TMS · Relapse detection
Detect TMS relapse before the patient disappears.
TMS produces strong acute response, but durability is the hard part. Roughly 37.5% of responders lose response within 12 months, and in standard care the first sign is a missed appointment, not a clinical signal. Early relapse detection turns that blind spot into an intervention window.
Why TMS relapse goes undetected
After the acute course, patients enter maintenance and visits get further apart. Between them, the only common monitoring tool is the PHQ-9 by email, which 60 to 70% of patients never complete. The patients least likely to respond to that email are the ones who are relapsing, because reduced motivation and executive function are themselves symptoms. So the monitoring system fails precisely where it matters most.
37.5%
of TMS responders relapse within 12 months
Dunner 2014, 42 US sites
60%
of relapses go undetected until the patient drops off
48h
earlier warning from passive signal vs. symptom report
How Emobot detects TMS relapse early
Continuous, passive signal
Facial expression, vocal biomarkers, activity, and digital behavior are measured every day, with no patient effort after a 3-minute install.
Objective trend, not recall
You see a continuous wellness score that correlates with MADRS at r=0.89, instead of a single self-reported snapshot every six to eight weeks.
48-hour lead time
Deterioration in the digital signal precedes clinically significant symptom worsening by roughly 48 hours, creating time to re-engage the patient.
In-app nudge to rebook
When the signal shifts, the patient is prompted inside their own app to reconnect with your clinic before they fall out of care.
The evidence
- Validated against MADRS at r=0.89 and PHQ-9 at r=0.83 across 11 clinical studies.
- 75% of patients activate when their physician recommends it, versus 30 to 40% completion for PHQ-9 by email.
- Facial analysis runs on-device; raw video and audio never leave the phone.
Frequently asked questions
What is the TMS relapse rate?
In the largest US naturalistic dataset, about 37.5% of TMS responders relapse within 12 months (Dunner 2014). Relapse risk is highest in the first several months after the acute course, which is why continuous monitoring during maintenance matters.
How do you detect TMS relapse early?
By monitoring objective behavioral signals continuously between visits rather than relying on periodic questionnaires. Passive, multimodal monitoring flags deterioration roughly 48 hours before a patient would endorse worsening on a standardized scale.
Does relapse detection require the patient to do anything?
No. After a one-time 3-minute setup, monitoring is completely passive. There are no surveys or daily check-ins, which is why activation reaches 75% when a physician recommends it.
Go deeper in the full guide: TMS Patient Monitoring: The Complete Guide
See it on a real patient case.
A 30-minute demo walks through the dashboard and how TMS patients show up in the data between visits.