Search Results

EndeavorRx is a digital therapeutic indicated to improve attention as measured by computer-based testing in children ages 8-17 years old with primarily inattentive or combined-type ADHD, who have demonstrated attention issue. Patients who engage with EndeavorRx demonstrate improvements in a digitally assessed measure Tests of Variables of Attention (TOVA) of sustained and selective attention and may not display benefits in typical behavioral symptoms, such as hyperactivity. EndeavorRx should be considered for use as part of a therapeutic program that may include: clinician-directed therapy, medication, and/ or educational programs, which further address symptoms of the disorder.

EndeavorRx is a modification to the previously granted EndeavorRx (DEN200026) with the primary difference being the expansion of the indicated patient population from 8-12 years old to 8-17 years old. In addition, minor software changes were made to improve app accessibility and user engagement. The core therapeutic software technology was not changed.

EndeavorRx is a prescription-only digital therapeutic software indicated for use in the treatment of attention impairment in pediatric patients (8-17 years of age) with primarily inattentive or combinedtype ADHD. EndeavorRx is a software-as-a-medical device (SaMD) that resides on the user's mobile device and can be executed at home.

EndeavorRx is engineered as a therapeutically active treatment for attention in pediatric patients affected by ADHD. EndeavorRx is built on Akili's proprietary, patented, technology platform and uses adaptive algorithms (also known as Selective Stimulus Management Engine, SSME™) to deliver stimuli that are designed to engage the patient in a manner that improves their attention function. In a closedloop system, the adaptive SSME™ algorithms automatically adjust the difficulty level for a personalized treatment experience that is tailored to the needs of each individual patient.

EndeavorRx is delivered through a video game experience which leverages art, music, storytelling, and reward cycles to keep patients engaged. The adaptive algorithm constantly pushes patients precisely at predefined performance bounds relative to each individual, such that they are continuously encouraged to exceed their historic performance. The science behind EndeavorRx was developed at the University of California, San Francisco by Adam Gazzaley, M.D., Ph.D., Founding Director of the University of California San Francisco's Neuroscape and Akili's Chief Science Advisor.

The basic program inputs are steering, which is accomplished by using the mobile device's internal accelerometer to measure the degree to which it is tilted, and tapping, which is accomplished using the touch screen to measure correct and incorrect targeting. The basic outputs are the visual display of the game progression along with audio, which is accomplished by using the internal high resolution display and internal speaker. The program includes features to ensure it is used per the prescribed regimen (approximately 25 minutes per day, 5 days per week, for 4 weeks or as recommended by the health care provider).

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided FDA 510(k) summary for EndeavorRx (K231337):

Device Name: EndeavorRx
Indication for Use: Digital therapeutic indicated to improve attention as measured by computer-based testing in children ages 8-17 years old with primarily inattentive or combined-type ADHD, who have demonstrated attention issue.

1. Table of Acceptance Criteria and Reported Device Performance

The provided document primarily details efficacy results rather than explicitly stated "acceptance criteria" as pass/fail thresholds for regulatory approval outside of the comparison tables for substantial equivalence. However, the core acceptance criterion for efficacy appears to be a significant positive mean change from baseline in the TOVA-ACS score.

2. Sample Sizes and Data Provenance

• Test Set (Clinical Study for Subject Device EndeavorRx v3.0 - K231337):

  • "Efficacy Population" (Complete Case Analysis - CCA): N = 146
  • "Safety Population" (Intent-to-Treat with Multiple Imputation - ITT with MI): N = 162
  • Provenance: Multi-center open-label study across 14 sites in the US.
  • Retrospective/Prospective: The description "multi-center open-label study enrolled 162 adolescents" indicates a prospective study.

• Predicate Device Study (STARS - DEN200026):

  • "ITT Population": N = 179 (for TOVA-ACS primary endpoint)
  • Provenance: Multi-site study across 20 sites in the US.
  • Retrospective/Prospective: Described as "Randomized, controlled, parallel arm," indicating a prospective study.

3. Number of Experts and Qualifications for Ground Truth (Test Set)

The ground truth for the device's efficacy is based on the objective, computer-based testing measure, TOVA-ACS (Test of Variables of Attention - ADHD Composite Score). This is a normed, standardized test, not subjective expert assessment requiring multiple readers or qualifications.

• Diagnosis of ADHD: Confirmed by Mini-International Neuropsychiatric Interview for Children and Adolescents (MINI-Kid) Version 7.0.2, administered by a trained clinician. The document does not specify the number or specific qualifications (e.g., years of experience) of these clinicians, but implies they are qualified to administer the MINI-Kid and make diagnoses based on DSM-5 criteria.
• High inattention: Measured by baseline TOVA-ACS score ≤ -1.8. This is a predefined numerical threshold.

4. Adjudication Method for the Test Set

Since the primary efficacy outcome (TOVA-ACS) is an objective, computer-based measure, and ADHD diagnosis was established by trained clinicians using a standardized interview (MINI-Kid), there is no mention of a traditional "adjudication method" involving multiple human readers reviewing results and resolving discrepancies. The data is either directly collected from the TOVA test or from structured diagnostic interviews.

The study design for the subject device was a "Single arm, open-label" study. This means there was no control group, and no 2+1 or 3+1 adjudication for comparing against a different intervention or sham.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No MRMC study was done. This device is a digital therapeutic, not a diagnostic imaging AI that assists human readers. Its primary claim is to improve attention as measured by a computer-based test, not to aid human interpretation of various cases. The clinical study for the subject device was:

• Standalone (algorithm only without human-in-the-loop performance): Yes, the device itself is the intervention, and its effect is measured directly.
• Human Reader Improvement with AI vs. Without AI Assistance (Effect Size): Not applicable, as this is not an AI-assisted diagnostic tool for human readers. The clinical study evaluated the direct therapeutic effect of the device on patients.

6. Standalone (Algorithm only) Performance

Yes, the primary efficacy endpoint (TOVA-ACS change) reflects the standalone performance of the EndeavorRx device in its effect on attention, as it is a direct therapeutic intervention.

7. Type of Ground Truth Used

• Primary Efficacy: Objective, digitally assessed measure of sustained and selective attention (Test of Variables of Attention - TOVA-ACS). This is a quantitative, standardized neurocognitive test.
• Secondary Outcomes: Clinical scales like ADHD-RS (Attention-Deficit/Hyperactivity Disorder Rating Scale), which are subjective clinician/parent-reported measures designed to assess symptom severity.
• Diagnosis (for inclusion criteria): Standardized diagnostic interview using DSM-5 criteria (MINI-Kid), performed by trained clinicians.

8. Sample Size for the Training Set

The document does not specify the sample size for the training set of the EndeavorRx device's algorithms (Selective Stimulus Management Engine, SSME™). It only describes the adaptive nature of the algorithm that "automatically adjust the difficulty level for a personalized treatment experience." This suggests an on-device adaptive learning system rather than a pre-trained machine learning model in the conventional sense where a large, separate training dataset is explicitly used for model development and then locked. However, the underlying "proprietary, patented, technology platform" and SSME™ algorithms would have been developed and refined using clinical and behavioral data over time, but details on such training data are not provided in this 510(k) summary.

9. How the Ground Truth for the Training Set was Established

As the document does not detail a distinct "training set" in the context of a fixed machine learning model, it also does not explain how ground truth was established for such a set.

The "adaptive algorithms" (SSME™) adjust difficulty based on the individual patient's real-time performance ("closed-loop system," "automatically adjust the difficulty level," "pushes patients precisely at predefined performance bounds relative to each individual"). This implies that the individual patient's performance within the game itself serves as the dynamic "ground truth" to adapt the therapeutic experience, rather than a separate, pre-labeled training dataset.

Ask a specific question about this device