01
Low completion rates
Digital PROMS and ePRO are often under-completed. Visibility on clinical evolution between visits remains fragmented.
Clinical voice infrastructure
Patient follow-up
should not stay silent.
Voice AI that applies your protocols, triggers structured follow-up, and surfaces actionable signals under medical supervision.
HDS
certified hosting in France
No
patient action required
Call in progress · Day 7 post-op follow-up
M
Marie D. · 68 years old
Orthopedic surgery · Dr. Leroy protocol
In progress
Doctinum Agent
Hello Marie, this is your weekly follow-up call. How would you rate your pain today from 0 to 10?
Patient
I would say 6, maybe 7. Nights are the hardest.
Doctinum Agent
Got it. Were you able to resume light activity as planned in your protocol?
7/10
Reported pain
↗
vs Day 3
⚑
Threshold alert
Live test
Field insight
Some warning signals remain silent.
That is often where risk starts.
02
Manual outreach burden
Coordinators, CRAs, and medical teams spend increasing time on manual outreach and documentation.
03
Unstructured data
Incoming calls, SMS, emails: information arrives in silos, without consistent traceability or usable structure.
04
Digital exclusion
Apps and portals structurally exclude fragile, older, or low-connectivity populations, exactly those needing the most vigilance.
40–50% average lost-to-follow-up in longitudinal clinical studies · The channel was the issue.
Our approach
Active follow-up, triggered by your teams.
You define the protocol. Voice AI executes follow-up. Each exchange is structured and delivered as an actionable signal for the medical team.
How it works
01
You define the protocol
Questions, cadence, alert thresholds, and escalation rules are configured by the medical team.
02
Patient consent captured
Patients explicitly consent to follow-up and voice channel before the first call. Logged and timestamped.
03
Structured calls triggered
The system calls on schedule, handles no-answers, rescheduling, and third-party contacts.
04
Structured data delivered
Responses are extracted, normalized, and delivered as clean data, not raw transcripts.
05
Alerts remain supervised
Alerts follow your thresholds. The medical team remains the sole decision-maker.
What the clinician receives
Summary · Marie D. · Day 7
Reported pain7/10↗ Threshold
Physical activityPartial restartDelay
TreatmentCompliantOK
Sleep qualityDisturbedSignal
Next callDay 10 · Scheduled
No autonomous medical decisions.The team arbitrates.
Field feedback
“What convinced me is that the system calls the patient, not the opposite. For my 70+ post-op patients, this is the only thing that really works.”
Dr. · Orthopedic surgeon · Doctinum pilot
Clinical infrastructure
Four engines. Not just a voice bot.
The difference between a generic voice bot and reliable clinical infrastructure comes from solving engineering problems that most solutions still leave unresolved.
Identity & consent
Patient verification before collection, anti-impersonation, in-call consent capture, and native GDPR traceability.
Native GDPRCall orchestration
Handles no-answers, voicemails, third parties, refusals, rescheduling, and time windows with full auditability.
Audit trailProtocol fidelity
Maintains clinical logic through conversational deviation. Knowledge base configured by clinicians.
Clinician protocolData structuring
Extracts clean clinical data, routes threshold alerts, and exports structured HL7 FHIR payloads.
HL7 FHIRLongitudinal continuity
Clinical memory across calls and trend detection over time to keep a usable trajectory.
Longitudinal follow-upHDS hosting · France
Data hosted in France by HDS-certified providers. Documented GDPR compliance. Supervised organizational and collection tool.
Use cases
Adapted to multiple clinical contexts.
Surgery · Aesthetic medicine · Sports medicine
Pre and post intervention follow-up
Between consultation and procedure, then through following weeks, voice follow-up structures what patients report without requiring teams to manually chase every check-in.
→Track pain, healing, and rehab protocol adherence
→Early detection of self-reported complication signals
→Automatically collect structured PROMS
→Provide concise summary before each follow-up consultation
Sample exchange · Day 5 post-arthroplasty
Doctinum
Hello, this is your daily follow-up. How would you rate your pain out of 10?
Patient
Around 6. Nights are harder.
Doctinum
Were you able to complete your physical therapy exercises today?
Patient
No, the pain was too strong this morning.
Signal sent · Pain ≥ 6 + stopped physical therapy · Dr. Martin threshold
Interactive simulation
Listen to what your patients experience between consultations.
Select a clinical context and trigger a realistic call simulation by a Doctinum agent. In 3 minutes, hear exactly what the patient hears and what your team receives.
Non-clinical simulation for demonstration only. Real protocols are defined and validated by medical teams.
Doctinum Agent
Simulation · Choose context
Security & regulatory framework
HDS hosting · France✦Native GDPR✦Logged consent✦Outside medical-device classification✦Medical supervision✦HDS hosting · France✦Native GDPR✦Logged consent✦Outside medical-device classification✦Medical supervision✦HDS hosting · France✦Native GDPR✦Logged consent✦Outside medical-device classification✦Medical supervision✦HDS hosting · France✦Native GDPR✦Logged consent✦Outside medical-device classification✦Medical supervision✦
Define your first protocol in 20 minutes.
One call to calibrate your use case, protocol, and pilot conditions according to your clinical context.