Mental health care has always depended on what patients can recall and report between sessions. Connected devices change that by adding continuous signals: how someone slept, how much they moved, how their heart rate varied, whether they used a prescribed therapy program. That data can sharpen care, and it can also feel like surveillance if it is handled carelessly.
This guide looks at the Internet of Medical Things (IoMT) in mental health from a builder's point of view. It covers which devices and sensors are used, what they can and cannot tell a clinician, how prescription digital therapy devices are now paid for by Medicare, and how to design consent, data flows and clinical review so the program helps rather than harms.
What is IoMT in mental health?
Adding device data to a mental health service?
Agnotic builds mental health apps and clinician tools that bring wearable, phone and therapy-device data into measurement-based care.
IoMT in mental health means connected devices and sensors that collect behavioral or physiological data for psychiatric and psychological care. It covers consumer wearables tracking sleep and activity, smartphone sensing of movement and phone use, and prescription digital therapy devices, with the data shared to clinicians to support assessment, treatment and follow-up between visits.
Unlike a blood pressure cuff, most of these signals do not measure the condition directly. They measure behavior that tends to change with the condition, so their value depends on comparing a person with their own baseline and on a clinician interpreting the change.
Which devices and sensors are used in mental health care?
Four groups of devices are common in mental health programs, and each collects a different kind of evidence. Most programs start with one group rather than all four.
- Wearables: Watches and rings that record sleep duration and timing, step counts, resting heart rate and heart rate variability.
- Smartphone sensing: Location variety, time at home, screen-on time, call and message counts (not content) and typing patterns, gathered in the background.
- Prescription digital therapy devices: Software that delivers a structured behavioral therapy program, prescribed and tracked by a clinician.
- Clinic and home devices: Equipment used in specific treatments, where the device or its companion software logs sessions and settings.
Active data still matters. Short mood check-ins and validated questionnaires, such as the PHQ-9 and GAD-7 that many practices already use, give the passive signals meaning. A drop in sleep alongside a rising questionnaire score tells a much clearer story than either alone.
What can passive sensing tell a clinician?
Passive sensing can show a clinician change over time in sleep, activity, routine and social contact, which often moves with mood, anxiety and relapse risk. It cannot diagnose a condition, and a single day's data rarely means anything on its own.
- Sleep shifts: Later sleep onset, broken sleep or much longer time in bed can precede or accompany a depressive or manic episode.
- Reduced movement: Fewer places visited and more time at home can track withdrawal.
- Routine disruption: Irregular daily patterns can matter for conditions where stable routine is part of treatment.
- Treatment engagement: Whether the patient opened the prescribed program and completed its exercises.
These patterns differ widely between people. Shift workers, new parents and people with physical illness all show changes that have nothing to do with mood. That is why per-patient baselines and clinician review come before any automated alert, and why the program should be explained to patients as support for the conversation, not as a test they can fail.
How does Medicare pay for digital mental health treatment devices?
Medicare pays for digital mental health treatment (DMHT) devices through three HCPCS codes, G0552, G0553 and G0554, created in the calendar year 2025 Physician Fee Schedule final rule. CMS's summary describes them as covering approved devices furnished incident to professional behavioral health services, used with ongoing treatment under a plan of care.
The device rules are specific. The CMS summary for 2026 states that the codes apply to devices cleared under section 510(k) or granted De Novo authorization and classified under 21 CFR 882.5801, and expands them to ADHD digital therapy devices under 21 CFR 882.5803. Section 882.5801 describes a prescription-only computerized behavioral therapy device used as an adjunct to clinician-supervised outpatient treatment.
- G0552: Supply of the device with initial education and onboarding, per course of treatment.
- G0553: The first 20 minutes of monthly treatment management related to the patient's use of the device.
- G0554: Each additional 20 minutes of that monthly management.
- Practitioner cost: The 2026 summary says the billing practitioner must incur the cost of furnishing the device.
For a software team, these rules turn into features. The clinician console needs onboarding records, device usage data the clinician can review, time tracking for management work and a log of interactive communication with the patient, all tied to the patient's plan of care.
How should device data reach the care team?
Device data should reach the care team as a short, reviewed summary inside the tools they already use, not as a stream of raw readings. Behavioral health clinicians rarely have time to study charts of heart rate variability between sessions.
- 1. Collect: The app reads from the phone's health store and approved wearables, and the therapy program reports its own usage.
- 2. Clean: The backend removes duplicates, fills known gaps and computes daily features such as sleep midpoint and step count.
- 3. Compare: Each feature is compared with the patient's own rolling baseline rather than a population norm.
- 4. Summarize: Before each session, the clinician sees a one-page trend view with the questionnaire scores alongside.
- 5. Record: The clinician's interpretation, not the raw data, goes into the note, and the trend view is linked for reference.
If the program uses alerts between sessions, they should be rare, reviewed by a person and paired with a defined response. Passive data should never be the only route by which a patient in crisis is noticed; the app must always offer direct crisis resources. Patients should also be told plainly, at enrollment and in the app, that the data is not watched around the clock.
Which privacy rules apply to mental health device data?
When a provider or health plan runs the program, HIPAA governs the data. When a consumer company runs it outside HIPAA, the FTC Health Breach Notification Rule can apply instead, because 16 CFR 318.2 counts apps and connected devices that track sleep, mental health and similar measures as health care services or supplies.
Either way, passive sensing raises the bar for consent. Location and phone-use data are sensitive even outside a health context, and in mental health they can reveal visits to a clinic, a change in relationships or a period of crisis.
- Per-signal choice: Patients turn each signal on separately, such as sleep yes, location no, and the program still works with fewer signals.
- Derived, not raw: Compute features on the device where possible and send summaries, such as number of places visited, instead of coordinates.
- No secondary use: Behavioral signals are not used for marketing, and analytics tools never receive them.
- Visible to the patient: Patients can see the same summaries the clinician sees.
- Easy pause: A single control pauses all sensing, for example during a hospital stay or a holiday, without leaving care.
How should a practice start a device program?
A practice should start with one clinical question, one patient group and one or two signals, then widen only after clinicians have used the data in real sessions. A narrow pilot shows quickly whether the signal changes any decisions.
- Clinical question: For example, "Can we see early sleep disruption in patients with bipolar disorder between monthly visits?"
- Patient group: A defined cohort who agree to take part, with a clear explanation of what is collected and who sees it.
- Signals: Sleep timing from a wearable plus a weekly questionnaire, rather than every sensor the phone offers.
- Review rhythm: Clinicians look at the trend view before each scheduled session, and the team agrees when an unscheduled contact is warranted.
- Exit criteria: The team decides in advance what result would justify expanding, changing or stopping the pilot.
Device choice follows from the question. If sleep timing is the signal, pick devices whose cloud APIs expose sleep stages and timestamps reliably, and test them on the patients' own phones before enrollment. Patients who already own a supported watch may join more readily than those asked to wear something new.
Equity deserves an early look too. Patients without recent smartphones, stable data plans or the money for a wearable can be left out of a device program by accident. Some practices lend devices, and the program should still work for patients who share only questionnaire data.
Where do mental health device programs go wrong?
Mental health device programs go wrong when the data is collected before anyone has decided how it will change care. The failure points below show up in pilots far more often than technical faults.
- Data without a decision: Signals are gathered, but no clinician has agreed what change would prompt what action.
- Population thresholds: Alerts fire for people whose normal sleep or activity simply differs from the average.
- Silent gaps: A dead watch battery looks the same as a patient who has stopped moving, unless missing data is shown as missing.
- Trust erosion: Patients who feel watched stop wearing the device or uninstall the app, and the program loses the people it most wanted to reach.
- Workload creep: Reviewing device data adds unpaid time unless it is built into session preparation or billable management work.
How Agnotic can help
Agnotic Technologies builds mental health apps and clinician tools that use device data responsibly. We integrate wearables and phone health stores, compute per-patient baselines, build the clinician trend view and connect it to the EHR, and design consent controls that let patients choose each signal.
For programs using prescription digital therapy devices, Agnotic builds the onboarding, usage review and time tracking that monthly management and billing depend on.
Design a device-informed mental health program
See our mental health app development work, or browse the wearables and devices hub.
Related reading on Agnotic
Read about mental health app builds, mental health telehealth, wearable biosensors, personalized RPM baselines and wearable technology in care.
