Development and Validation of a Machine Learning Model for Hepatitis C Virus Exposure: A Demographic Screening Approach for the US Population

Development and Validation of a Machine Learning Model for Hepatitis C Virus Exposure: A Demographic Screening Approach for the US Population | 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 Development and Validation of a Machine Learning Model for Hepatitis C Virus Exposure: A Demographic Screening Approach for the US Population Dorian G Ding, Taoyi Chen, Yu Sheng, Jeffrey S.H. Lin, Ye Yuan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9361729/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Hepatitis C virus (HCV) remains underdiagnosed in the United States despite recommendations for universal screening. A simple approach based on readily available demographic information may help target screening in settings where screening implementation continues to be incomplete. Methods We analyzed 10 NHANES cycles (1999–2014 and 2017–2023) and defined HCV exposure as a positive HCV antibody or RNA result. Using sex, birth year, race/ethnicity, birthplace, and income-to-poverty ratio, we trained and compared logistic regression (LR) and machine learning models in training and validation cohorts (48,434 and 20,762 participants, respectively). Model performance was evaluated based on sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiver operating characteristic curve (AUROC). A web-based calculator was developed to facilitate bedside HCV screening. Results 69,196 participants were included, with 967 showing evidence of HCV exposure. Weighted HCV prevalence remained relatively stable across cycles, ranging from 1.22% to 1.93%. The prevalence did not change significantly after the pandemic. Earlier birth year, male sex, non-Hispanic Black race, US birth, and lower income-to-poverty ratio were independently associated with HCV exposure. XGBoost performed better than LR in the validation cohort (AUROC 0.860 vs 0.762, p < 0.001). Predicted risk separated the population clearly: observed HCV prevalence increased from 0.05% in the lowest-risk decile to 7.95% in the highest, with the top decile containing 58.3% of participants with HCV exposure and the top three deciles containing 85.5%. Conclusions Five demographic variables were sufficient to build a useful HCV risk model in a nationally representative US sample. Most HCV-exposed individuals were concentrated in the highest predicted-risk groups, suggesting that this approach could help prioritize and optimize testing where universal screening uptake remains incomplete. As no laboratory data is required, it may also be practical in data-limited settings and adaptable in other health systems. Hepatitis C Prevalence Demographics Screening Machine Learning Figures Figure 1 Figure 2 Figure 3 Figure 4 INTRODUCTION Hepatitis C virus (HCV) remains a significant public health concern worldwide. In 2020, an estimated 57.8 million people were living with chronic HCV infection, and in 2022, there were 1 million newly infected individuals and 242,000 HCV-related deaths reported 1 , 2 . In the United States, it was estimated that roughly 2.2 million adults were previously or currently infected with HCV 3 . Following acute infection, approximately 70% of HCV-infected individuals progress to chronic infection 2 , and among them, about one-third of the patients will progress to cirrhosis, and the risk of liver-related mortality is 12-fold higher than the mortality of all other causes 4 , 5 . HCV is composed of six major genotypes, of which genotypes 1 and 3 accounted for approximately 46% and 36% of global infections, respectively 6 . Although the development of direct-acting anti-viral (DAA) treatment has made virologic cure achievable across genotypes, with phase 3 sustained virologic response rates exceeding 95%, HCV remains a substantial source of morbidity and mortality worldwide 7 . Accordingly, the WHO set the elimination of viral hepatitis globally by 2030 as a major international goal, with HCV being one of its primary focuses 2 . The estimated number of HCV infections worldwide declined from 63.6 million in 2015 to 56.8 million in 2020, an 11% reduction; however, the WHO elimination target cannot be achieved at this pace 8 . Progress toward HCV elimination has been uneven across regions. In metropolitan Italy, HCV prevalence decreased, and the awareness of HCV infection increased 9 . In sub-Saharan Africa, which accounts for approximately 11 million of the global infections, the elimination effort remained mild due to limited health infrastructure, low diagnostic coverage, and insufficient treatment access 8 , 10 . Even in the US, underdiagnosis remains a central problem, with about 30% to 40% of US adults with past or present infection being unaware of their HCV-infected status 11 , 12 . Thus, identifying high-risk groups in the general population can help with the case-finding process in a cost-effective way and further contribute to achieving the 2030 target. Historically, HCV detection relied on risk-based testing, which targeted individuals with recognized exposure or conditions, including people who inject drugs, men who have sex with men (MSM), recipients of blood transfusions or organ transplants, and patients undergoing hemodialysis, before the universal adult screening was recommended 13 – 15 . However, the real-world implementation of the screening recommendations has remained incomplete. Only 40.6% of pregnant individuals were screened in 2021, despite being a clearly recommended population for testing 16 . Similarly, strengthening screening among the incarcerated populations is necessary, since their HCV positivity rates are about ten times higher than the general population 17 . Thus, a simple, low-cost, and easily replicable approach for early HCV risk stratification is needed to complement the universal screening recommendation. Recent advancements in data science have expanded the use of machine learning methods such as random forest in clinical predictions due to their ability to capture the complex nonlinear relationships between variables, where classical statistical methods such as logistic regression (LR) often act as a benchmark 18 , 19 . In structured tabular datasets, which are typical for electronic health records, tree-based machine learning methods were shown to outperform deep learning in medium-sized (~ 10k) datasets under most scenarios 20 . In a previous HCV risk prediction study, the best-performing machine-learning model was trained on 238 different predictors, with a C statistic of 0.916 21 . Because the model relied on extensive amounts of variables, its direct application may be less practical. Therefore, we developed, validated, and compared the performance of five different machine learning models and also LR that were trained on only five demographic variables, using NHANES as a proof-of-concept dataset to demonstrate a framework that could be replicated with local data in other countries. METHODS Study Design and Population The National Health and Nutrition Examination Survey (NHANES) is a nationally representative, repeated cross-sectional survey of the US population that collects questionnaire data, examination measures, and laboratory results, and uses sampling weights to generate population-level estimates. The newly released 2021–2023 cycle makes it possible to extend prior HCV studies into the post-pandemic period 3 . This study used data from ten cycles (1999–2014, 2017–2023) from the NHANES database, with cycle 2015–2016 excluded because HCV-related laboratory data were not available. The protocol of NHANES was approved by the National Center for Health Statistics, and all participants in the database provided written informed consent 22 . 109,584 distinct participants were identified from 1999–2000 through 2021–2023. 31,863 participants were excluded due to unresolved HCV status (missing both HCV antibody status and RNA status), and 8,525 participants were excluded due to missing demographic data (birth year, sex, race/ethnicity, birthplace, or income-to-poverty ratio). This exclusion criterion resulted in a final modeling cohort of 69,196 participants (Supplementary Fig. 1) . Outcome Definition Outcome Definition The primary outcome of this study was HCV exposure, defined as either a positive HCV antibody test or a positive HCV RNA test. Participants were classified as HCV-negative if all available HCV-related marker(s) were negative. Participants were classified as current infection if they tested positive for RNA, regardless of the presentation of antibody or not. Participants were defined as resolved HCV infection if they were positive for antibody test but negative for RNA test. Assessment of Covariates Covariates for baseline characterization were derived from official NHANES variables and standardized across cycles. Sex was recorded as male or female, and race/ethnicity was grouped as White, Black, Hispanic, and Asian/Other. Birthplace was dichotomized to US-born versus non-US-born. The income-to-poverty ratio (IPR) was defined as the ratio of family income to the federal poverty threshold. Education was categorized as less than college versus college or above, and marital status as married/living with partner versus other from NHANES questionnaire responses. Diabetes Mellitus (DM) and Coronary Artery Disease (CAD) were defined by self-reported physician diagnosis. Current smoking was defined as having smoked at least 100 cigarettes in life and currently smoking every day or some days. Alcohol use was classified as > 1 versus ≤ 1 drink on drinking days in the past 12 months. Probable depression was defined as a Patient Health Questionnaire-9 (PHQ-9) total score of ≥ 10 23 . Model Training and Validation The study cohort was divided into a training cohort in a 7:3 ratio according to the cycle information 24 . The training cohort included NHANES participants from cycles 1999–2000, 2003–2004, 2007–2008, 2009–2010, 2011–2012, 2013–2014, and 2017-March 2020 pre-pandemic, with a sample size of 48,434 25 . The validation cohort included NHANES participants from the cycles 2001–2002, 2005–2006, and 2021–2023, with a population of 20,762. Five demographic and socioeconomic variables, including sex, birth year, race/ethnicity, birthplace, and income-to-poverty ratio, were used to train machine learning models to predict the HCV exposure status of the individuals. These variables were selected a priori because birth year, birthplace, sex, race/ethnicity, and poverty-related measures had been associated with HCV prevalence in the US in previous studies 11 , 12 , 26 , 27 . Six models were trained, validated, and compared on the training and validation cohorts using the selected features. The models included LR, weighted logistic regression (WLR), random forest (RF), extreme gradient boosting (XGB), generalized additive model (GAM), and elastic-net logistic regression (GLMNET). The LR was chosen as an interpretable benchmark. RF and XGB were decision-tree-based models, whereas GAM modeled nonlinear associations using smooth functions; WLR and GLMNET were regression-based models. Both tree-based models and regularized logistic regression methods were well-suited to structured tabular health data and can capture nonlinear associations 18 , 19 , 28 . Technical details of these models can be found in the Supplementary Table 1 . Predicted probabilities were calibrated using Platt scaling fit on the training cohort, and calibrated probabilities were used for decile-based display. Web-based Calculator To facilitate the practical use of the results of this study, the best-performing model was deployed as a web-based HCV-risk calculator, available at https://sylviesaiko.github.io/hcv-risk-calculator/ . The tool provided an easy-to-use user interface in which the HCV exposure risk was calculated from the five demographic variables entered by healthcare personnel ( Supplementary Fig. 4 ). The web calculator applies the same Platt-scaling calibration fitted in the fixed-cycle training cohort for probability display 29 – 31 . Statistical Analysis The weighted analyses adhered to NHANES analytic guidelines, which accounted for survey strata, primary sampling units, and mobile examination center (MEC) weights, and the resultant sample is representative of a large portion of the US population. Model performance metrics included sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiver operating characteristic curve (AUROC). Threshold-dependent performance metrics were calculated using a model-specific cutoff, defined as the 80th percentile of predicted risk within the corresponding dataset, such that the top 20% of participants were classified as high risk. p-values were unweighted for model-development contexts (train, validation, or modeling cohort compatibility), while survey-weighted p-values were used for the pre/post prevalence and weighted regression. The Pearson chi-square test was used for all categorical variables, and the Wilcoxon rank-sum test was used for all continuous variables in baseline comparisons. The DeLong test was used to generate p-values for AUROC comparison between different machine-learning models and LR. Model interpretability was assessed using Shapley Additive Explanations (SHAP) in the validation cohort. For each fitted model, the SHAP value was computed on a random validation subsample (up to 400 participants) using Monte Carlo estimation. The feature importance was summarized as the mean absolute SHAP value, and both global-level and participant-level SHAP distributions were visualized. As a sensitivity analysis, model training and validation were repeated for all models after redefining the outcome as current HCV infection. Participants with negative HCV antibody results or those with positive antibody but negative RNA results were classified as controls, whereas antibody-positive participants without available RNA results were excluded from the analysis. The same five predictors were used, and the participants were split into training and validation cohorts in the same way as the main analysis. All analyses were performed using R (version 4.5.2; http://www.R-project.org ). RESULTS Prevalence of HCV in the United States A total of 69,196 participants were included in the analysis, with 967 individuals showing evidence of HCV exposure (Supplementary Fig. 1) . The weighted HCV prevalence remained relatively stable across most cycles, ranging from 1.22% to 1.93%. The weighted prevalence did not differ significantly before and after the pandemic, with estimates of 1.71% (95% CI 1.08%-2.69%) in 2017–2020 and 1.38% (95% CI 0.98%-1.95%) in 2021–2023 ( Fig. 1 ) . The HCV prevalence was higher in men than in women consistently over cycles (Supplementary Fig. 2a) . The HCV prevalence was generally the highest in non-Hispanic Black participants, although non-Hispanic White prevalence also increased in the 2017–2020 pre-pandemic cycle. (Supplementary Fig. 2b) Baseline Characteristics of the Analytical Sample The overall cohort was 48.84% male, 42.23% non-Hispanic White, 22.47% non-Hispanic Black, 26.77% Hispanic, 8.53% Asian or other, and 82.01% US-born ( Table 1 ) . Compared with participants without HCV exposure, HCV-exposed participants were more likely to be male, older, black, born in the USA, have diabetes, or have coronary artery disease. Participants with HCV exposure also had higher Hemoglobin A1C, glucose levels, alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), and creatinine. They were also more socioeconomically disadvantaged, with a lower income-to-poverty ratio, a lower level of education, lower insurance coverage, and higher rates of current smoking and probable depression. LDL was slightly lower rather than higher in the HCV-exposed group, and alkaline phosphatase did not differ significantly. Table 1 Baseline characteristics by HCV exposure / No HCV exposure Variable Overall (n = 69,196) HCV exposure (n = 967) No HCV exposure (n = 68,229) p value Male sex 48.84 62.15 48.65 < 0.001 Birth year 1976 (1955–1991) 1958 (1952–1964) 1976 (1955–1991) < 0.001 1911–1930 4.57 2.28 4.60 1931–1950 16.37 18.92 16.34 1951–1970 23.82 67.53 23.18 1971–1990 30.16 10.13 30.45 1991–2010 25.08 1.14 25.43 Race/ethnicity < 0.001 Non-Hispanic White 42.23 41.26 42.24 Non-Hispanic Black 22.47 34.85 22.30 Hispanic 26.77 17.79 26.90 Asian or other 8.53 6.10 8.56 Birth place: USA-born 82.01 92.04 81.87 < 0.001 Education (n = 54,663) < 0.001 Less than college 79.64 93.24 79.40 College or above 20.36 6.76 20.60 Health insurance (n = 48,516) < 0.001 Yes 83.15 75.07 83.27 No 16.85 24.93 16.73 Marital status (n = 51,622) < 0.001 Married or living with partner 53.61 47.38 53.73 Widowed, divorced, separated, or never married 46.39 52.62 46.27 Alcohol use in past 12 months (n = 29,693) 1/day 65.12 82.50 64.77 <=1/day 34.88 17.50 35.23 Smoking (n = 35,220) < 0.001 Every day or some days 27.02 73.74 26.08 Never 72.98 26.26 73.92 Depression (n = 32,990) < 0.001 Probable case 9.35 20.65 9.12 Not a probable case 90.65 79.35 90.88 Diabetes (n = 68,046) 8.33 15.96 8.23 < 0.001 CAD (n = 46,223) 4.41 6.58 4.36 < 0.001 Income to poverty ratio (n = 69,196) 2.0 (1.0-3.9) 1.3 (0.8–2.4) 2.0 (1.0-3.9) < 0.001 BMI, kg/m2 (n = 68,202) 26.0 (21.6–30.8) 27.1 (23.5–31.1) 25.9 (21.6–30.8) < 0.001 HbA1c, % (n = 60,318) 5.4 (5.1–5.7) 5.5 (5.2–5.9) 5.4 (5.1–5.7) < 0.001 Glucose, mg/dl (n = 29,644) 97.1 (90.8-106.4) 102.5 (94.0-113.0) 97.0 (90.7–106.0) < 0.001 HDL, mg/dl (n = 55,039) 51.0 (43.0–62.0) 50.0 (41.0–62.0) 51.0 (43.0–62.0) 0.039 LDL, mg/dl (n = 29,569) 104.0 (83.0-130.0) 98.0 (80.0-126.0) 104.0 (83.0-130.0) 0.004 TG, mg/dl (n = 30,485) 96.0 (66.0-144.0) 98.5 (71.0-145.2) 96.0 (66.0-144.0) 0.136 ALT, U/L (n = 59,613) 19.0 (15.0–26.0) 33.0 (21.0–55.0) 19.0 (15.0–26.0) < 0.001 AST, U/L (n = 59,568) 22.0 (18.0–26.0) 33.0 (23.0–51.0) 22.0 (18.0–26.0) < 0.001 ALP, U/L (n = 59,700) 74.0 (59.0–95.0) 75.0 (61.0–95.0) 74.0 (59.0–95.0) 0.349 GGT, U/L (n = 59,699) 18.0 (13.0–28.0) 39.0 (22.0–86.0) 18.0 (13.0–27.0) < 0.001 TBil, mg/dl (n = 53,233) 0.6 (0.4–0.8) 0.6 (0.4–0.8) 0.6 (0.4–0.8) 0.041 Creatinine, mg/dl (n = 59,704) 0.8 (0.7-1.0) 0.9 (0.7-1.0) 0.8 (0.7-1.0) < 0.001 Platelets, x10^9/L (n = 69,106) 258.0 (217.0-305.0) 230.5 (181.0-278.0) 258.0 (218.0-305.0) < 0.001 Categorical variables are shown as percentages; continuous variables are shown as median (Q1-Q3). HCV exposure was defined as positivity for HCV antibody or HCV RNA. Variable-specific sample sizes are shown in the row labels because data were not available for all participants across all NHANES cycles. P values compare HCV-exposed and non-exposed participants. HCV, hepatitis C virus; CAD, coronary artery disease; BMI, body mass index; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglycerides; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ALP, alkaline phosphatase; GGT, gamma-glutamyl transferase; TBil, total bilirubin. Characteristics of Current vs. Previous HCV Infection Current-versus-previous infection status could be determined for 888 participants who were HCV antibody positive and had available RNA results. Among the individuals, 580 had current HCV infection, and 308 had previous HCV infection (Supplementary Table 2) . Compared with previous infection, current HCV infection was more common in men and non-Hispanic Black participants and was associated with less insurance coverage, lower income-to-poverty ratio, and higher smoking prevalence. Current infection also showed more adverse liver-related laboratory results, including higher alanine aminotransferase, aspartate aminotransferase, gamma-glutamyl transferase, and total bilirubin, together with lower platelet counts. Across cycles, the weighted proportion of current infection among antibody-positive participants declined from over 80% in the early cycles to 40.9% in 2017–2020 and 31.9% in 2021–2023. Key Predictor Identification All six models were trained on five sociodemographic variables, including race/ethnicity, biological sex, birth year, birthplace, and income-to-poverty ratio. In a multivariable LR analysis ( Table 2 ) , later birth year was associated with a lower chance of HCV exposure (adjusted OR per 10-year increase, 0.72, 95% CI 0.70–0.75, p < 0.001), whereas male sex (adjusted OR 1.91, 95% CI 1.63–2.23, p < 0.001), non-Hispanic Black race versus White (adjusted OR 1.87, 95% CI 1.56–2.24, p < 0.001), U.S. birth (adjusted OR 3.04, 95% CI 2.23–4.13, p < 0.001), and lower income-to-poverty ratio (adjusted OR per 1-unit increase, 0.69, 95% CI 0.65–0.73, p < 0.001) were independently associated with HCV exposure. Hispanic ethnicity (adjusted OR 1.08, 95% CI 0.85–1.38, p = 0.523) and Asian/Other race (adjusted OR 1.30, 95% CI 0.90–1.89, p = 0.162) were not independently associated with HCV exposure after adjustment. Table 2 Unadjusted and adjusted logistic-regression predictors for HCV exposure Term Unadjusted OR p-value Adjusted OR p value Birth year (per 10 year) 0.76 (0.73, 0.78) < 0.001 0.72 (0.70, 0.75) < 0.001 Male sex 1.76 (1.51, 2.06) < 0.001 1.91 (1.63, 2.23) < 0.001 Black vs White 1.59 (1.34, 1.89) < 0.001 1.87 (1.56, 2.24) < 0.001 Hispanic vs White 0.64 (0.51, 0.79) < 0.001 1.08 (0.85, 1.38) 0.523 Asian/Other vs White 0.54 (0.38, 0.77) < 0.001 1.30 (0.90, 1.89) 0.162 US-born vs non-US-born 2.53 (1.93, 3.33) < 0.001 3.04 (2.23, 4.13) < 0.001 Income/poverty-line ratio (per 1 unit) 0.76 (0.72, 0.80) < 0.001 0.69 (0.65, 0.73) < 0.001 Odds ratios (ORs) and 95% CIs were estimated using logistic regression. The adjusted model included birth year, sex, race/ethnicity, birthplace, and income-to-poverty ratio. Birth year was modeled per 10-year increase, and income-to-poverty ratio was modeled per 1-unit increase. White race/ethnicity and non-US-born status were used as reference categories where applicable. Model Performance The fixed-cycle split yielded 48,434 training participants and 20,762 validation participants (Supplementary Table 3) . In the training set, RF achieved the highest apparent discrimination (AUROC 0.915), followed by XGBoost (AUROC 0.893), whereas LR achieved an AUROC of 0.783 ( Table 3 , Fig. 2 ) . In the validation set, LR achieved an AUROC of 0.762, sensitivity 0.527, specificity 0.804, PPV 0.036, and NPV 0.992, whereas XGBoost showed the best overall generalization with an AUROC of 0.860, sensitivity 0.760, specificity 0.808, PPV 0.052, and NPV 0.996 ( Table 3 , Fig. 2 ) . The RF achieved the highest training AUROC (0.915) but lower validation performance (0.848), suggesting less stable generalization than XGBoost. Table 3 Model performance for HCV exposure prediction Cohort Model Sensitivity Specificity PPV NPV AUROC Training LR 0.583 0.805 0.0412 0.993 0.783 GAM 0.757 0.808 0.0535 0.996 0.868 GLMNET 0.522 0.805 0.0368 0.992 0.785 RF * 0.864 0.810 0.0610 0.998 0.915 WLR 0.577 0.805 0.0408 0.993 0.778 XGB 0.797 0.809 0.0563 0.996 0.893 Validation LR 0.527 0.804 0.0359 0.992 0.762 GAM 0.753 0.808 0.0513 0.996 0.856 GLMNET 0.519 0.804 0.0354 0.992 0.759 RF 0.731 0.807 0.0499 0.995 0.848 WLR 0.519 0.804 0.0354 0.992 0.757 XGB * 0.760 0.808 0.0517 0.996 0.860 Model performance was evaluated in the training and validation cohorts. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated using the cohort-specific 80th -percentile threshold for each model. Asterisks indicate the best-performing model in each cohort according to AUROC. In the sensitivity analysis, where the outcome was changed to HCV RNA positivity, the overall pattern of model performance was similar (Supplementary Table 4, Supplementary Fig. 3) . The fixed-cycle split resulted in 47,585 training participants and 20,688 validation participants, with 425 and 160 current HCV infections, respectively. In the training cohort, LR achieved an AUROC of 0.802, whereas RF was the best-performing model with an AUROC of 0.921. In the validation cohort, LR achieved an AUROC of 0.794, whereas XGBoost again showed the best performance with an AUROC of 0.867. In the validation cohort, decile-based risk stratification showed a clear gradient in HCV exposure prevalence ( Fig. 3 ) . For the LR model, the observed prevalence increased from 0.14% in the lowest decile to 4.14% in the highest decile, suggesting a ~ 30-fold increase, whereas for XGBoost, the observed prevalence increased from 0.05% in the lowest decile to 7.95% in the highest decile, corresponding to an approximately 160-fold increase. The top validation decile captured 30.4% of participants with HCV exposure for LR and 58.3% for XGBoost. The top three deciles together captured approximately 85.5% of cases in XGBoost, compared to 67.1% in LR. SHAP analyses across all six models consistently identified birth year (mean |SHAP value| = 0.22) as the dominant predictor, with income-to-poverty ratio (mean |SHAP value| = 0.075) ranking second in the best-performing tree-based models ( Fig. 4 a ) . DISCUSSION In this study, the prevalence of HCV exposure was stable across all cycles. There was no statistically significant difference between the 2017–2020 pre-pandemic period and the 2021–2023 cycle, suggesting that the pandemic did not have a significant impact on national HCV burden 11 . The prevalence was consistent with the previous reports of HCV prevalence of approximately 1% 3 . Notably, the proportion of current infection among affected individuals dropped steadily, falling from above 80% in previous cycles to 31.9% in 2021–2023, especially in recent years. This trend likely reflects the synergy between the expanded DAA treatment in 2014 and ongoing prevention efforts 8 , 32 . HCV exposure prevalence was consistently higher in males than in females, and non-Hispanic Black participants had the highest weighted prevalence across all cycles except for 2017–2020. The prevalence shift was probably explained by the opioid epidemic, where the new HCV infections surged among young white adults 33 . Evidence of HCV exposure was more common among males, non-Hispanic Black participants, US-born participants, and participants with greater socioeconomic disadvantage, which is consistent with previous findings and thus expected 11 , 12 , 26 , 27 . HCV-exposed individuals also had significantly higher rates of current smoking and probable depression, showing the association between HCV infection and psychiatric comorbidity, which was shown to be both a risk factor for infection and a barrier to treatment 34 . The liver-related laboratory profile was notably more adverse in HCV-exposed groups than in the general population. ALT, AST, and GGT were higher, indicating an ongoing hepatic inflammation. Creatinine was also higher in the HCV-exposed group, which may suggest the HCV-associated glomerulonephritis 35 . Interestingly, LDL was slightly lower in the HCV-exposed group, consistent with a prior result that showed how HCV disrupts lipid metabolism by utilizing lipoprotein pathways for cell entry and assembly 36 . In the cohort, current HCV infection was associated with a more adverse socioeconomic and liver-related laboratory profile than those with resolved infection, which may reflect differences in viral clearance, treatment uptake, and access to care. A central finding of this study is that a model based on five readily available demographic variables can still provide accurate HCV risk stratification. XGBoost had the best validation performance and identified high-risk groups more effectively than LR. Most previous HCV prediction studies have relied on much more detailed clinical data. Jang et al. reported stronger discrimination with an electronic health record (EHR) based model that included hundreds of candidate predictors in a state-level dataset 21 . Our goal, however, was not to maximize predictive performance at all costs, but to evaluate how far a simple demographic model could go in a nationally representative sample. In that context, NHANES offers an important advantage because it is nationally representative, and the required sociodemographic variables are widely available in both clinical and public health settings. The SHAP analyses also highlighted the importance of the birth-cohort effect in HCV epidemiology, with birth year ranking as the most influential predictor across all models, consistent with prior findings 12 . The similar performance observed in the sensitivity analysis further suggests that these demographic predictors retain useful discriminatory value even when the outcome is restricted to clinically active infection. This study has several strengths. It leveraged multiple NHANES cycles, including the newly released 2021–2023 cycle, which could be weighted to represent the US population. Additionally, all models were interpreted using SHAP, which improves the transparency and interpretability of the results. Moreover, the best-performing model was published online with open access, which offers a practical application opportunity in clinical settings. In addition, the main finding remained mostly unchanged in a sensitivity analysis where only HCV RNA-positive individuals were considered positive, indicating the robustness of the model. The study also has important limitations. NHANES is cross-sectional for HCV infection status, so incident infection and longitudinal transitions cannot be assessed. Furthermore, NHANES excluded certain populations with high HCV burden, so the true national burden is likely underestimated 3 . The models were also only internally validated due to the lack of other nationally representative datasets. Finally, the models should be interpreted as complementary tools for case-finding rather than replacements for guideline-recommended universal screening. In summary, this study demonstrates that a parsimonious set of five sociodemographic variables was sufficient to stratify HCV exposure risk in the US population. The same framework could be adopted and recalibrated using local data in other US healthcare systems or non-US populations, contributing to the global progress towards the WHO goal. Importantly, this model aligns with the current universal recommendation, and a high predicted risk score provides an additional reason for prioritizing HCV testing in certain participants, particularly in resource-limited settings where universal screening has not yet been adopted. Declarations Ethics approval and consent to participate: The study was conducted in accordance with the Declaration of Helsinki. The protocol of NHANES was approved by the National Center for Health Statistics Research Ethics Review Board, and all participants in the database provided written informed consent. Consent for publication: Not applicable. Availability of data and materials: The datasets analyzed in this study are publicly available from the NHANES repository at the CDC (https://www.cdc.gov/nchs/nhanes/index.html, accessed 01 March 2026). The analysis code and web calculator are openly available on GitHub (https://sylviesaiko.github.io/hcv-risk-calculator/). Co mpeting Interest: None. Funding: None. Author s’ Contributions: D.D. and Y.Y. conceived this study and designed the research and analytical plan. D.D. and Y.Y. wrote the first draft of the manuscript. D.D. conducted the statistical analysis and prepared all tables and figures. T.C., Y.S., and J.L. reviewed and edited the manuscript. All authors approved the final version of the manuscript. Acknowledgements: We gratefully acknowledge all those whose dedication and contributions were essential to this research project. Guarantor Statement: Ye Yuan and Dorian G Ding are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. References Cui F, Blach S, Manzengo Mingiedi C, et al. Global reporting of progress towards elimination of hepatitis B and hepatitis C. Lancet Gastroenterol Hepatol. 2023;8(4):332–42. 10.1016/s2468-1253(22)00386-7 . World Health Organization, Hepatitis C, World Health O. 2024. Accessed April 1, 2026. https://www.who.int/news-room/fact-sheets/detail/hepatitis-c Hall EW, Bradley H, Barker LK, et al. Estimating hepatitis C prevalence in the United States, 2017–2020. Hepatology. 2025;81(2):625–36. 10.1097/hep.0000000000000927 . Mahajan R, Xing J, Liu SJ, et al. Mortality among persons in care with hepatitis C virus infection: the Chronic Hepatitis Cohort Study (CHeCS), 2006–2010. Clin Infect Dis. 2014;58(8):1055–61. 10.1093/cid/ciu077 . Gordon SC, Lamerato LE, Rupp LB, et al. Prevalence of cirrhosis in hepatitis C patients in the Chronic Hepatitis Cohort Study (CHeCS): a retrospective and prospective observational Study. Am J Gastroenterol. 2015;110(8):1169–77. 10.1038/ajg.2015.203 . Messina JP, Humphreys I, Flaxman A, et al. Global distribution and prevalence of hepatitis C virus genotypes. Hepatology. 2015;61(1):77–87. 10.1002/hep.27259 . Feld JJ, Jacobson IM, Hézode C, et al. Sofosbuvir and Velpatasvir for HCV Genotype 1, 2, 4, 5, and 6 Infection. N Engl J Med. 2015;373(27):2599–607. 10.1056/NEJMoa1512610 . Polaris Observatory HCV, Collaborators. Global change in hepatitis C virus prevalence and cascade of care between 2015 and 2020: a modelling study. Lancet Gastroenterol Hepatol. 2022;7(5):396–415. 10.1016/s2468-1253(21)00472-6 . Andriulli A, Stroffolini T, Mariano A, et al. Declining prevalence and increasing awareness of HCV infection in Italy: A population-based survey in five metropolitan areas. Eur J Intern Med. 2018;53:79–84. 10.1016/j.ejim.2018.02.015 . Sonderup MW, Afihene M, Ally R, et al. Hepatitis C in sub-Saharan Africa: the current status and recommendations for achieving elimination by 2030. Lancet Gastroenterol Hepatol. 2017;2(12):910–9. 10.1016/S2468-1253(17)30249-2 . Lewis KC, Barker LK, Jiles RB, Gupta N. Estimated Prevalence and Awareness of Hepatitis C Virus Infection Among US Adults: National Health and Nutrition Examination Survey, January 2017-March 2020. Clin Infect Dis. 2023;77(10):1413–5. 10.1093/cid/ciad411 . Zou B, Yeo YH, Le MH, et al. Prevalence of Viremic Hepatitis C Virus Infection by Age, Race/Ethnicity, and Birthplace and Disease Awareness Among Viremic Persons in the United States, 1999–2016. J Infect Dis. 2020;221(3):408–18. 10.1093/infdis/jiz479 . US Preventive Services Task Force, Owens DK, Davidson KW, et al. Screening for Hepatitis C Virus Infection in Adolescents and Adults: US Preventive Services Task Force Recommendation Statement. JAMA. 2020;323(10):970–5. 10.1001/jama.2020.1123 . Schillie S, Wester C, Osborne M, Wesolowski L, Ryerson AB. CDC Recommendations for Hepatitis C Screening Among Adults - United States, 2020. MMWR Recomm Rep. 2020;69(2):1–17. 10.15585/mmwr.rr6902a1 . Published 2020 Apr 10. Jin F, Dore GJ, Matthews G, et al. Prevalence and incidence of hepatitis C virus infection in men who have sex with men: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol. 2021;6(1):39–56. 10.1016/S2468-1253(20)30303-4 . Kaufman HW, Osinubi A, Meyer WA 3rd, et al. Hepatitis C Virus Testing During Pregnancy After Universal Screening Recommendations. Obstet Gynecol. 2022;140(1):99–101. 10.1097/AOG.0000000000004822 . McNamara M, Furukawa N, Cartwright EJ, Advancing Hepatitis C. Elimination through Opt-Out Universal Screening and Treatment in Carceral Settings, United States. Emerg Infect Dis. 2024;30(13):S80–7. 10.3201/eid3013.230859 . Khosravi B, Weston AD, Nugen F, et al. Demystifying Statistics and Machine Learning in Analysis of Structured Tabular Data. J Arthroplasty. 2023;38(10):1943–7. 10.1016/j.arth.2023.08.045 . Goldstein BA, Navar AM, Pencina MJ, Ioannidis JP. Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review. J Am Med Inf Assoc. 2017;24(1):198–208. 10.1093/jamia/ocw042 . Borisov V, Leemann T, Sebler K, Haug J, Pawelczyk M, Kasneci G. Deep Neural Networks and Tabular Data: A Survey. IEEE Trans Neural Netw Learn Syst. 2024;35(6):7499–519. 10.1109/TNNLS.2022.3229161 . Jang SC, Lo-Ciganic WH, Hernandez-Con P, et al. Development and Validation of a Machine Learning-Based Screening Algorithm to Predict High-Risk Hepatitis C Infection. Open Forum Infect Dis. 2025;12(8):ofaf496. 10.1093/ofid/ofaf496 . Published 2025 Aug 15. National Center for Health Statistics, Centers for Disease Control and Prevention. NHANES Research Ethics Review Board Approval. National Health and Nutrition Examination Survey. 2024. Accessed April 1, 2026. https://wwwn.cdc.gov/nchs/nhanes/default.aspx Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606–13. 10.1046/j.1525-1497.2001.016009606.x . Lin L, Zhang L, Zhang J, Ding D. A Novel Depression Risk Prediction Model Using NHANES Data With Mendelian Randomization Validation. Brain Behav. 2025;15(7):e70674. 10.1002/brb3.70674 . Ramrakhiani NS, Chen VL, Le M, et al. Optimizing hepatitis B virus screening in the United States using a simple demographics-based model. Hepatology. 2022;75(2):430–7. 10.1002/hep.32142 . Rosenberg ES, Rosenthal EM, Hall EW, et al. Prevalence of Hepatitis C Virus Infection in US States and the District of Columbia, 2013 to 2016. JAMA Netw Open. 2018;1(8):e186371. 10.1001/jamanetworkopen.2018.6371 . Published 2018 Dec 7. Hofmeister MG, Rosenthal EM, Barker LK, et al. Estimating Prevalence of Hepatitis C Virus Infection in the United States, 2013–2016. Hepatology. 2019;69(3):1020–31. 10.1002/hep.30297 . Dentamaro V, Giglio P, Impedovo D, Pirlo G, Ciano MD. An Interpretable Adaptive Multiscale Attention Deep Neural Network for Tabular Data. IEEE Trans Neural Netw Learn Syst. 2025;36(4):6995–7009. 10.1109/TNNLS.2024.3392355 . Van Calster B, McLernon DJ, van Smeden M, Wynants L, Steyerberg EW. Topic Group ‘Evaluating diagnostic tests and prediction models’ of the STRATOS initiative. Calibration: the Achilles heel of predictive analytics. BMC Med. 2019;17(1):230. Published 2019 Dec 16. 10.1186/s12916-019-1466-7 Collins GS, Moons KGM, Dhiman P, et al. TRIPOD + AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378. 10.1136/bmj-2023-078378 . Published 2024 Apr 16. Mou W, Shan W, Yu S, et al. Interpretable machine learning model for predicting recurrence in patients with diabetic foot ulcers. BMJ Open Diabetes Res Care. 2025;13(6):e005242. 10.1136/bmjdrc-2025-005242 . Published 2025 Nov 12. Hoenigl M, Abramovitz D, Flores Ortega RE, Martin NK, Reau N. Sustained Impact of the Coronavirus Disease 2019 Pandemic on Hepatitis C Virus Treatment Initiations in the United States. Clin Infect Dis. 2022;75(1):e955–61. 10.1093/cid/ciac175 . Zibbell JE, Asher AK, Patel RC, et al. Increases in Acute Hepatitis C Virus Infection Related to a Growing Opioid Epidemic and Associated Injection Drug Use, United States, 2004 to 2014. Am J Public Health. 2018;108(2):175–81. 10.2105/AJPH.2017.304132 . Schaefer M, Capuron L, Friebe A, et al. Hepatitis C infection, antiviral treatment and mental health: a European expert consensus statement. J Hepatol. 2012;57(6):1379–90. 10.1016/j.jhep.2012.07.037 . Cacoub P, Comarmond C, Domont F, Savey L, Desbois AC, Saadoun D. Extrahepatic manifestations of chronic hepatitis C virus infection. Ther Adv Infect Dis. 2016;3(1):3–14. 10.1177/2049936115585942 . Bassendine MF, Sheridan DA, Felmlee DJ, Bridge SH, Toms GL, Neely RD. HCV and the hepatic lipid pathway as a potential treatment target. J Hepatol. 2011;55(6):1428–40. 10.1016/j.jhep.2011.06.004 . Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial20260408.docx Cite Share Download PDF Status: Posted Version 1 posted 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-9361729","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":622204222,"identity":"bcfc4526-0157-4180-9dca-4daa46c50f46","order_by":0,"name":"Dorian G Ding","email":"","orcid":"","institution":"Emory University","correspondingAuthor":false,"prefix":"","firstName":"Dorian","middleName":"G","lastName":"Ding","suffix":""},{"id":622204223,"identity":"4e306e02-a772-4ccc-8dab-aa5ea0fd32a2","order_by":1,"name":"Taoyi Chen","email":"","orcid":"","institution":"Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Taoyi","middleName":"","lastName":"Chen","suffix":""},{"id":622204224,"identity":"c55248ee-4bc9-40fb-9511-e98078e2497b","order_by":2,"name":"Yu Sheng","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Sheng","suffix":""},{"id":622204225,"identity":"5b752d6e-5af1-415e-b4c6-3bee6d3a0ebb","order_by":3,"name":"Jeffrey S.H. Lin","email":"","orcid":"","institution":"University of British Columbia","correspondingAuthor":false,"prefix":"","firstName":"Jeffrey","middleName":"S.H.","lastName":"Lin","suffix":""},{"id":622204226,"identity":"8a98176c-e2f7-4176-ba83-9233234cef69","order_by":4,"name":"Ye Yuan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIiWNgGAWjYPACGwbGBiDFQ4KWNNK1HIZQRGkxOMBjJvFzx3l75hkJjA/etjHImxOjRbL3zG1mxhkJzIZz2xgMdzYQYwtv2202oBY2ad42hgSDA8TY8rftHA9QC/tvorUADT8gAbKFmSgtkofZiq1l25INGHseNkvOOSdhuIGQFr7jzRtvvm2zszdsTz744U2ZjTxBWxQOcxiAGYYN4MiUIKAeCOQb2B9AGITVjoJRMApGwUgFAKSWOpzqWha4AAAAAElFTkSuQmCC","orcid":"","institution":"Zhejiang University","correspondingAuthor":true,"prefix":"","firstName":"Ye","middleName":"","lastName":"Yuan","suffix":""}],"badges":[],"createdAt":"2026-04-09 01:08:28","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9361729/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9361729/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106963021,"identity":"43bc1f78-ce23-4ade-9ce2-3d10dd46d90b","added_by":"auto","created_at":"2026-04-15 09:41:50","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":48237,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eWeighted HCV prevalence by NHANES cycle year.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9361729/v1/e30d8f6f71ca89d2b4892bca.png"},{"id":106964403,"identity":"8563e8db-89e9-44c7-b9a7-63c0e63d9645","added_by":"auto","created_at":"2026-04-15 09:50:09","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1159858,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eROC curve comparison across models.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9361729/v1/ddde03433d80de93c6c54c1b.png"},{"id":106962091,"identity":"ebbcaca7-5034-42c4-aa4a-dbc7da0d7a73","added_by":"auto","created_at":"2026-04-15 09:33:34","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":458407,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eObserved HCV exposure prevalence by predicted risk decile.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9361729/v1/dbc16310304d0a93045a15ff.png"},{"id":106962080,"identity":"9c2b25ab-00e5-4e5b-aa7a-64bcd6630e43","added_by":"auto","created_at":"2026-04-15 09:33:21","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":108050,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSHAP interpretation of the best-performing XGBoost model. (A) Mean absolute SHAP values. (B) Beeswarm plot.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9361729/v1/4b0483634125f8f09e44db24.png"},{"id":106994259,"identity":"116bf072-bbaf-4c0a-8a07-db59f501f657","added_by":"auto","created_at":"2026-04-15 15:06:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3252567,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9361729/v1/8549b45b-bc25-40bf-b119-2d259d80281c.pdf"},{"id":106962069,"identity":"855ec7ad-d71d-44b7-bc22-868e617c6cf1","added_by":"auto","created_at":"2026-04-15 09:32:56","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":982435,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial20260408.docx","url":"https://assets-eu.researchsquare.com/files/rs-9361729/v1/1e78a7769645170eced17da2.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and Validation of a Machine Learning Model for Hepatitis C Virus Exposure: A Demographic Screening Approach for the US Population","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eHepatitis C virus (HCV) remains a significant public health concern worldwide. In 2020, an estimated 57.8\u0026nbsp;million people were living with chronic HCV infection, and in 2022, there were 1\u0026nbsp;million newly infected individuals and 242,000 HCV-related deaths reported \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. In the United States, it was estimated that roughly 2.2\u0026nbsp;million adults were previously or currently infected with HCV \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Following acute infection, approximately 70% of HCV-infected individuals progress to chronic infection \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, and among them, about one-third of the patients will progress to cirrhosis, and the risk of liver-related mortality is 12-fold higher than the mortality of all other causes \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. HCV is composed of six major genotypes, of which genotypes 1 and 3 accounted for approximately 46% and 36% of global infections, respectively \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. Although the development of direct-acting anti-viral (DAA) treatment has made virologic cure achievable across genotypes, with phase 3 sustained virologic response rates exceeding 95%, HCV remains a substantial source of morbidity and mortality worldwide \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAccordingly, the WHO set the elimination of viral hepatitis globally by 2030 as a major international goal, with HCV being one of its primary focuses \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. The estimated number of HCV infections worldwide declined from 63.6\u0026nbsp;million in 2015 to 56.8\u0026nbsp;million in 2020, an 11% reduction; however, the WHO elimination target cannot be achieved at this pace \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Progress toward HCV elimination has been uneven across regions. In metropolitan Italy, HCV prevalence decreased, and the awareness of HCV infection increased \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. In sub-Saharan Africa, which accounts for approximately 11\u0026nbsp;million of the global infections, the elimination effort remained mild due to limited health infrastructure, low diagnostic coverage, and insufficient treatment access \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Even in the US, underdiagnosis remains a central problem, with about 30% to 40% of US adults with past or present infection being unaware of their HCV-infected status \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Thus, identifying high-risk groups in the general population can help with the case-finding process in a cost-effective way and further contribute to achieving the 2030 target.\u003c/p\u003e \u003cp\u003eHistorically, HCV detection relied on risk-based testing, which targeted individuals with recognized exposure or conditions, including people who inject drugs, men who have sex with men (MSM), recipients of blood transfusions or organ transplants, and patients undergoing hemodialysis, before the universal adult screening was recommended \u003csup\u003e\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. However, the real-world implementation of the screening recommendations has remained incomplete. Only 40.6% of pregnant individuals were screened in 2021, despite being a clearly recommended population for testing \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Similarly, strengthening screening among the incarcerated populations is necessary, since their HCV positivity rates are about ten times higher than the general population \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Thus, a simple, low-cost, and easily replicable approach for early HCV risk stratification is needed to complement the universal screening recommendation.\u003c/p\u003e \u003cp\u003eRecent advancements in data science have expanded the use of machine learning methods such as random forest in clinical predictions due to their ability to capture the complex nonlinear relationships between variables, where classical statistical methods such as logistic regression (LR) often act as a benchmark \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. In structured tabular datasets, which are typical for electronic health records, tree-based machine learning methods were shown to outperform deep learning in medium-sized (~\u0026thinsp;10k) datasets under most scenarios \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. In a previous HCV risk prediction study, the best-performing machine-learning model was trained on 238 different predictors, with a C statistic of 0.916 \u003csup\u003e21\u003c/sup\u003e. Because the model relied on extensive amounts of variables, its direct application may be less practical. Therefore, we developed, validated, and compared the performance of five different machine learning models and also LR that were trained on only five demographic variables, using NHANES as a proof-of-concept dataset to demonstrate a framework that could be replicated with local data in other countries.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Population\u003c/h2\u003e \u003cp\u003eThe National Health and Nutrition Examination Survey (NHANES) is a nationally representative, repeated cross-sectional survey of the US population that collects questionnaire data, examination measures, and laboratory results, and uses sampling weights to generate population-level estimates. The newly released 2021\u0026ndash;2023 cycle makes it possible to extend prior HCV studies into the post-pandemic period \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. This study used data from ten cycles (1999\u0026ndash;2014, 2017\u0026ndash;2023) from the NHANES database, with cycle 2015\u0026ndash;2016 excluded because HCV-related laboratory data were not available. The protocol of NHANES was approved by the National Center for Health Statistics, and all participants in the database provided written informed consent \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. 109,584 distinct participants were identified from 1999\u0026ndash;2000 through 2021\u0026ndash;2023. 31,863 participants were excluded due to unresolved HCV status (missing both HCV antibody status and RNA status), and 8,525 participants were excluded due to missing demographic data (birth year, sex, race/ethnicity, birthplace, or income-to-poverty ratio). This exclusion criterion resulted in a final modeling cohort of 69,196 participants \u003cb\u003e(Supplementary Fig.\u0026nbsp;1)\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eOutcome Definition\u003c/h3\u003e\n\u003cdiv class=\"Heading\"\u003eOutcome Definition\u003c/div\u003e \u003cp\u003eThe primary outcome of this study was HCV exposure, defined as either a positive HCV antibody test or a positive HCV RNA test. Participants were classified as HCV-negative if all available HCV-related marker(s) were negative. Participants were classified as current infection if they tested positive for RNA, regardless of the presentation of antibody or not. Participants were defined as resolved HCV infection if they were positive for antibody test but negative for RNA test.\u003c/p\u003e\n\u003ch3\u003eAssessment of Covariates\u003c/h3\u003e\n\u003cp\u003eCovariates for baseline characterization were derived from official NHANES variables and standardized across cycles. Sex was recorded as male or female, and race/ethnicity was grouped as White, Black, Hispanic, and Asian/Other. Birthplace was dichotomized to US-born versus non-US-born. The income-to-poverty ratio (IPR) was defined as the ratio of family income to the federal poverty threshold. Education was categorized as less than college versus college or above, and marital status as married/living with partner versus other from NHANES questionnaire responses. Diabetes Mellitus (DM) and Coronary Artery Disease (CAD) were defined by self-reported physician diagnosis. Current smoking was defined as having smoked at least 100 cigarettes in life and currently smoking every day or some days. Alcohol use was classified as \u0026gt;\u0026thinsp;1 versus \u0026le;\u0026thinsp;1 drink on drinking days in the past 12 months. Probable depression was defined as a Patient Health Questionnaire-9 (PHQ-9) total score of \u0026ge;\u0026thinsp;10 \u003csup\u003e23\u003c/sup\u003e.\u003c/p\u003e\n\u003ch3\u003eModel Training and Validation\u003c/h3\u003e\n\u003cp\u003eThe study cohort was divided into a training cohort in a 7:3 ratio according to the cycle information \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. The training cohort included NHANES participants from cycles 1999\u0026ndash;2000, 2003\u0026ndash;2004, 2007\u0026ndash;2008, 2009\u0026ndash;2010, 2011\u0026ndash;2012, 2013\u0026ndash;2014, and 2017-March 2020 pre-pandemic, with a sample size of 48,434 \u003csup\u003e25\u003c/sup\u003e. The validation cohort included NHANES participants from the cycles 2001\u0026ndash;2002, 2005\u0026ndash;2006, and 2021\u0026ndash;2023, with a population of 20,762. Five demographic and socioeconomic variables, including sex, birth year, race/ethnicity, birthplace, and income-to-poverty ratio, were used to train machine learning models to predict the HCV exposure status of the individuals. These variables were selected a priori because birth year, birthplace, sex, race/ethnicity, and poverty-related measures had been associated with HCV prevalence in the US in previous studies \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. Six models were trained, validated, and compared on the training and validation cohorts using the selected features. The models included LR, weighted logistic regression (WLR), random forest (RF), extreme gradient boosting (XGB), generalized additive model (GAM), and elastic-net logistic regression (GLMNET). The LR was chosen as an interpretable benchmark. RF and XGB were decision-tree-based models, whereas GAM modeled nonlinear associations using smooth functions; WLR and GLMNET were regression-based models. Both tree-based models and regularized logistic regression methods were well-suited to structured tabular health data and can capture nonlinear associations \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Technical details of these models can be found in the \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e. Predicted probabilities were calibrated using Platt scaling fit on the training cohort, and calibrated probabilities were used for decile-based display.\u003c/p\u003e\n\u003ch3\u003eWeb-based Calculator\u003c/h3\u003e\n\u003cp\u003eTo facilitate the practical use of the results of this study, the best-performing model was deployed as a web-based HCV-risk calculator, available at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://sylviesaiko.github.io/hcv-risk-calculator/\u003c/span\u003e\u003cspan address=\"https://sylviesaiko.github.io/hcv-risk-calculator/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The tool provided an easy-to-use user interface in which the HCV exposure risk was calculated from the five demographic variables entered by healthcare personnel (\u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e). The web calculator applies the same Platt-scaling calibration fitted in the fixed-cycle training cohort for probability display \u003csup\u003e\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003e The weighted analyses adhered to NHANES analytic guidelines, which accounted for survey strata, primary sampling units, and mobile examination center (MEC) weights, and the resultant sample is representative of a large portion of the US population. Model performance metrics included sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiver operating characteristic curve (AUROC). Threshold-dependent performance metrics were calculated using a model-specific cutoff, defined as the 80th percentile of predicted risk within the corresponding dataset, such that the top 20% of participants were classified as high risk. p-values were unweighted for model-development contexts (train, validation, or modeling cohort compatibility), while survey-weighted p-values were used for the pre/post prevalence and weighted regression. The Pearson chi-square test was used for all categorical variables, and the Wilcoxon rank-sum test was used for all continuous variables in baseline comparisons. The DeLong test was used to generate p-values for AUROC comparison between different machine-learning models and LR. Model interpretability was assessed using Shapley Additive Explanations (SHAP) in the validation cohort. For each fitted model, the SHAP value was computed on a random validation subsample (up to 400 participants) using Monte Carlo estimation. The feature importance was summarized as the mean absolute SHAP value, and both global-level and participant-level SHAP distributions were visualized. As a sensitivity analysis, model training and validation were repeated for all models after redefining the outcome as current HCV infection. Participants with negative HCV antibody results or those with positive antibody but negative RNA results were classified as controls, whereas antibody-positive participants without available RNA results were excluded from the analysis. The same five predictors were used, and the participants were split into training and validation cohorts in the same way as the main analysis. All analyses were performed using R (version 4.5.2; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.R-project.org\u003c/span\u003e\u003cspan address=\"http://www.R-project.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003ePrevalence of HCV in the United States\u003c/h2\u003e \u003cp\u003eA total of 69,196 participants were included in the analysis, with 967 individuals showing evidence of HCV exposure \u003cb\u003e(Supplementary Fig.\u0026nbsp;1)\u003c/b\u003e. The weighted HCV prevalence remained relatively stable across most cycles, ranging from 1.22% to 1.93%. The weighted prevalence did not differ significantly before and after the pandemic, with estimates of 1.71% (95% CI 1.08%-2.69%) in 2017\u0026ndash;2020 and 1.38% (95% CI 0.98%-1.95%) in 2021\u0026ndash;2023 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. The HCV prevalence was higher in men than in women consistently over cycles \u003cb\u003e(Supplementary Fig.\u0026nbsp;2a)\u003c/b\u003e. The HCV prevalence was generally the highest in non-Hispanic Black participants, although non-Hispanic White prevalence also increased in the 2017\u0026ndash;2020 pre-pandemic cycle. \u003cb\u003e(Supplementary Fig.\u0026nbsp;2b)\u003c/b\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eBaseline Characteristics of the Analytical Sample\u003c/h2\u003e \u003cp\u003eThe overall cohort was 48.84% male, 42.23% non-Hispanic White, 22.47% non-Hispanic Black, 26.77% Hispanic, 8.53% Asian or other, and 82.01% US-born \u003cb\u003e(\u003c/b\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Compared with participants without HCV exposure, HCV-exposed participants were more likely to be male, older, black, born in the USA, have diabetes, or have coronary artery disease. Participants with HCV exposure also had higher Hemoglobin A1C, glucose levels, alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), and creatinine. They were also more socioeconomically disadvantaged, with a lower income-to-poverty ratio, a lower level of education, lower insurance coverage, and higher rates of current smoking and probable depression. LDL was slightly lower rather than higher in the HCV-exposed group, and alkaline phosphatase did not differ significantly.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics by HCV exposure / No HCV exposure\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;69,196)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHCV exposure\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;967)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo HCV exposure\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;68,229)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMale sex\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e48.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e62.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBirth year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1976 (1955\u0026ndash;1991)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1958 (1952\u0026ndash;1964)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1976 (1955\u0026ndash;1991)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e1911\u0026ndash;1930\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e1931\u0026ndash;1950\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e18.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e1951\u0026ndash;1970\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e67.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e23.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e1971\u0026ndash;1990\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e1991\u0026ndash;2010\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e25.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRace/ethnicity\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNon-Hispanic White\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e42.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e41.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e42.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNon-Hispanic Black\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e22.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHispanic\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26.90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAsian or other\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBirth place: USA-born\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e82.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e92.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEducation (n\u0026thinsp;=\u0026thinsp;54,663)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLess than college\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e93.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e79.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCollege or above\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHealth insurance\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e(n\u0026thinsp;=\u0026thinsp;48,516)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eYes\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e83.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e75.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e83.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNo\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e24.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMarital status (n\u0026thinsp;=\u0026thinsp;51,622)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMarried or living with partner\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e53.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e53.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eWidowed, divorced, separated, or never married\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e46.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAlcohol use in past 12 months (n\u0026thinsp;=\u0026thinsp;29,693)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e\u0026gt;1/day\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e65.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e82.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e64.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt;=1/day\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e34.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e35.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSmoking (n\u0026thinsp;=\u0026thinsp;35,220)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEvery day or some days\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e73.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNever\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e72.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDepression (n\u0026thinsp;=\u0026thinsp;32,990)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eProbable case\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNot a probable case\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e90.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e79.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDiabetes (n\u0026thinsp;=\u0026thinsp;68,046)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCAD (n\u0026thinsp;=\u0026thinsp;46,223)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eIncome to poverty ratio\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e(n\u0026thinsp;=\u0026thinsp;69,196)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.0\u003c/p\u003e \u003cp\u003e(1.0-3.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.3\u003c/p\u003e \u003cp\u003e(0.8\u0026ndash;2.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.0\u003c/p\u003e \u003cp\u003e(1.0-3.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBMI, kg/m2 (n\u0026thinsp;=\u0026thinsp;68,202)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.0\u003c/p\u003e \u003cp\u003e(21.6\u0026ndash;30.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27.1\u003c/p\u003e \u003cp\u003e(23.5\u0026ndash;31.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e25.9\u003c/p\u003e \u003cp\u003e(21.6\u0026ndash;30.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHbA1c, % (n\u0026thinsp;=\u0026thinsp;60,318)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.4\u003c/p\u003e \u003cp\u003e(5.1\u0026ndash;5.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.5\u003c/p\u003e \u003cp\u003e(5.2\u0026ndash;5.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.4\u003c/p\u003e \u003cp\u003e(5.1\u0026ndash;5.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGlucose, mg/dl (n\u0026thinsp;=\u0026thinsp;29,644)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97.1\u003c/p\u003e \u003cp\u003e(90.8-106.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e102.5\u003c/p\u003e \u003cp\u003e(94.0-113.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e97.0\u003c/p\u003e \u003cp\u003e(90.7\u0026ndash;106.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHDL, mg/dl (n\u0026thinsp;=\u0026thinsp;55,039)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51.0\u003c/p\u003e \u003cp\u003e(43.0\u0026ndash;62.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.0\u003c/p\u003e \u003cp\u003e(41.0\u0026ndash;62.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e51.0\u003c/p\u003e \u003cp\u003e(43.0\u0026ndash;62.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.039\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLDL, mg/dl (n\u0026thinsp;=\u0026thinsp;29,569)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e104.0\u003c/p\u003e \u003cp\u003e(83.0-130.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e98.0\u003c/p\u003e \u003cp\u003e(80.0-126.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e104.0\u003c/p\u003e \u003cp\u003e(83.0-130.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTG, mg/dl (n\u0026thinsp;=\u0026thinsp;30,485)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e96.0\u003c/p\u003e \u003cp\u003e(66.0-144.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e98.5\u003c/p\u003e \u003cp\u003e(71.0-145.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e96.0\u003c/p\u003e \u003cp\u003e(66.0-144.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.136\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eALT, U/L (n\u0026thinsp;=\u0026thinsp;59,613)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19.0\u003c/p\u003e \u003cp\u003e(15.0\u0026ndash;26.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33.0\u003c/p\u003e \u003cp\u003e(21.0\u0026ndash;55.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19.0\u003c/p\u003e \u003cp\u003e(15.0\u0026ndash;26.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAST, U/L (n\u0026thinsp;=\u0026thinsp;59,568)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22.0\u003c/p\u003e \u003cp\u003e(18.0\u0026ndash;26.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33.0\u003c/p\u003e \u003cp\u003e(23.0\u0026ndash;51.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e22.0\u003c/p\u003e \u003cp\u003e(18.0\u0026ndash;26.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eALP, U/L (n\u0026thinsp;=\u0026thinsp;59,700)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e74.0\u003c/p\u003e \u003cp\u003e(59.0\u0026ndash;95.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e75.0\u003c/p\u003e \u003cp\u003e(61.0\u0026ndash;95.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e74.0\u003c/p\u003e \u003cp\u003e(59.0\u0026ndash;95.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.349\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGGT, U/L (n\u0026thinsp;=\u0026thinsp;59,699)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18.0\u003c/p\u003e \u003cp\u003e(13.0\u0026ndash;28.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39.0\u003c/p\u003e \u003cp\u003e(22.0\u0026ndash;86.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18.0\u003c/p\u003e \u003cp\u003e(13.0\u0026ndash;27.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTBil, mg/dl (n\u0026thinsp;=\u0026thinsp;53,233)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003cp\u003e(0.4\u0026ndash;0.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003cp\u003e(0.4\u0026ndash;0.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003cp\u003e(0.4\u0026ndash;0.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.041\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCreatinine, mg/dl\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e(n\u0026thinsp;=\u0026thinsp;59,704)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.8\u003c/p\u003e \u003cp\u003e(0.7-1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.9\u003c/p\u003e \u003cp\u003e(0.7-1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.8\u003c/p\u003e \u003cp\u003e(0.7-1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePlatelets, x10^9/L\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e(n\u0026thinsp;=\u0026thinsp;69,106)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e258.0\u003c/p\u003e \u003cp\u003e(217.0-305.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e230.5\u003c/p\u003e \u003cp\u003e(181.0-278.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e258.0\u003c/p\u003e \u003cp\u003e(218.0-305.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eCategorical variables are shown as percentages; continuous variables are shown as median (Q1-Q3). HCV exposure was defined as positivity for HCV antibody or HCV RNA. Variable-specific sample sizes are shown in the row labels because data were not available for all participants across all NHANES cycles. P values compare HCV-exposed and non-exposed participants. HCV, hepatitis C virus; CAD, coronary artery disease; BMI, body mass index; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglycerides; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ALP, alkaline phosphatase; GGT, gamma-glutamyl transferase; TBil, total bilirubin.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eCharacteristics of Current vs. Previous HCV Infection\u003c/h2\u003e \u003cp\u003eCurrent-versus-previous infection status could be determined for 888 participants who were HCV antibody positive and had available RNA results. Among the individuals, 580 had current HCV infection, and 308 had previous HCV infection \u003cb\u003e(Supplementary Table\u0026nbsp;2)\u003c/b\u003e. Compared with previous infection, current HCV infection was more common in men and non-Hispanic Black participants and was associated with less insurance coverage, lower income-to-poverty ratio, and higher smoking prevalence. Current infection also showed more adverse liver-related laboratory results, including higher alanine aminotransferase, aspartate aminotransferase, gamma-glutamyl transferase, and total bilirubin, together with lower platelet counts. Across cycles, the weighted proportion of current infection among antibody-positive participants declined from over 80% in the early cycles to 40.9% in 2017\u0026ndash;2020 and 31.9% in 2021\u0026ndash;2023.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eKey Predictor Identification\u003c/h2\u003e \u003cp\u003eAll six models were trained on five sociodemographic variables, including race/ethnicity, biological sex, birth year, birthplace, and income-to-poverty ratio. In a multivariable LR analysis \u003cb\u003e(\u003c/b\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e, later birth year was associated with a lower chance of HCV exposure (adjusted OR per 10-year increase, 0.72, 95% CI 0.70\u0026ndash;0.75, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), whereas male sex (adjusted OR 1.91, 95% CI 1.63\u0026ndash;2.23, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), non-Hispanic Black race versus White (adjusted OR 1.87, 95% CI 1.56\u0026ndash;2.24, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), U.S. birth (adjusted OR 3.04, 95% CI 2.23\u0026ndash;4.13, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and lower income-to-poverty ratio (adjusted OR per 1-unit increase, 0.69, 95% CI 0.65\u0026ndash;0.73, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) were independently associated with HCV exposure. Hispanic ethnicity (adjusted OR 1.08, 95% CI 0.85\u0026ndash;1.38, p\u0026thinsp;=\u0026thinsp;0.523) and Asian/Other race (adjusted OR 1.30, 95% CI 0.90\u0026ndash;1.89, p\u0026thinsp;=\u0026thinsp;0.162) were not independently associated with HCV exposure after adjustment.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnadjusted and adjusted logistic-regression predictors for HCV exposure\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTerm\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnadjusted OR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAdjusted OR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBirth year\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e(per 10 year)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.76 (0.73, 0.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.72 (0.70, 0.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMale sex\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.76 (1.51, 2.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.91 (1.63, 2.23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBlack vs White\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.59 (1.34, 1.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.87 (1.56, 2.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHispanic vs White\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.64 (0.51, 0.79)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.08 (0.85, 1.38)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.523\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAsian/Other vs White\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.54 (0.38, 0.77)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.30 (0.90, 1.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.162\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eUS-born vs non-US-born\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.53 (1.93, 3.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.04 (2.23, 4.13)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eIncome/poverty-line ratio\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e(per 1 unit)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.76 (0.72, 0.80)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.69 (0.65, 0.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eOdds ratios (ORs) and 95% CIs were estimated using logistic regression. The adjusted model included birth year, sex, race/ethnicity, birthplace, and income-to-poverty ratio. Birth year was modeled per 10-year increase, and income-to-poverty ratio was modeled per 1-unit increase. White race/ethnicity and non-US-born status were used as reference categories where applicable.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eModel Performance\u003c/h2\u003e \u003cp\u003eThe fixed-cycle split yielded 48,434 training participants and 20,762 validation participants \u003cb\u003e(Supplementary Table\u0026nbsp;3)\u003c/b\u003e. In the training set, RF achieved the highest apparent discrimination (AUROC 0.915), followed by XGBoost (AUROC 0.893), whereas LR achieved an AUROC of 0.783 \u003cb\u003e(\u003c/b\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. In the validation set, LR achieved an AUROC of 0.762, sensitivity 0.527, specificity 0.804, PPV 0.036, and NPV 0.992, whereas XGBoost showed the best overall generalization with an AUROC of 0.860, sensitivity 0.760, specificity 0.808, PPV 0.052, and NPV 0.996 \u003cb\u003e(\u003c/b\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. The RF achieved the highest training AUROC (0.915) but lower validation performance (0.848), suggesting less stable generalization than XGBoost.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel performance for HCV exposure prediction\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAUROC\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTraining\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.583\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.805\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0412\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.993\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.783\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGAM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0535\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.996\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.868\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGLMNET\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.522\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.805\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.992\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRF\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.864\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0610\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.998\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.915\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.577\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.805\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0408\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.993\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.778\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eXGB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.797\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.809\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0563\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.996\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.893\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eValidation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.527\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.804\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0359\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.992\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.762\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGAM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.753\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0513\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.996\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.856\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGLMNET\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.519\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.804\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0354\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.992\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.759\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.731\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.807\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0499\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.995\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.848\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWLR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.519\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.804\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0354\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.992\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.757\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eXGB\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.760\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.0517\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.996\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.860\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eModel performance was evaluated in the training and validation cohorts. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated using the cohort-specific 80th -percentile threshold for each model. Asterisks indicate the best-performing model in each cohort according to AUROC.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn the sensitivity analysis, where the outcome was changed to HCV RNA positivity, the overall pattern of model performance was similar \u003cb\u003e(Supplementary Table\u0026nbsp;4, Supplementary Fig.\u0026nbsp;3)\u003c/b\u003e. The fixed-cycle split resulted in 47,585 training participants and 20,688 validation participants, with 425 and 160 current HCV infections, respectively. In the training cohort, LR achieved an AUROC of 0.802, whereas RF was the best-performing model with an AUROC of 0.921. In the validation cohort, LR achieved an AUROC of 0.794, whereas XGBoost again showed the best performance with an AUROC of 0.867.\u003c/p\u003e \u003cp\u003eIn the validation cohort, decile-based risk stratification showed a clear gradient in HCV exposure prevalence \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. For the LR model, the observed prevalence increased from 0.14% in the lowest decile to 4.14% in the highest decile, suggesting a\u0026thinsp;~\u0026thinsp;30-fold increase, whereas for XGBoost, the observed prevalence increased from 0.05% in the lowest decile to 7.95% in the highest decile, corresponding to an approximately 160-fold increase. The top validation decile captured 30.4% of participants with HCV exposure for LR and 58.3% for XGBoost. The top three deciles together captured approximately 85.5% of cases in XGBoost, compared to 67.1% in LR. SHAP analyses across all six models consistently identified birth year (mean |SHAP value| = 0.22) as the dominant predictor, with income-to-poverty ratio (mean |SHAP value| = 0.075) ranking second in the best-performing tree-based models \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn this study, the prevalence of HCV exposure was stable across all cycles. There was no statistically significant difference between the 2017\u0026ndash;2020 pre-pandemic period and the 2021\u0026ndash;2023 cycle, suggesting that the pandemic did not have a significant impact on national HCV burden \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. The prevalence was consistent with the previous reports of HCV prevalence of approximately 1% \u003csup\u003e3\u003c/sup\u003e. Notably, the proportion of current infection among affected individuals dropped steadily, falling from above 80% in previous cycles to 31.9% in 2021\u0026ndash;2023, especially in recent years. This trend likely reflects the synergy between the expanded DAA treatment in 2014 and ongoing prevention efforts \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. HCV exposure prevalence was consistently higher in males than in females, and non-Hispanic Black participants had the highest weighted prevalence across all cycles except for 2017\u0026ndash;2020. The prevalence shift was probably explained by the opioid epidemic, where the new HCV infections surged among young white adults \u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eEvidence of HCV exposure was more common among males, non-Hispanic Black participants, US-born participants, and participants with greater socioeconomic disadvantage, which is consistent with previous findings and thus expected \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. HCV-exposed individuals also had significantly higher rates of current smoking and probable depression, showing the association between HCV infection and psychiatric comorbidity, which was shown to be both a risk factor for infection and a barrier to treatment \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. The liver-related laboratory profile was notably more adverse in HCV-exposed groups than in the general population. ALT, AST, and GGT were higher, indicating an ongoing hepatic inflammation. Creatinine was also higher in the HCV-exposed group, which may suggest the HCV-associated glomerulonephritis \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Interestingly, LDL was slightly lower in the HCV-exposed group, consistent with a prior result that showed how HCV disrupts lipid metabolism by utilizing lipoprotein pathways for cell entry and assembly \u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. In the cohort, current HCV infection was associated with a more adverse socioeconomic and liver-related laboratory profile than those with resolved infection, which may reflect differences in viral clearance, treatment uptake, and access to care.\u003c/p\u003e \u003cp\u003eA central finding of this study is that a model based on five readily available demographic variables can still provide accurate HCV risk stratification. XGBoost had the best validation performance and identified high-risk groups more effectively than LR. Most previous HCV prediction studies have relied on much more detailed clinical data. Jang et al. reported stronger discrimination with an electronic health record (EHR) based model that included hundreds of candidate predictors in a state-level dataset \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Our goal, however, was not to maximize predictive performance at all costs, but to evaluate how far a simple demographic model could go in a nationally representative sample. In that context, NHANES offers an important advantage because it is nationally representative, and the required sociodemographic variables are widely available in both clinical and public health settings. The SHAP analyses also highlighted the importance of the birth-cohort effect in HCV epidemiology, with birth year ranking as the most influential predictor across all models, consistent with prior findings \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. The similar performance observed in the sensitivity analysis further suggests that these demographic predictors retain useful discriminatory value even when the outcome is restricted to clinically active infection.\u003c/p\u003e \u003cp\u003eThis study has several strengths. It leveraged multiple NHANES cycles, including the newly released 2021\u0026ndash;2023 cycle, which could be weighted to represent the US population. Additionally, all models were interpreted using SHAP, which improves the transparency and interpretability of the results. Moreover, the best-performing model was published online with open access, which offers a practical application opportunity in clinical settings. In addition, the main finding remained mostly unchanged in a sensitivity analysis where only HCV RNA-positive individuals were considered positive, indicating the robustness of the model. The study also has important limitations. NHANES is cross-sectional for HCV infection status, so incident infection and longitudinal transitions cannot be assessed. Furthermore, NHANES excluded certain populations with high HCV burden, so the true national burden is likely underestimated \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. The models were also only internally validated due to the lack of other nationally representative datasets. Finally, the models should be interpreted as complementary tools for case-finding rather than replacements for guideline-recommended universal screening.\u003c/p\u003e \u003cp\u003eIn summary, this study demonstrates that a parsimonious set of five sociodemographic variables was sufficient to stratify HCV exposure risk in the US population. The same framework could be adopted and recalibrated using local data in other US healthcare systems or non-US populations, contributing to the global progress towards the WHO goal. Importantly, this model aligns with the current universal recommendation, and a high predicted risk score provides an additional reason for prioritizing HCV testing in certain participants, particularly in resource-limited settings where universal screening has not yet been adopted.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u0026nbsp;\u003c/strong\u003eThe study was conducted in accordance with the Declaration of Helsinki. The protocol of NHANES was approved by the National Center for Health Statistics Research Ethics Review Board, and all participants in the database provided written informed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u0026nbsp;\u003c/strong\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials:\u0026nbsp;\u003c/strong\u003eThe datasets analyzed in this study are publicly available from the NHANES repository at the CDC (https://www.cdc.gov/nchs/nhanes/index.html, accessed 01 March 2026). The analysis code and web calculator are openly available on GitHub (https://sylviesaiko.github.io/hcv-risk-calculator/).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCo\u003c/strong\u003e\u003cstrong\u003empeting\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;Interest:\u0026nbsp;\u003c/strong\u003eNone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e None.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor\u003c/strong\u003e\u003cstrong\u003es\u0026rsquo;\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;Contributions:\u003c/strong\u003e D.D. and Y.Y. conceived this study and designed the research and analytical plan. D.D. and Y.Y. wrote the first draft of the manuscript. D.D. conducted the statistical analysis and prepared all tables and figures. T.C., Y.S., and J.L. reviewed and edited the manuscript. All authors approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e We gratefully acknowledge all those whose dedication and contributions were essential to this research project.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGuarantor Statement:\u0026nbsp;\u003c/strong\u003eYe Yuan and Dorian G Ding are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eCui F, Blach S, Manzengo Mingiedi C, et al. Global reporting of progress towards elimination of hepatitis B and hepatitis C. Lancet Gastroenterol Hepatol. 2023;8(4):332\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s2468-1253(22)00386-7\u003c/span\u003e\u003cspan address=\"10.1016/s2468-1253(22)00386-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Health Organization, Hepatitis C, World Health O. 2024. Accessed April 1, 2026. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/news-room/fact-sheets/detail/hepatitis-c\u003c/span\u003e\u003cspan address=\"https://www.who.int/news-room/fact-sheets/detail/hepatitis-c\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHall EW, Bradley H, Barker LK, et al. Estimating hepatitis C prevalence in the United States, 2017\u0026ndash;2020. Hepatology. 2025;81(2):625\u0026ndash;36. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/hep.0000000000000927\u003c/span\u003e\u003cspan address=\"10.1097/hep.0000000000000927\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMahajan R, Xing J, Liu SJ, et al. Mortality among persons in care with hepatitis C virus infection: the Chronic Hepatitis Cohort Study (CHeCS), 2006\u0026ndash;2010. Clin Infect Dis. 2014;58(8):1055\u0026ndash;61. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/cid/ciu077\u003c/span\u003e\u003cspan address=\"10.1093/cid/ciu077\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGordon SC, Lamerato LE, Rupp LB, et al. Prevalence of cirrhosis in hepatitis C patients in the Chronic Hepatitis Cohort Study (CHeCS): a retrospective and prospective observational Study. Am J Gastroenterol. 2015;110(8):1169\u0026ndash;77. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/ajg.2015.203\u003c/span\u003e\u003cspan address=\"10.1038/ajg.2015.203\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMessina JP, Humphreys I, Flaxman A, et al. Global distribution and prevalence of hepatitis C virus genotypes. Hepatology. 2015;61(1):77\u0026ndash;87. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hep.27259\u003c/span\u003e\u003cspan address=\"10.1002/hep.27259\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeld JJ, Jacobson IM, H\u0026eacute;zode C, et al. Sofosbuvir and Velpatasvir for HCV Genotype 1, 2, 4, 5, and 6 Infection. N Engl J Med. 2015;373(27):2599\u0026ndash;607. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa1512610\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa1512610\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePolaris Observatory HCV, Collaborators. Global change in hepatitis C virus prevalence and cascade of care between 2015 and 2020: a modelling study. Lancet Gastroenterol Hepatol. 2022;7(5):396\u0026ndash;415. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s2468-1253(21)00472-6\u003c/span\u003e\u003cspan address=\"10.1016/s2468-1253(21)00472-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndriulli A, Stroffolini T, Mariano A, et al. Declining prevalence and increasing awareness of HCV infection in Italy: A population-based survey in five metropolitan areas. Eur J Intern Med. 2018;53:79\u0026ndash;84. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ejim.2018.02.015\u003c/span\u003e\u003cspan address=\"10.1016/j.ejim.2018.02.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSonderup MW, Afihene M, Ally R, et al. Hepatitis C in sub-Saharan Africa: the current status and recommendations for achieving elimination by 2030. Lancet Gastroenterol Hepatol. 2017;2(12):910\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S2468-1253(17)30249-2\u003c/span\u003e\u003cspan address=\"10.1016/S2468-1253(17)30249-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLewis KC, Barker LK, Jiles RB, Gupta N. Estimated Prevalence and Awareness of Hepatitis C Virus Infection Among US Adults: National Health and Nutrition Examination Survey, January 2017-March 2020. Clin Infect Dis. 2023;77(10):1413\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/cid/ciad411\u003c/span\u003e\u003cspan address=\"10.1093/cid/ciad411\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZou B, Yeo YH, Le MH, et al. Prevalence of Viremic Hepatitis C Virus Infection by Age, Race/Ethnicity, and Birthplace and Disease Awareness Among Viremic Persons in the United States, 1999\u0026ndash;2016. J Infect Dis. 2020;221(3):408\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/infdis/jiz479\u003c/span\u003e\u003cspan address=\"10.1093/infdis/jiz479\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUS Preventive Services Task Force, Owens DK, Davidson KW, et al. Screening for Hepatitis C Virus Infection in Adolescents and Adults: US Preventive Services Task Force Recommendation Statement. JAMA. 2020;323(10):970\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jama.2020.1123\u003c/span\u003e\u003cspan address=\"10.1001/jama.2020.1123\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchillie S, Wester C, Osborne M, Wesolowski L, Ryerson AB. CDC Recommendations for Hepatitis C Screening Among Adults - United States, 2020. MMWR Recomm Rep. 2020;69(2):1\u0026ndash;17. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.15585/mmwr.rr6902a1\u003c/span\u003e\u003cspan address=\"10.15585/mmwr.rr6902a1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2020 Apr 10.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin F, Dore GJ, Matthews G, et al. Prevalence and incidence of hepatitis C virus infection in men who have sex with men: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol. 2021;6(1):39\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S2468-1253(20)30303-4\u003c/span\u003e\u003cspan address=\"10.1016/S2468-1253(20)30303-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaufman HW, Osinubi A, Meyer WA 3rd, et al. Hepatitis C Virus Testing During Pregnancy After Universal Screening Recommendations. Obstet Gynecol. 2022;140(1):99\u0026ndash;101. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/AOG.0000000000004822\u003c/span\u003e\u003cspan address=\"10.1097/AOG.0000000000004822\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcNamara M, Furukawa N, Cartwright EJ, Advancing Hepatitis C. Elimination through Opt-Out Universal Screening and Treatment in Carceral Settings, United States. Emerg Infect Dis. 2024;30(13):S80\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3201/eid3013.230859\u003c/span\u003e\u003cspan address=\"10.3201/eid3013.230859\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhosravi B, Weston AD, Nugen F, et al. Demystifying Statistics and Machine Learning in Analysis of Structured Tabular Data. J Arthroplasty. 2023;38(10):1943\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.arth.2023.08.045\u003c/span\u003e\u003cspan address=\"10.1016/j.arth.2023.08.045\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoldstein BA, Navar AM, Pencina MJ, Ioannidis JP. Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review. J Am Med Inf Assoc. 2017;24(1):198\u0026ndash;208. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/jamia/ocw042\u003c/span\u003e\u003cspan address=\"10.1093/jamia/ocw042\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBorisov V, Leemann T, Sebler K, Haug J, Pawelczyk M, Kasneci G. Deep Neural Networks and Tabular Data: A Survey. IEEE Trans Neural Netw Learn Syst. 2024;35(6):7499\u0026ndash;519. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/TNNLS.2022.3229161\u003c/span\u003e\u003cspan address=\"10.1109/TNNLS.2022.3229161\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJang SC, Lo-Ciganic WH, Hernandez-Con P, et al. Development and Validation of a Machine Learning-Based Screening Algorithm to Predict High-Risk Hepatitis C Infection. Open Forum Infect Dis. 2025;12(8):ofaf496. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/ofid/ofaf496\u003c/span\u003e\u003cspan address=\"10.1093/ofid/ofaf496\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2025 Aug 15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Center for Health Statistics, Centers for Disease Control and Prevention. NHANES Research Ethics Review Board Approval. National Health and Nutrition Examination Survey. 2024. Accessed April 1, 2026. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://wwwn.cdc.gov/nchs/nhanes/default.aspx\u003c/span\u003e\u003cspan address=\"https://wwwn.cdc.gov/nchs/nhanes/default.aspx\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1046/j.1525-1497.2001.016009606.x\u003c/span\u003e\u003cspan address=\"10.1046/j.1525-1497.2001.016009606.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin L, Zhang L, Zhang J, Ding D. A Novel Depression Risk Prediction Model Using NHANES Data With Mendelian Randomization Validation. Brain Behav. 2025;15(7):e70674. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/brb3.70674\u003c/span\u003e\u003cspan address=\"10.1002/brb3.70674\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamrakhiani NS, Chen VL, Le M, et al. Optimizing hepatitis B virus screening in the United States using a simple demographics-based model. Hepatology. 2022;75(2):430\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hep.32142\u003c/span\u003e\u003cspan address=\"10.1002/hep.32142\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosenberg ES, Rosenthal EM, Hall EW, et al. Prevalence of Hepatitis C Virus Infection in US States and the District of Columbia, 2013 to 2016. JAMA Netw Open. 2018;1(8):e186371. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2018.6371\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2018.6371\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2018 Dec 7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHofmeister MG, Rosenthal EM, Barker LK, et al. Estimating Prevalence of Hepatitis C Virus Infection in the United States, 2013\u0026ndash;2016. Hepatology. 2019;69(3):1020\u0026ndash;31. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hep.30297\u003c/span\u003e\u003cspan address=\"10.1002/hep.30297\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDentamaro V, Giglio P, Impedovo D, Pirlo G, Ciano MD. An Interpretable Adaptive Multiscale Attention Deep Neural Network for Tabular Data. IEEE Trans Neural Netw Learn Syst. 2025;36(4):6995\u0026ndash;7009. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/TNNLS.2024.3392355\u003c/span\u003e\u003cspan address=\"10.1109/TNNLS.2024.3392355\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Calster B, McLernon DJ, van Smeden M, Wynants L, Steyerberg EW. Topic Group \u0026lsquo;Evaluating diagnostic tests and prediction models\u0026rsquo; of the STRATOS initiative. Calibration: the Achilles heel of predictive analytics. \u003cem\u003eBMC Med.\u003c/em\u003e 2019;17(1):230. Published 2019 Dec 16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12916-019-1466-7\u003c/span\u003e\u003cspan address=\"10.1186/s12916-019-1466-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollins GS, Moons KGM, Dhiman P, et al. TRIPOD\u0026thinsp;+\u0026thinsp;AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmj-2023-078378\u003c/span\u003e\u003cspan address=\"10.1136/bmj-2023-078378\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2024 Apr 16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMou W, Shan W, Yu S, et al. Interpretable machine learning model for predicting recurrence in patients with diabetic foot ulcers. BMJ Open Diabetes Res Care. 2025;13(6):e005242. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmjdrc-2025-005242\u003c/span\u003e\u003cspan address=\"10.1136/bmjdrc-2025-005242\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2025 Nov 12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoenigl M, Abramovitz D, Flores Ortega RE, Martin NK, Reau N. Sustained Impact of the Coronavirus Disease 2019 Pandemic on Hepatitis C Virus Treatment Initiations in the United States. Clin Infect Dis. 2022;75(1):e955\u0026ndash;61. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/cid/ciac175\u003c/span\u003e\u003cspan address=\"10.1093/cid/ciac175\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZibbell JE, Asher AK, Patel RC, et al. Increases in Acute Hepatitis C Virus Infection Related to a Growing Opioid Epidemic and Associated Injection Drug Use, United States, 2004 to 2014. Am J Public Health. 2018;108(2):175\u0026ndash;81. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2105/AJPH.2017.304132\u003c/span\u003e\u003cspan address=\"10.2105/AJPH.2017.304132\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchaefer M, Capuron L, Friebe A, et al. Hepatitis C infection, antiviral treatment and mental health: a European expert consensus statement. J Hepatol. 2012;57(6):1379\u0026ndash;90. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jhep.2012.07.037\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2012.07.037\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCacoub P, Comarmond C, Domont F, Savey L, Desbois AC, Saadoun D. Extrahepatic manifestations of chronic hepatitis C virus infection. Ther Adv Infect Dis. 2016;3(1):3\u0026ndash;14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/2049936115585942\u003c/span\u003e\u003cspan address=\"10.1177/2049936115585942\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBassendine MF, Sheridan DA, Felmlee DJ, Bridge SH, Toms GL, Neely RD. HCV and the hepatic lipid pathway as a potential treatment target. J Hepatol. 2011;55(6):1428\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jhep.2011.06.004\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2011.06.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Hepatitis C, Prevalence, Demographics, Screening, Machine Learning","lastPublishedDoi":"10.21203/rs.3.rs-9361729/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9361729/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHepatitis C virus (HCV) remains underdiagnosed in the United States despite recommendations for universal screening. A simple approach based on readily available demographic information may help target screening in settings where screening implementation continues to be incomplete.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe analyzed 10 NHANES cycles (1999–2014 and 2017–2023) and defined HCV exposure as a positive HCV antibody or RNA result. Using sex, birth year, race/ethnicity, birthplace, and income-to-poverty ratio, we trained and compared logistic regression (LR) and machine learning models in training and validation cohorts (48,434 and 20,762 participants, respectively). Model performance was evaluated based on sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiver operating characteristic curve (AUROC). A web-based calculator was developed to facilitate bedside HCV screening.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e69,196 participants were included, with 967 showing evidence of HCV exposure. Weighted HCV prevalence remained relatively stable across cycles, ranging from 1.22% to 1.93%. The prevalence did not change significantly after the pandemic. Earlier birth year, male sex, non-Hispanic Black race, US birth, and lower income-to-poverty ratio were independently associated with HCV exposure. XGBoost performed better than LR in the validation cohort (AUROC 0.860 vs 0.762, p \u0026lt; 0.001). Predicted risk separated the population clearly: observed HCV prevalence increased from 0.05% in the lowest-risk decile to 7.95% in the highest, with the top decile containing 58.3% of participants with HCV exposure and the top three deciles containing 85.5%.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFive demographic variables were sufficient to build a useful HCV risk model in a nationally representative US sample. Most HCV-exposed individuals were concentrated in the highest predicted-risk groups, suggesting that this approach could help prioritize and optimize testing where universal screening uptake remains incomplete. As no laboratory data is required, it may also be practical in data-limited settings and adaptable in other health systems.\u003c/p\u003e","manuscriptTitle":"Development and Validation of a Machine Learning Model for Hepatitis C Virus Exposure: A Demographic Screening Approach for the US Population","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-15 07:15:23","doi":"10.21203/rs.3.rs-9361729/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2e9db109-c7ce-442e-a7b9-5210356cf10d","owner":[],"postedDate":"April 15th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-27T15:54:14+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-15 07:15:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9361729","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9361729","identity":"rs-9361729","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}