
1. The Product
Kinesis, Speech Vitals, DCR, “Move Together”
- AI powered clinical intelligence tools for remote patient monitoring and cognitive assessments
- Biomarkers:
- Speech
- Eye tracking
- Mobility — gait, balance, and fall risk
- Motor movement
- Hearing
- Integration capabilities: Speech assessment tools, RPM devices, and EHR systems
- Target users: Healthcare providers, clinical researchers, and health system administrators, Enterprise Customers
- Scale: 500K+ patient lives across multiple health systems and 300+ clinical specialists
2. Context & Functionalities Introduced
- Problem:
- Clinicians overwhelmed by RPM data volume, leading to delayed interventions and missed critical health events
- Multiple medical devices meant that interpretation of results combined with previous EHR data was difficult for Providers
- The introduction of more medical devices via acquisitions increased the complexity of data interpretation and care optimization
- Solution
- Built LLM-powered clinical support tool that reads and contextualizes RPM reports, lab results, and clinical notes
- Implemented predictive analytics engine using cognitive assessment data to detect intervention opportunities
- Developed automated risk stratification with intelligent alert system for clinical prioritization
- Created a centralized ML data platform enabling real-time, high-fidelity data access for providers and researchers
- How
- Working with Clinical teams, Data Science (AI/ML Engineers) Design, Engineering, & Regulatory
- Taking proprietary data, session data, and our proprietary clinical decision support tree and fine tuning it over OpenAI's commercial api (Open AI o3)
- Model performance tracking - monitoring accuracy, hallucination detection, and response consistency across clinical scenarios & error handling
- weekly synchs with data science, clinical teams & engineering to review outputs
- Data monitoring - ensuring training data adherence and patient privacy compliance
- System Health & Reliability (AWS Suite)
- API Response Times
- Medical device
- EHR
- Open AI
- System Uptime/Availability
- Error Rates (by error type and severity)
- Data Pipeline Health (ingestion success rates)
- EHR Integration Status (connection health, sync delays)
- Concurrent User Load (peak usage patterns)
- Decision Latency (time from input to next steps in platform)
3. Impact
- Diagnostic accuracy improvement: 31% increase in accuracy through AI-enhanced clinical decision support
- Care delivery efficiency: increasing proactive care interventions by 27%.
- Provider productivity: Enabled management of larger patient populations through intelligent automation
- Population health scale: Successfully deployed across a health system customer managing 500K+ lives
Demo:
smart-care-dashboard
Untitled