In mental health care, remote monitoring usually means questionnaires that patients complete between visits, sometimes alongside app or sensor data. Medicare's remote monitoring codes are narrower. They require a connected device that meets the FDA definition of a device. Sending a questionnaire link does not by itself make a service billable as remote monitoring.
What remote monitoring looks like in mental health
Most practices start with scheduled questionnaires sent by email or text. The PHQ-9 and GAD-7 describe depression and anxiety symptoms. The PHQ-2 and GAD-2 are their two-item screening versions. All four ask about the past two weeks.
Other methods collect different data. Ecological momentary assessment sends brief prompts several times a day or week, usually through an app. Passive sensing uses phone or wearable data, such as activity or sleep timing, with the patient's consent. A sensor signal is not a symptom score, and it needs its own evidence before it informs a clinical decision.
None of these methods is continuous monitoring unless someone reviews the data on a defined schedule. Tell patients when results are read and what to do if they need help sooner.
Match the interval to the recall period
Weekly PHQ-9 or GAD-7 administration is a practice choice, not a standard. Each form covers the past two weeks, so weekly results share half their window with the previous one. Two consecutive totals can reflect the same difficult week, which makes the change between them harder to read.
A shorter interval can suit a care phase, such as the weeks after a medication change. Record why, and read consecutive results with the overlap in mind. Otherwise, set the interval from the questionnaire's timeframe, the next decision the result could inform, and patient burden. The assessment scheduling guide lists the fields to decide before a schedule starts.
Decide in advance what size of change prompts review. In a 2004 study of 434 older adults in the IMPACT trial, a five-point PHQ-9 difference was the estimated minimum clinically important change. Treat a change of that size as a review signal in a similar context, not a universal rule for response or worsening. The PHQ-9 tracking guide explains how to compare results.
Set the review and safety plan before enrolling
Name who reviews each result, how soon, and who covers absences. Write down what triggers outreach. A rising total, a missed questionnaire, and an answer about self-harm each need a different response. A missing result is not evidence that symptoms improved.
Any answer above zero on PHQ-9 item 9 needs direct follow-up from a qualified clinician, whatever the total. A questionnaire submitted at night may not be read until the next business day. Patients need to know it is not a way to ask for urgent help.
If you need help now. In the US, Call 988 or text 988. Call 911 if you are in immediate danger. Outside the US, contact your local emergency number or find support in your country.
The between-appointment assessment guide covers the review contract in more detail, including late and incomplete results.
Medicare RPM and RTM billing
Medicare has two remote monitoring code families. Remote physiologic monitoring (RPM) covers physiologic data such as blood pressure, weight, or oxygen saturation. Remote therapeutic monitoring (RTM) allows non-physiologic data, including therapy adherence and therapy response.
The current CMS telehealth and remote monitoring booklet (MLN901705, December 2025) states these requirements:
- RPM, but not RTM, requires an established patient relationship.
- RPM and RTM cannot be billed together.
- Either one, but not both, can be billed alongside chronic care management, transitional care management, behavioral health integration, principal care management, or chronic pain management.
- RPM data must be collected electronically and uploaded automatically to a secure location.
- Auxiliary personnel can furnish the services under the billing practitioner's general supervision.
- The patient must consent at the time RPM services are provided.
Two rulemaking details decide whether questionnaire-based monitoring can fit. CMS adopted RTM in the CY 2022 physician fee schedule final rule. It noted that RTM data may be patient reported or digitally uploaded, while RPM data must be physiologic and digitally uploaded. For both code families, the device must meet the FDA definition of a device.
RTM has device supply codes for monitoring cognitive behavioral therapy. CPT 98978 covers 16 to 30 days of data in a 30-day period, and 98986 covers 2 to 15 days. The CY 2026 final rule keeps both contractor priced, which means each Medicare Administrative Contractor (MAC) sets payment. CMS said the technologies are still evolving. The same rule says the 2-to-15-day and 16-to-30-day device codes are not additive. Bill one code for the days of data in each 30-day period.
Before billing questionnaire-based monitoring as RTM, confirm three things with the MAC and each payer. The technology must qualify as a device. The record must meet the current CPT descriptor and day count. The payer must cover and price the code. Medicaid and commercial rules can differ from Medicare.
Privacy and consent
Explain what you will collect, how often, who reviews it, how long it is kept, and how the patient can stop. Keep message content minimal, because texts and emails can appear on shared devices and lock screens.
A vendor that creates, receives, maintains, or transmits protected health information on your behalf is generally a business associate. One example in HHS business associate guidance is a health app developer that handles a covered entity's patient data for services such as patient messaging. The agreement is required, but it does not show that the service is configured or used safely. The HIPAA assessment workflow guide covers that work.
Start with a small pilot
Begin with one questionnaire, one patient group, and one accountable reviewer. Track delivery, completion, time to review, and the action that followed each result. Results that nobody reviews cost patients effort and give the team nothing, so fix gaps in review before adding patients or measures.
