Large-Scale Data Mining to Optimize Patient-Centered Scheduling at Health Centers

Area of Science:

• Health Informatics
• Data Mining in Healthcare
• Operations Research

Background:

• Patient-centered appointment access is crucial for community health centers (CHCs).
• Effective scheduling relies on advanced data analytics and Electronic Health Record/Practice Management (EHR/PM) systems.
• Optimizing scheduling requires understanding factors influencing patient appointment adherence.

Purpose of the Study:

• To optimize patient-centered appointment scheduling in CHCs.
• To identify critical factors influencing appointment adherence using data mining.
• To improve open-access scheduling through data-driven insights.

Main Methods:

• Collected data from EHR/PM systems across three Indiana CHCs.
• Integrated data into a multidimensional data warehouse.
• Employed decision tree modeling, logistic regression, visual analytics, and n-gram modeling for data mining.

Main Results:

• Appointment adherence strongly correlated with scheduling time dimensions, especially lead time.
• Time of day, season, patient demographics, and clinical characteristics were significant predictors.
• Implementing findings reduced missed appointment rates by 16% in an interventional CHC.
• N-gram text mining revealed reasons for same-day appointment requests.

Conclusions:

• Variability exists in factors affecting patient appointment adherence across different healthcare settings.
• Optimizing open-access scheduling requires continuous data monitoring and analysis.
• Enhanced in-CHCs data analytic capabilities are needed to redesign care delivery for improved access and efficiency.