Beyond Time-Based Billing: An AI-Enabled Complexity Model for Secure Message Reimbursement | JoVE Visualize

Area of Science:

Health Informatics
Artificial Intelligence in Medicine
Healthcare Administration

Background:

Current telehealth reimbursement relies on evaluation and management duration, which does not accurately reflect the complexity of provider responses to secure patient messages.
There is a need for reimbursement models that better represent the intricate nature of asynchronous telehealth clinical care.
Existing frameworks fail to capture the cognitive effort and clinical judgment involved in digital patient communication.

Purpose of the Study:

To develop and validate an AI-enabled model for telehealth reimbursement that quantifies clinical complexity.
To create a more accurate billing framework for asynchronous telehealth services based on provider cognitive effort and clinical content.
To establish a foundation for reforming telehealth reimbursement to ensure fair provider compensation.

Main Methods:

Analysis of 149,499 secure patient messages from a large health system (2023-2024).
Utilized supervised pseudo-labeling with rigorous annotation guidelines (Cohen's κ = 0.816) on a subset of messages.
Developed a gradient boosting classifier using features of electronic health record engagement (Domain Engagement) and linguistic complexity (Cognitive Judgment), validated through cross-validation and leave-one-provider-out testing.

Main Results:

The complexity-based model achieved an area under the curve (AUC) of approximately 0.82, significantly outperforming a time-based model (AUC ~0.57).
Key predictors of complexity included unique medical concepts, breadth of clinical content, and emotional urgency, not message length or duration.
Billable threads demonstrated higher medical density and emotional intensity, validating the Domain Engagement and Cognitive Judgment metrics.

Conclusions:

AI-driven complexity metrics more accurately represent provider cognitive effort in telehealth compared to time-based measures.
The developed model offers a practical, accurate, and explainable approach for telehealth billing reform.
This framework supports fairer provider reimbursement and enhances the sustainability of telehealth services.