Machine learning based assessment of preclinical health questionnaires

Area of Science:

• Health Informatics
• Machine Learning Applications
• Public Health

Background:

• Increasing availability of patient health data from diverse sources like wearables and medical devices.
• Need for advanced analytical techniques to interpret complex health datasets.
• Focus on preclinical health assessment using patient-reported information.

Purpose of the Study:

• To investigate the application of modern Machine Learning (ML) techniques for preclinical health assessment.
• To develop and evaluate an ML model for pattern detection in risk assessment using questionnaire data.
• To assess the health conditions of pregnant women through data analysis.

Main Methods:

• Development of a questionnaire distributed to pregnant women in Mureș County, Romania.
• Extraction and analysis of data from completed questionnaires.
• Implementation of a Machine Learning (ML) model for pattern detection and risk assessment.

Main Results:

• Data collected from 1278 women, with specific focus on smoking habits during pregnancy (381 smokers, 216 quitters).
• The developed ML model achieved a high accuracy of 98% for the case study.
• Demonstrated feasibility of using ML for analyzing questionnaire data in a clinical setting.

Conclusions:

• The proposed solution enables efficient digitization and analysis of questionnaire data.
• The ML approach requires reduced computational resources (memory and processing power).
• Offers a simple and effective method for leveraging patient data for health assessment.