(213 days)
Freespira is intended for use as a relaxation treatment for the reduction of stress by leading the user through guided and monitored breathing exercises. The device is indicated as an adjunctive treatment of symptoms associated with panic disorder (PD) and/or posttraumatic stress disorder (PTSD), to be used under the direction of a healthcare professional, together with other pharmacological and/or non-pharmacological interventions.
Freespira® is intended for use as a relaxation treatment for the reduction of stress by leading the user through guided and monitored breathing exercises. The device is indicated as an adjunctive treatment of symptoms associated with panic disorder (PD) and/or posttraumatic stress disorder (PTSD), to be used under the direction of a healthcare professional, together with other pharmacological and/or nonpharmacological interventions. Freespira is authorized and overseen by a licensed healthcare provider. Patients are trained to use the Freespira Sensor and the Freespira Mobile App to measure and display their exhaled carbon dioxide (EtCO₂) level and respiration rate (RR) and how different breathing habits affect EtCO2 levels.
1. A table of acceptance criteria and the reported device performance
| Acceptance Criteria (Primary Effectiveness Hypothesis) | Reported Device Performance (2-month post-treatment follow-up) |
|---|---|
| At least 50% of study participants will experience a response to the treatment (defined as a 6-point reduction in CAPS-5 score from baseline). | 93% of participants demonstrated a response (95% CI: 77-99%). |
2. Sample size used for the test set and the data provenance
The document refers to a "prospective, single arm, un-blinded investigation of Freespira". While a specific sample size for a "test set" (implying a separate validation cohort) is not explicitly stated in the context of AI device testing, the clinical study enrolled patients with a primary diagnosis of PTSD. The number of participants in this clinical study is not directly specified by a single number, but results are presented as percentages and confidence intervals, implying a participant cohort. Given the confidence intervals provided (e.g., 95% CI between 77-99% for 93% response), the sample size is likely small, potentially in the range of 20-40 participants for this type of study.
The data provenance is prospective clinical data collected from patients with PTSD. The country of origin is not explicitly stated, but the company, Palo Alto Health Sciences, Inc., is based in Kirkland, WA, U.S.A., which suggests the study was likely conducted within the United States.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
The ground truth was established using the Clinician Administered PTSD Scale (CAPS-5). This is a clinician-administered assessment, implying that a qualified clinician (an "expert") was involved in administering and scoring the scale. The document does not specify the exact number of clinicians involved in assessing each patient or their specific qualifications (e.g., years of experience, specific medical specialty), but it implies the use of healthcare professionals capable of administering and interpreting this validated psychiatric assessment tool.
4. Adjudication method for the test set
The document does not detail an adjudication method for the CAPS-5 scores (e.g., 2+1, 3+1). Since CAPS-5 is a standardized, clinician-administered scale, it's typically administered by a single trained clinician. If there were multiple clinicians involved in assessments over time, standard clinical practice would involve training and inter-rater reliability checks, but these are not explicitly described as formal adjudication steps for the ground truth in this submission.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance
No, an MRMC comparative effectiveness study involving human readers and AI assistance was not mentioned or presented in this document. The study evaluates the Freespira device as a standalone treatment, not as an assistive tool for human readers.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
Yes, a standalone performance study was done in the context of the device's clinical effectiveness. The Freespira device, which guides users through breathing exercises and measures EtCO2 and RR, was evaluated in a clinical trial to determine its efficacy as an adjunctive treatment for PTSD symptoms. The primary endpoint related to the reduction in CAPS-5 scores directly measures the outcome of using the device.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)
The primary ground truth used was clinically validated outcome data based on the Clinician Administered PTSD Scale (CAPS-5) scores. This scale is administered by clinicians and reflects a standardized assessment of PTSD symptoms and severity. Additionally, patient-reported questionnaires like PHQ-9, SF-36, and PDSS were used as secondary assessments.
8. The sample size for the training set
The document does not explicitly mention a training set in the context of AI/algorithm development. The clinical study described appears to be an evaluation of the device as a whole (hardware + software + therapeutic protocol) rather than a specific validation of an AI algorithm trained on a dataset. The device's function involves measuring EtCO2 and RR and providing biofeedback, which might rely on pre-programmed logic rather than a machine learning model that requires a distinct "training set."
9. How the ground truth for the training set was established
As no specific "training set" for an AI algorithm is mentioned, the method for establishing its ground truth is not applicable based on the provided text. The device primarily works by guiding users based on physiological measurements, rather than making diagnostic or predictive outputs that would require a ground truth established for an AI model.
§ 882.5050 Biofeedback device.
(a)
Identification. A biofeedback device is an instrument that provides a visual or auditory signal corresponding to the status of one or more of a patient's physiological parameters (e.g., brain alpha wave activity, muscle activity, skin temperature, etc.) so that the patient can control voluntarily these physiological parameters.(b)
Classification. Class II (special controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter when it is a prescription battery powered device that is indicated for relaxation training and muscle reeducation and prescription use, subject to § 882.9.