AI CPT Codes in the Wild

A discussion and analysis on:

Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance Claims

NEJM AI, Published November 9, 2023

DOI: 10.1056/AIoa2300030

Introduction

This study investigates the real-world clinical adoption of FDA-approved medical AI devices by analyzing U.S. insurance claims data between early 2018 and mid 2023. By identifying specific Current Procedural Terminology (CPT) codes associated with these devices, the research provides insights into their utilization across various medical specialties and geographic regions.

Detailed Analysis

1. Identification of AI-Related CPT Codes

The researchers compiled a list of CPT codes corresponding to FDA-approved medical AI devices. This involved reviewing official sources, web resources, and insurance company policies to accurately map AI devices to their respective billing codes.

Why This Matters: Establishing a comprehensive list of AI-related CPT codes is essential for tracking the adoption and utilization of AI technologies in clinical practice.

CareSight lens: When evaluating AI tools, we should consider whether the device or software has an associated CPT code for its intended clinical use, as this can influence reimbursement and adoption rates.

2. Analysis of Insurance Claims Data

Utilizing a national insurance claims database, the study quantified the usage of each identified CPT code, shedding light on the frequency and distribution of AI device utilization across different specialties and regions.

Why This Matters: Analyzing claims data offers a real-world perspective on how widely and effectively AI devices are being integrated into clinical workflows.

CareSight lens: By assessing CPT-based clinical adoption of AI tools as reflected via insurance claims data, we can provide informed recommendations regarding its market penetration and acceptance.

3. Specialty and Geographic Distribution

The study found variability in the adoption of AI devices across medical specialties and geographic locations, indicating that certain areas and fields are more inclined to integrate AI technologies than others.

Why This Matters: Understanding the distribution patterns can help identify barriers to adoption and areas where AI integration may be particularly beneficial or underutilized.

CareSight lens: When assessing AI tools, we should consider specialty-specific and regional adoption trends to tailor recommendations that address unique challenges and opportunities in different contexts.

Practical Implications for Recommendations

1. Evaluating Reimbursement Potential:

   • We should guide clients in understanding the importance of obtaining specific CPT codes for their AI devices, as this can significantly impact reimbursement and adoption.

2. Assessing Market Penetration:

   • By analyzing insurance claims data, we can provide insights into the real-world utilization of AI tools, helping clients gauge market acceptance and identify areas for growth.

3. Identifying Adoption Barriers:

   • Understanding specialty and geographic variability bolsters strategies to address potential barriers to, such as targeted education or addressing regional regulatory differences.

Concluding Reflection

This study underscores the value of insurance claims data in assessing the clinical adoption of medical AI devices. For GPT-driven recommendations, incorporating such real-world evidence is crucial in evaluating an AI tool's market presence, reimbursement landscape, and identifying factors that influence its integration into clinical practice.