Using Wearables and Machine Learning to Enable Personalized Lifestyle Recommendations to Improve Blood Pressure

Area of Science:

• Cardiovascular Health
• Biomedical Informatics
• Machine Learning in Healthcare

Background:

• Blood pressure (BP) is a critical health indicator influenced by lifestyle factors like activity and sleep.
• Individual impact of lifestyle factors on BP is not well understood, leading to generalized recommendations.
• Personalized BP management requires understanding individual-specific factor influences.

Purpose of the Study:

• To investigate the relationships between BP and various lifestyle factors.
• To develop a system for personalized BP modeling and precise lifestyle recommendations.
• To move beyond general lifestyle advice towards individualized health interventions.

Main Methods:

• Automated data collection via home BP monitors and wearable activity trackers.
• Application of Random Forest with Shapley-Value-based Feature Selection for personalized modeling.
• Generation of precise lifestyle recommendations based on identified key factors.

Main Results:

• A clinical study with 25 patients demonstrated accurate personalized BP models and identification of varying top influencing factors.
• Personalized recommendations led to significant systolic and diastolic BP reductions (3.8 and 2.3 mmHg).
• Control group showed minimal BP decrease (0.3 and 0.9 mmHg), validating recommendation effectiveness.

Conclusions:

• Wearable technology and machine learning can create effective personalized models for BP management.
• Precise, data-driven lifestyle recommendations show significant potential for improving cardiovascular health.
• The study highlights a shift towards individualized healthcare strategies for hypertension.