A decision tree-based classifier for E-visit service provision

Area of Science:

• Health Informatics
• Decision Science
• Health Policy

Background:

• Healthcare systems face increasing demand for efficient service delivery.
• Subsidization of e-visit services by health insurance necessitates accurate patient selection.
• Identifying suitable candidates for e-visits is crucial for optimizing healthcare resource allocation.

Purpose of the Study:

• To propose and validate a decision tree-based e-visit classification approach (DTEVCA).
• To identify clinic visits and patients who would benefit most from e-visit services.
• To provide a clear guideline for insurance agencies regarding e-visit implementation.

Main Methods:

• Utilized a decision tree-based classification model.
• Employed clinic medical records and patient demographic data for analysis.
• Validated the approach using a large dataset from Taiwan's National Health Insurance.

Main Results:

• The DTEVCA accurately classifies in-office visits suitable for conversion to e-visits.
• The decision tree provides straightforward rules for insurance agencies.
• The approach demonstrates efficiency and validity in classifying e-visit candidates.

Conclusions:

• The DTEVCA is an effective tool for classifying e-visit suitability.
• The approach supports policy formulation for e-visit implementation and physician adoption.
• The DTEVCA framework is adaptable for other new services beyond e-visits.