Machine Learning for Risk Group Identification and User Data Collection in a Herpes Simplex Virus Patient Registry:...

Area of Science:

• Computational epidemiology
• Machine learning in public health
• Viral infectious disease research

Background:

• Herpes simplex virus (HSV) research faces challenges due to poor data quality, low user engagement, and stigma.
• Accurate data collection is crucial for understanding HSV infection patterns and risk factors.

Purpose of the Study:

• To enhance data collection for a real-world HSV registry.
• To identify key predictors of HSV infection.
• To minimize the number of questions for new users to assess HSV infection risk.

Main Methods:

• Utilized the US National Health and Nutrition Examination Survey (NHANES) database (2015-2016) for HSV-1 and HSV-2 status.
• Developed a random forest machine learning model using Python to analyze demographic and health data.
• Reduced the number of lifestyle-based questions required for risk assessment.

Main Results:

• The model achieved high accuracy (0.91 for HSV-1, 0.96 for HSV-2) and recall (0.88 for HSV-1, 0.98 for HSV-2) in predicting infection risk.
• Significantly reduced the number of data collection questions from 150 to an average of 40.
• Demonstrated high predictability of infection risk with minimal user input.

Conclusions:

• The machine learning algorithm can improve real-world evidence registries by efficiently collecting lifestyle data and identifying HSV risk.
• Future work includes integrating real user data, electronic medical records, and ensuring compliance with data protection regulations (GDPR, HIPAA).