Predicting changes in hypertension control using electronic health records from a chronic disease management program

Area of Science:

• Clinical informatics
• Health informatics
• Predictive analytics in healthcare

Background:

• Hypertension management is complex and costly.
• Electronic health records (EHRs) offer valuable data for chronic disease management.
• Predicting deterioration in hypertension control is crucial for timely intervention.

Purpose of the Study:

• To predict transition points of hypertension control using EHR data.
• To identify key clinical features that indicate changes in hypertension control status.
• To lay the groundwork for personalized hypertension management strategies.

Main Methods:

• A cohort of 1294 hypertension patients was analyzed.
• Clinical data features were extracted and refined using information gain criteria.
• A random forest classifier was employed to predict transition points.

Main Results:

• Key predictors included assessment patterns, comorbidities, procedures, and medication history.
• The random forest model achieved a c-statistic of 0.836.
• The model demonstrated an accuracy of 0.773 in predicting control transitions.

Conclusions:

• Accurate prediction of hypertension control transition points was achieved.
• This predictive capability is a vital step towards personalized hypertension management.
• The findings support the use of EHR data for proactive chronic disease management.