Decoding FDA Labeling of Prescription Digital Therapeutics: A Cross-Sectional Regulatory Study

TL;DR

This study analyzes FDA labeling of digital therapeutics to understand how they are regulated and described for medical use.

Contribution

The first systematic regulatory labeling analysis of all FDA-cleared prescription digital therapeutics.

Findings

Most PDTs were cleared via the 510(k) pathway, with a focus on neurological and psychiatric conditions.

Labeling language varied significantly, with only one PDT receiving unambiguous treatment language.

Sponsors were all US-based, concentrated in digital health hubs like San Francisco.

Abstract

Background

Prescription digital therapeutics (PDTs) are software-only, FDA-regulated medical devices prescribed to prevent, manage, or treat disease. Despite increasing FDA clearance, there remains limited understanding of how PDTs are regulated and labeled from a product, sponsor, and indication standpoint.

Objective

This study aims to conduct the first systematic regulatory labeling analysis of all FDA-cleared PDTs, characterizing their approval pathways, sponsor profiles, clinical indications, and therapeutic language.

Methods

We performed a retrospective descriptive analysis of all software-only PDTs cleared by the FDA as of May 2025. Publicly available decision summaries, classification orders, and device listings were reviewed. Each PDT was examined by regulatory pathway, reviewing office, product code, sponsor geography, ICD-11 mapping, and FDA-approved labeling language, with a focus on terms of therapeutic intent and age-based eligibility.

Results

Thirteen PDTs were identified, with eight (61.5%) cleared via the 510(k) pathway and five (38.5%) via de novo classification. The most targeted neurological or psychiatric conditions were reviewed by the corresponding FDA offices. Sponsors were all US-based and concentrated in digital health hubs, particularly San Francisco. Therapeutic indications ranged from insomnia and diabetes to migraine and opioid use disorder. Labeling language varied: 11 PDTs included treatment claims, although most used modifiers such as “symptom improvement” or "aid in the management." Only one PDT, CT-132 for migraine, received a clean treatment label, defined as unambiguous treatment language without qualifiers, reflecting direct disease-targeting intent. Two PDTs deviated notably: reSET-O was labeled solely to increase outpatient treatment retention for opioid use disorder, and EndeavorRx was indicated to improve attention function without claiming to treat ADHD. Age-based eligibility spanned pediatric to adult definitions, consistent with FDA device criteria.

Conclusions

This study reveals meaningful variation in how PDTs are classified, labeled, and geographically distributed. FDA-sanctioned language plays a critical role in defining therapeutic scope and impacts how PDTs are interpreted by clinicians, payers, and patients. These insights mark a step forward in understanding the regulatory architecture of digital medicine.

Figures