Search Results

Stanza is a prescription digital therapeutic that provides Acceptance and Commitment Therapy, a form of Cognitive Behavioral Therapy, and is indicated for the treatment of fibromyalgia symptoms in adult patients.

Stanza is a prescription-only smartphone-based digital therapy that delivers a self-guided program of Acceptance and Commitment Therapy (ACT), a form of Cognitive Behavioral Therapy (CBT), for patients with fibromyalgia.

Stanza is designed to be a 12-week treatment, followed by open access to the program materials for a period of time. Stanza contains 42 unique daily sessions, organized into 8 chapters, which comprise the core program. Once the core program is completed, patients have access to reinforcement sessions, organized into playlists, to strengthen the skill building and reinforce the core program.

To use Stanza, patients first download Stanza from the App store or Google Play store and then securely log in and are authorized. Patients go through a brief tutorial of Stanza's major features and exit the tutorial onto the Today page , which is a patient's home base for the treatment program.

The core program is organized into eight chapters (Table 1). Each of the eight chapters are intended to train the patient on the core ACT skills: Acceptance, Values, Mindfulness, Defusion, Self-as-Context, and Willingness/Committed Action.

Each daily Session may include the following types of interactive activities:

• A lesson to build an ACT skill .
• . A guided audio awareness activity
• A journaling activity to help patients reflect on incorporating ACT skills in their daily lives .
• . Self-guided pacing or exercise activity to help patients gradually increase their activity level and physical functioning

Here's a summary of the acceptance criteria and the studies that prove Stanza meets those criteria:

1. Table of Acceptance Criteria and Reported Device Performance

The provided text doesn't explicitly state quantitative acceptance criteria in the typical sense (e.g., "device must achieve X% sensitivity"). Instead, the regulatory approval hinges on demonstrating:

• Effectiveness: Improvement in fibromyalgia symptoms compared to a control.
• Safety: No device-related adverse events.
• Software Reliability: Performs as intended.
• Labeling Sufficiency: Clear and accurate information for users and prescribers.

Here's how the device performance maps to these broader criteria based on the clinical study summaries:

2. Sample Size for Test Set and Data Provenance

The primary clinical evidence comes from three studies: PROSPER-FM, SMART-FM, and REACT-FM. These studies collectively serve as the "test set" for the device's performance.

• PROSPER-FM Study:
  • Sample Size: Interim look results reported for the first 90 randomized subjects (44 in Digital ACT arm, 46 in Digital Symptom Tracker arm). 159 participants were screened at 13 study sites.
  • Data Provenance: Multicenter (up to 30 North American sites planned for full study), randomized, active-controlled. The study description implies a prospective design for this interim analysis.
• SMART-FM Study:
  • Sample Size: 67 participants were enrolled and randomized (39 in Tempo Digital ACT arm, 28 in Tempo Digital ST arm). 106 participants were screened.
  • Data Provenance: Multicenter (up to 10 North American sites planned), randomized, controlled, pilot study. Implies a prospective design. "Tempo" was the previous version of Stanza.
• REACT-FM Study:
  • Sample Size: Interim assessment performed on the first 95 enrolled patients. "Completer analysis" was done on N=54 participants who completed end-of-treatment outcome data collection.
  • Data Provenance: Virtual, single-arm investigation, assessing effectiveness in a real-world setting. This is a prospective, ongoing study.

3. Number of Experts Used to Establish Ground Truth and Qualifications

The studies described are clinical trials measuring patient-reported outcomes (PROs) and adverse events, not diagnostic performance where "ground truth" is typically established by medical experts reviewing images or pathology.

• Ground Truth Establishment: The primary endpoints (PGIC and FIQ-R scores) are patient-reported. These instruments themselves are validated measures designed to capture the patient's subjective experience of their condition and improvement.
• Expert Oversight: While not explicitly stated as "experts establishing ground truth for individual cases," the studies involved "study sites" and "clinical study staff" (in the case of REACT-FM). These would implicitly include healthcare professionals (e.g., physicians, nurses, study coordinators) qualified to manage the trials, ensure proper patient selection based on diagnostic criteria (2016 American College of Rheumatology Preliminary Diagnostic Criteria for fibromyalgia), and assess safety events. The PROSPER-FM study conducted safety assessments in clinic or via videoconference.

4. Adjudication Method for the Test Set

The concept of an "adjudication method" (like 2+1 or 3+1) is typically used in diagnostic studies where multiple reviewers interpret data and discrepancies are resolved by a higher-level expert. This is not directly applicable to Stanza's clinical trials, as the primary outcomes are patient-reported questionnaires (PGIC, FIQ-R) and adverse event reporting.

• PGIC and FIQ-R: These questionnaires are self-administered by patients, meaning the "adjudication" is inherent in the design of the validated scales themselves and aggregated statistically.
• Adverse Events: Adverse events were reported by participants and assessed by study staff. While the text states that none were attributed to the device, it doesn't detail a formal, multi-expert adjudication process for adverse events beyond typical clinical trial monitoring.

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

No, an MRMC comparative effectiveness study was not conducted in the context typically understood for diagnostic devices. MRMC studies are designed to compare the performance of human readers (e.g., radiologists) with and without AI assistance on a set of cases.

Stanza is a direct-to-patient digital therapeutic. The clinical studies (PROSPER-FM, SMART-FM) are comparative effectiveness studies comparing patients using the Stanza app (Digital ACT arm) to patients using a control app (Digital Symptom Tracker arm), not comparing human readers' performance with and without AI. Therefore, the concept of "effect size of how much human readers improve with AI vs without AI assistance" does not apply to this device.

6. Standalone Performance Study (Algorithm Only)

Yes, a standalone performance was effectively done. Stanza is the algorithm/digital therapeutic being evaluated directly in the PROSPER-FM and SMART-FM studies. Its performance (i.e., its ability to improve fibromyalgia symptoms) was measured without requiring human interaction from a clinician during its use. The device itself delivers the therapy. The primary endpoints (PGIC, FIQ-R) directly reflect the patient's response to the therapeutic intervention provided solely by the Stanza app. The comparator arm (Digital Symptom Tracker) serves as a control for this standalone performance.

7. Type of Ground Truth Used

The ground truth for evaluating Stanza's effectiveness and safety was based on:

• Patient-Reported Outcomes (PROs):
  • Patient Global Impression of Change (PGIC): A 7-point scale reflecting the patient's subjective assessment of overall improvement.
  • Revised Fibromyalgia Impact Questionnaire (FIQ-R): A comprehensive questionnaire assessing the impact of fibromyalgia symptoms on daily life (total scores, function, impact, symptoms).
  • Additional PROs like PROMIS Fatigue, PROMIS Sleep, PIPS, CAQ-8, and BDI-II were also used to assess secondary outcomes and patient perspectives.
• Adverse Event Reporting: Self-reported and clinically monitored adverse events during the study period.

These are standard and widely accepted methods for establishing "ground truth" regarding the clinical efficacy and safety of therapeutic interventions, particularly for chronic conditions like fibromyalgia where subjective experience plays a significant role.

8. Sample Size for the Training Set

The provided text does not specify a separate training set size for Stanza as an AI/ML model in the typical sense of a diagnostic or predictive algorithm.

Stanza delivers a pre-defined Acceptance and Commitment Therapy program. While the development of ACT as a therapeutic approach is based on extensive psychological research and clinical practice (analogous to "training data" for a human therapist), the text describes Stanza as a "self-guided program" and a "digital therapeutic" that delivers a fixed curriculum. It's not presented as a machine learning model that continuously learns from patient data to adapt its therapy.

Therefore, there is no explicit "training set" of patient data used to train an underlying AI model within the device in the way this question typically implies for a machine learning algorithm.

9. How Ground Truth for the Training Set Was Established

As noted above, there's no explicitly defined "training set" for an AI model within Stanza. However, if one were to consider the development of the ACT program itself as analogous to "training," the 'ground truth' for ACT is established through:

• Extensive psychological and clinical research: Decades of research and clinical practice in Acceptance and Commitment Therapy and Cognitive Behavioral Therapy contribute to the theoretical framework and practical exercises incorporated into Stanza.
• Expert consensus: The ACT curriculum is designed by mental health professionals with expertise in this therapeutic modality.
• Previous clinical trials: The effectiveness of ACT as a therapeutic approach in general, and potentially prior pilot studies (like SMART-FM, though framed as an efficacy study here) or iterative development work, would have guided the design and content of the Stanza program.

The text also highlights that "Tempo Digital ACT device is the previous version of Stanza," suggesting an iterative development process informed by earlier results, but it doesn't describe a formal "training set" with ground truth in the AI/ML context.

Ask a specific question about this device