Identifying patterns and predictors of lifestyle modification in electronic health record documentation using...

Area of Science:

• Cardiology
• Health Informatics
• Preventive Medicine

Background:

• Hypertension affects millions of US adults, with a significant portion having uncontrolled blood pressure.
• Uncontrolled hypertension elevates risks for severe cardiovascular events, including death, stroke, heart failure, and myocardial infarction.
• Current hypertension management guidelines emphasize lifestyle modifications like diet and exercise.

Purpose of the Study:

• To identify predictors and timing of lifestyle modification assessment or advice in adult hypertension patients.
• To inform tailored interventions for improving lifestyle modification documentation and hypertension control.
• To leverage electronic health record (EHR) data for enhanced understanding of hypertension care processes.

Main Methods:

• Analysis of EHR data from 14,360 adult hypertension patients at an academic medical center.
• Application of statistical and machine learning methods, including Random Forest and logistic regression.
• Evaluation of multiple time points for lifestyle modification documentation.

Main Results:

• Multiple patient, clinic, and provider variables were statistically significant predictors of lifestyle modification documentation.
• Random Forest achieved an Area Under the Receiver Operator Curve (AUROC) of 0.831 for classifying documentation at any time.
• Logistic regression demonstrated an AUROC of 0.685 for classifying documentation within ≤3 months.

Conclusions:

• Analyzing EHR data, including narrative and coded information, enhances understanding of lifestyle modification timing and associated factors.
• Identified predictors can guide the development of targeted interventions to improve hypertension care processes.
• Improved documentation of lifestyle modifications has the potential to enhance overall hypertension control.