.6558, 0.6607, and 0.6118 for predicting chlamydia, genital herpes, genital warts, and gonorrhea infections; it also achieved an AUC value of 0.6932 for overall STIs. For females, the classifier attained AUC values of 0.7082, 0.6470, 0.6767, 0.8459 for chlamydia, genital herpes, genital warts, and gonorrhea infections; it also reached AUC values of 0.6929 for HPV infection and 0.7005 for overall STIs.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eClassification Performance of CatBoost classifier in male populations\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMale-Label\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAUC\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRecall\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePrec.\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eF1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eKappa\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMCC\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eChlamydia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9925\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.7891\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0013\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGenital herpes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.6558\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0007\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGenital warts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9738\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.6607\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0091\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0167\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0158\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.029\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGonorrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9947\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.6118\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSTIs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.6932\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0045\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eClassification Performance of CatBoost classifier in female populations\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eFemale-Label\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAUC\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRecall\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePrec.\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eF1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eKappa\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMCC\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eChlamydia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9849\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.7082\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0006\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGenital herpes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9448\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.647\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0083\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0154\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0129\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0235\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGenital warts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9485\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.6767\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0045\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0087\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0066\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0176\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGonorrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9968\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.8459\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.0002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHPV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9075\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.6929\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0262\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.255\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.0472\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0327\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0573\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSTIs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.8153\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.7005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.1031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.3726\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.1612\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0917\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1172\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003e3.3 Model interpretation: Shapley Additive exPlanations (SHAP)\u003c/h2\u003e\n \u003cp\u003eUtilizing the SHAP algorithm, the feature ranking interpretation of the CatBoost classifier reveals the top 20 most influential characteristics for predicting outcomes in both male and female populations (Figs. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e and \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eIn general, the top three significant predictors of male chlamydia infection risk are identified as sxd510 (female sex partners per year), sxq827 (female vaginal sex partners per year), and sxq251_2 (instances of sex without a condom per year). The top three important predictors for male genital herpes risk include sxq824 (female vaginal sex partners in lifetime), dmdeduc2_5 (education level), and sxq610 (instances of vaginal or anal sex per year). For male genital warts risk, the top three important predictors are dmdeduc2_4 (education level), sxq251_1 (instances of sex without a condom per year), and sxq806_2 (ever having had anal sex with a woman). The top three important predictors for male gonorrhea risk consist of sxq280_1 (circumcision status), sxd510 (female sex partners per year), and sxq645_2 (protection use during oral sex). Lastly, the top three important predictors for total male STI risk include sxq610_3 (instances of vaginal or anal sex per year), sxd806_1 (ever having had anal sex with a woman), and sxq171(female sex partners in lifetime).\u003c/p\u003e\n \u003cp\u003eThe top three significant predictors of female chlamydia infection risk are identified as dmdmartl_5 (marital status), sxq294_1 (self-described sexual orientation for females), and ridageye (age per year). The top three important predictors for female genital herpes risk include sxd101 (male sex partners in lifetime), sxq706_1 (ever having had anal sex with a man), and dmdeduc2_4 (education level). For female genital warts risk, the top three important predictors are sxq706_2 (ever having had anal sex with a man), ridageye (age per year), and dmdeduc2_5 (education level). The top three important predictors for female gonorrhea risk consist of sxq727 (male vaginal sex partners per year), sxd648_2 (instances of sex with a new partner per year), and dmdmartl_5 (marital status). The top three significant predictors of female HPV infection risk include ridageye(age per year), sxq706_2(ever having had anal sex with a man) and sxq624(male oral sex partners in lifetime). Lastly, the top three important predictors for total female STI risk include sxd101(male sex partners in lifetime), sxq706_2(ever having had anal sex with a man) and dmdeduc2_4(education level).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eWe developed risk prediction models for chlamydia, genital herpes, genital warts, and gonorrhea in male populations, as well as for chlamydia, genital herpes, genital warts, gonorrhea, and HPV infection in female populations using the CatBoost algorithm. The AUC values of these models range from 0.6 to 0.85, with overall STI prediction AUC values of 0.6932 and 0.7005 for males and females respectively. Lastly, we conducted an interpretability analysis on the models and obtained feature importance rankings for various prediction models.\u003c/p\u003e \u003cp\u003ePrevious studies have employed machine learning to predict the risk of STI occurrence. For example, risk prediction tools have been developed to forecast HIV and STIs over the next 12 months [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e], demonstrating acceptable performance for HIV (AUC\u0026thinsp;=\u0026thinsp;75.0), syphilis (AUC\u0026thinsp;=\u0026thinsp;73.0), gonorrhea (AUC\u0026thinsp;=\u0026thinsp;67.12), and chlamydia (AUC\u0026thinsp;=\u0026thinsp;0.67) infection prediction in test datasets. Xianglong Xu et al.[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] developed a machine learning-based STI risk prediction tool, MySTIRisk, which exhibits promising performance on the testing dataset (AUC for HIV\u0026thinsp;=\u0026thinsp;0.78; AUC for syphilis\u0026thinsp;=\u0026thinsp;0.84; AUC for gonorrhea\u0026thinsp;=\u0026thinsp;0.78; AUC for chlamydia\u0026thinsp;=\u0026thinsp;0.70). Furthermore, it demonstrated stable performance on both external validation data from 2019 (AUC for HIV\u0026thinsp;=\u0026thinsp;0.79; AUC for syphilis\u0026thinsp;=\u0026thinsp;0.85; AUC for gonorrhea\u0026thinsp;=\u0026thinsp;0.81; AUC for chlamydia\u0026thinsp;=\u0026thinsp;0.69) and data from 2020\u0026ndash;2021 (AUC for HIV\u0026thinsp;=\u0026thinsp;0.71; AUC for syphilis\u0026thinsp;=\u0026thinsp;0.84; AUC for gonorrhea\u0026thinsp;=\u0026thinsp;0.79; AUC for chlamydia\u0026thinsp;=\u0026thinsp;0.69). These studies enable individuals to comfortably predict their own risk of HIV and STIs from home. Given that HIV poses higher risks than other STIs, more research has focused on early detection and identification of HIV [\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOur results are comparable to those of the aforementioned studies in terms of predictive performance. We conducted a subgroup analysis based on gender since the likelihood of contracting STIs differs between males and females due to differences in reproductive system structures, aiming to improve our predictive model's accuracy. Additionally, we carried out an interpretability analysis on our models to assist clinical practitioners in better understanding the models and asking more targeted questions (focusing on the top-ranking features) during actual consultations and screening processes.\u003c/p\u003e \u003cp\u003eNonetheless, our study presents several limitations: 1. In the classification models, numerous models exhibit extremely low recall and precision rates, some even as low as 0. This primarily results from the highly imbalanced ratio of positive and negative data in the samples. Although we employed the SMOTE algorithm to address imbalanced data, the issue remains significant; 2. Factors influencing STIs may vary across different races. Furthermore, this study did not conduct external validation of the model on distinct datasets; hence, the model's generalizability has not been tested; 3. The questionnaire data in the database lacks information on HIV and syphilis infection, rendering it impossible to predict associated risks.\u003c/p\u003e \u003cp\u003eTo mitigate the aforementioned limitations, future research can implement the following improvements: 1. Utilize additional sample augmentation methods, such as enhanced SMOTE algorithms [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] and Generative Adversarial Networks (GANs) [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], to tackle the issue of imbalanced data; 2. Collect more data from diverse races and regions for external validation and generalization testing of the model; 3. In designing sexual behavior questionnaires, incorporate more data collection on various sexually transmitted diseases to enhance the model's overall predictive capacity for related infection risks.\u003c/p\u003e \u003cp\u003eIn future research, the focus could be directed towards the prevention of STIs in high-risk populations and the intelligent management of STIs-affected individuals. On one hand, developing high-performance early screening models for STIs can expedite the identification of affected populations. On the other hand, for existing diagnosed STIs populations, personalized treatment methods employing artificial intelligence can be adopted to reduce management costs and enhance treatment success rates across different population groups.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis study found that the CatBoost classifier achieved good classification performance in predicting the risk of different STIs among both male and female populations. The SHAP algorithm identified several important predictors for each STI, with certain demographic characteristics and sexual behaviors being consistently significant across different infections. These findings can inform targeted prevention and intervention efforts to reduce the burden of STIs in the population.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe present study utilized data from the National Health and Nutrition Examination Survey (NHANES), which is conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC). NHANES is a publicly available database that collects health information from a nationally representative sample of the US population.\u003c/p\u003e\n\u003cp\u003eEthical approval for NHANES was obtained by NCHS, and all participants provided written informed consent prior to their participation in the survey. The consent form explained the purpose of the survey, procedures involved, potential risks and benefits, confidentiality measures, and the right to withdraw from the survey at any time without penalty. Participants were also informed that their data would be kept confidential and used only for research purposes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analyzed during the current study are available in the NHANES database (https://www.cdc.gov/nchs/nhanes/index.htm). The data used in this study were accessed through a public access repository and no identifiable information was obtained.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMengjie Hu, Han Peng and Xuan Zhang wrote the main manuscript text and Lefeng Wang prepared figures 1,2. Jingjing Ren provide research ideas. All authors reviewed the manuscript.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eRamchandani MS, Golden MR. Confronting Rising STIs in the Era of PrEP and Treatment as Prevention. Curr HIV/AIDS Rep. 2019;16:244\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang J, Ma B, Han X, Ding S, Li Y. Global, regional, and national burdens of HIV and other sexually transmitted infections in adolescents and young adults aged 10\u0026ndash;24 years from 1990 to 2019: a trend analysis based on the Global Burden of Disease Study 2019. Lancet Child Adolesc Health. 2022 Nov;6(11):763\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLemoh C, Guy R, Yohannes K, Lewis J, Street A, Biggs B, Hellard M. Delayed diagnosis of HIV infection in Victoria 1994 to 2006. Sex Health. 2009 Jun;6(2):117\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng Y, Yu Q, Lin Y, Zhou Y, Lan L, Yang S, Wu J. Global burden and trends of sexually transmitted infections from 1990 to 2019: an observational trend study. Lancet Infect Dis. 2022 Apr;22(4):541\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDu M, Yan W, Jing W, Qin C, Liu Q, Liu M, Liu J. Increasing incidence rates of sexually transmitted infections from 2010 to 2019: an analysis of temporal trends by geographical regions and age groups from the 2019 Global Burden of Disease Study. BMC Infect Dis. 2022 Jun;26(1):574.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSangani P, Rutherford G, Wilkinson D. Population-based interventions for reducing sexually transmitted infections, including HIV infection. Cochrane Database Syst Rev. 2004;(2):CD001220.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBao Y, Medland NA, Fairley CK, Wu J, Shang X, Chow EPF, Xu X, Ge Z, Zhuang X, Zhang L. Predicting the diagnosis of HIV and sexually transmitted infections among men who have sex with men using machine learning approaches. J Infect. 2021 Jan;82(1):48\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFieggen J, Smith E, Arora L, Segal B. The role of machine learning in HIV risk prediction. Front Reprod Health 2022 Dec 22;4:1062387.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu X, Chow EPF, Fairley CK, Chen M, Aguirre I, Goller J, Hocking J, Carvalho N, Zhang L, Ong JJ. Determinants and prediction of Chlamydia trachomatis re-testing and re-infection within 1 year among heterosexuals with chlamydia attending a sexual health clinic. Front Public Health 2023 Jan 13;10:1031372.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndresen S, Balakrishna S, Mugglin C, Schmidt AJ, Braun DL, Marzel A, Doco Lecompte T, Darling KE, Roth JA, Schmid P, Bernasconi E, G\u0026uuml;nthard HF, Rauch A, Kouyos RD, Salazar-Vizcaya L, Swiss HIV. Cohort Study. Unsupervised machine learning predicts future sexual behaviour and sexually transmitted infections among HIV-positive men who have sex with men. PLoS Comput Biol 2022 Oct 27;18(10):e1010559.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZipf G, Chiappa M, Porter KS, Ostchega Y, Lewis BG, Dostal J. National health and nutrition examination survey: plan and operations, 1999\u0026ndash;2010. Vital Health Stat. 2013;1(56):1\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHancock JT, Khoshgoftaar TM. CatBoost for big data: an interdisciplinary review. J Big Data. 2020;7(1):94.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu X, Ge Z, Chow EPF, Yu Z, Lee D, Wu J, Ong JJ, Fairley CK, Zhang L. A Machine-Learning-Based Risk-Prediction Tool for HIV and Sexually Transmitted Infections Acquisition over the Next 12 Months. J Clin Med. 2022 Mar;25(7):1818.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu X, Yu Z, Ge Z, Chow EPF, Bao Y, Ong JJ, Li W, Wu J, Fairley CK, Zhang L. Web-Based Risk Prediction Tool for an Individual's Risk of HIV and Sexually Transmitted Infections Using Machine Learning Algorithms: Development and External Validation Study. J Med Internet Res. 2022 Aug 25;24(8):e37850.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBao Y, Medland NA, Fairley CK, Wu J, Shang X, Chow EPF, Xu X, Ge Z, Zhuang X, Zhang L. Predicting the diagnosis of HIV and sexually transmitted infections among men who have sex with men using machine learning approaches. J Infect. 2021 Jan;82(1):48\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe J, Li J, Jiang S, Cheng W, Jiang J, Xu Y, Yang J, Zhou X, Chai C, Wu C. Application of machine learning algorithms in predicting HIV infection among men who have sex with men: Model development and validation. Front Public Health 2022 Aug 25;10:967681.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKosolwattana T, Liu C, Hu R, Han S, Chen H, Lin Y. A self-inspected adaptive SMOTE algorithm (SASMOTE) for highly imbalanced data classification in healthcare. BioData Min 2023 Apr 25;16(1):15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwon C, Park S, Ko S, Ahn J. Increasing prediction accuracy of pathogenic staging by sample augmentation with a GAN. PLoS One 2021 Apr 27;16(4):e0250458.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLan T, Hu Q, Liu X, He K, Yang C. Arrhythmias Classification Using Short-Time Fourier Transform and GAN Based Data Augmentation. Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:308\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Sexually transmitted infections, CatBoost algorithm, NHANES data, SHAP algorithm","lastPublishedDoi":"10.21203/rs.3.rs-3020338/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3020338/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eAims\u003c/h2\u003e \u003cp\u003eSexually transmitted infections (STIs) are a significant global public health challenge due to their high incidence rate and potential for severe consequences when early intervention is neglected. Research shows an upward trend in absolute cases and DALY numbers of STIs, with syphilis, chlamydia, trichomoniasis, and genital herpes exhibiting an increasing trend in age-standardized rate (ASR) from 2010 to 2019. Machine learning (ML) presents significant advantages in disease prediction, with several studies exploring its potential for STI prediction. The objective of this study is to build males-based and females-based STI risk prediction models based on the CatBoost algorithm using data from the National Health and Nutrition Examination Survey (NHANES) for training and validation, with sub-group analysis performed on each STI. The female sub-group also includes human papilloma virus (HPV) infection.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThe study utilized data from the National Health and Nutrition Examination Survey (NHANES) program to build males-based and females-based STI risk prediction models using the CatBoost algorithm. Data was collected from 12,053 participants aged 18 to 59 years old, with general demographic characteristics and sexual behavior questionnaire responses included as features. The SMOTE algorithm was used to address data imbalance, and 15 machine learning algorithms were evaluated before ultimately selecting the CatBoost algorithm. The SHAP method was employed to enhance interpretability by identifying feature importance in the model's STIs risk prediction.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe CatBoost classifier achieved AUC values of 0.7891, 0.6558, 0.6607, 0.6118 and 0.6932 for predicting chlamydia, genital herpes, genital warts, gonorrhea, and overall STIs infections among males.The CatBoost classifier achieved AUC values of 0.7082, 0.647, 0.6767, 0.8459, 0.6929 and 0.7005 for predicting chlamydia, genital herpes, genital warts, gonorrhea, HPV and overall STIs infections among females.\u003c/p\u003e","manuscriptTitle":"Building Gender-Specific Sexually Transmitted Infection Risk Prediction Models Using CatBoost Algorithm and NHANES Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-06-13 15:23:09","doi":"10.21203/rs.3.rs-3020338/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorAssigned","content":"","date":"2023-06-13T10:00:21+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2023-06-11T05:52:06+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Informatics and Decision Making","date":"2023-06-04T10:23:34+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"71675181-5486-414e-b1b7-f04e5dfe56da","owner":[],"postedDate":"June 13th, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-01-29T15:25:33+00:00","versionOfRecord":{"articleIdentity":"rs-3020338","link":"https://doi.org/10.1186/s12911-024-02426-1","journal":{"identity":"bmc-medical-informatics-and-decision-making","isVorOnly":false,"title":"BMC Medical Informatics and Decision Making"},"publishedOn":"2024-01-24 15:16:32","publishedOnDateReadable":"January 24th, 2024"},"versionCreatedAt":"2023-06-13 15:23:09","video":"","vorDoi":"10.1186/s12911-024-02426-1","vorDoiUrl":"https://doi.org/10.1186/s12911-024-02426-1","workflowStages":[]},"version":"v1","identity":"rs-3020338","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3020338","identity":"rs-3020338","version":["v1"]},"buildId":"-HB7Z8yhvgn0wM9Nzuekk","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}