507 digital therapeutics are approved in four countries. Effect sizes aren't comparable

Digital therapeutics — software-driven medical interventions intended to treat a condition rather than to inform a clinician about it — have been approved in growing numbers for a decade without anyone being able to say confidently how many exist. A dataset published in Scientific Data on 23 April closes that particular gap [s1].

The count

The authors set out to build a standardised dataset covering approved digital therapeutic devices in four countries: the United States, China, Germany and Belgium [s1]. Devices were identified through systematic searches of national public databases combined with a keyword strategy, then put through multi-stage screening, standardised information extraction and expert arbitration [s1].

The result is 507 devices meeting international digital therapeutics criteria, described across 18 product characteristics [s1].

The paper's motivating observation is worth stating plainly, because it explains why a dataset is a publishable contribution in this field at all: digital therapeutics are developing rapidly with regulatory frameworks being established in multiple countries, and there has been no unified, structured data resource to support cross-national comparative or industry analysis [s1]. Until now the category has been governed by four or more regulators, each with its own definitions, databases and disclosure norms, and no one had assembled the result into a single comparable table.

Why the count doesn't answer the question people want answered

A total of 507 approvals invites the inference that digital therapeutics are an established treatment class. Two randomised trials published within a week of each other in May show why that inference outruns the evidence. Both tested a digital therapeutic. Both were positive. They are not describing the same order of clinical effect.

Migraine. An integrated analysis in npj Digital Medicine on 13 May pooled two similarly designed randomised controlled trials of CT-132, a digital therapeutic delivering cognitive behavioural interventions as an adjunct to existing migraine therapy [s2]. ReMMi-D enrolled 568 patients on prescription migraine medications that were not migraine-specific; ReMMiD-C enrolled 110 patients taking migraine-specific medications, specifically CGRP inhibitors [s2]. Across the pooled 678 participants, CT-132 significantly reduced monthly migraine days, with a least-squares mean difference of -0.86 days and P = 0.003 [s2]. The analysis reports high engagement and no device-related safety concerns [s2].

Tinnitus. A double-blind, sham-controlled randomised trial published in JAMA Otolaryngology-Head & Neck Surgery on 7 May tested a prototype app delivering educational counselling and cognitive behavioural therapy elements for chronic tinnitus [s3]. Sixty participants — 33 (55%) female, median age 58.5 years (IQR 52.0-64.0) — were assessed, 30 in the therapeutic app group and 30 in a sham-control group whose app lacked the therapeutic functions, across a 16-week treatment phase and an 8-week follow-up [s3]. Mean baseline Tinnitus Handicap Inventory score was 41.3 (SD 16.9) [s3]. At week 16 the therapeutic group improved by 20.4 THI points more than sham (95% CI -28.2 to -12.6), and the difference was still 18.3 points at week 24 (95% CI -26.4 to -10.1) [s3]. Tinnitus worsened in one patient in the sham group; no serious adverse events or hazardous device malfunctions were observed [s3].

Reading the two results together

The migraine result is a post-hoc integrated analysis of two trials, and the paper says so [s2]. Its effect — 0.86 fewer migraine days per month — is real, statistically robust across 678 patients, and small. For a patient with fifteen migraine days a month it is a reduction of under a day.

The tinnitus result comes from 60 patients, which is small enough that the confidence interval on the primary outcome spans 12.6 to 28.2 THI points [s3]. But it is a double-blind trial against an active sham that stripped out the therapeutic content while keeping the app, which is a more demanding comparator than a waitlist, and the effect persisted eight weeks after the therapeutic functions were switched off [s3].

Those are different kinds of evidence supporting different magnitudes of claim, and both sit inside the same regulatory category. That is the substantive problem the 507-device count exposes rather than solves. "Approved digital therapeutic" is a regulatory status, not a statement about effect size, comparator quality, or durability.

What the dataset makes possible

The value of the Scientific Data release is that questions of this kind become answerable at scale for the first time. With 507 devices described across 18 standardised characteristics [s1], it becomes possible to ask which conditions the field has concentrated on, how evidence requirements differ across the four regulators covered, and whether devices approved in one jurisdiction appear in the others.

The obvious limitation is coverage. Four countries is not the world, and the paper is explicit that this is a dataset assembled to support cross-national comparison and regulatory decision-making [s1] — a research instrument, not a register.

What to watch

Whether the dataset is extended beyond the four countries covered, and whether it is kept current rather than becoming a snapshot. Whether regulators begin publishing effect-size and comparator information alongside approvals in a structured form. And whether CT-132's 0.86-day effect proves sufficient for payers, which is a separate test from the one the trials were designed to pass.

Sources

1. [s1] A dataset of digital therapeutic device approvals to support regulatory decisions and industry research. Scientific Data, 23 April 2026. https://doi.org/10.1038/s41597-026-07221-3
2. [s2] Integrated analysis of two randomized controlled trials of a digital therapeutic for episodic migraine prevention. npj Digital Medicine, 13 May 2026. ClinicalTrials.gov NCT05853900, NCT06004388. https://doi.org/10.1038/s41746-026-02675-6
3. [s3] Mobile Application as a Digital Therapeutic for Chronic Tinnitus: A Randomized Clinical Trial. JAMA Otolaryngology-Head & Neck Surgery, published online 7 May 2026. jrct.mhlw.go.jp identifier 032230359. https://doi.org/10.1001/jamaoto.2026.0858