Proof-of-concept NHANES model

NHANES-derived HCV marker risk calculator

This page scores a single patient locally from a 5-variable XGBoost model derived from NHANES data for the paper's fixed-cycle primary analysis using birth year, sex, birthplace, race/ethnicity, and income-to-poverty ratio. Risk is displayed on the paper-aligned calibrated probability scale for evidence of current or past HCV infection, not confirmed active viremia. This proof-of-concept calculator has not been externally validated for clinical decision-making.

Inputs 5 demographic variables
Calibration Training-derived Platt scaling
Reference Fixed-cycle validation deciles
Model Loading...
Validation Cohort Loading...
Primary AUROC Loading...

Patient Inputs

Patient demographics

Use notes

  • This score estimates evidence of current or past HCV infection, meaning the patient may be currently infected or may have been infected previously.
  • This calculator is intended for research, education, and outreach-prioritization illustration, not for diagnosis or standalone clinical decision-making.
  • Calibration and decile context are anchored to the paper's fixed-cycle split (training n = 48,434; validation n = 20,762).
  • The NHANES-derived model should not be assumed to transport to other healthcare systems without local validation and recalibration.
  • A positive status screen still needs standard HCV testing workflow, typically antibody testing with confirmatory HCV RNA testing.
  • The calculator runs entirely in the browser after the page is opened; users should not enter names, record numbers, or other direct identifiers.
  • Income should be entered as poverty-income ratio, not raw annual dollars.

Model Output

Risk estimate

Enter the five variables to estimate calibrated infection risk and place the patient within the paper's fixed-cycle validation deciles.