Evidence
Transparent reporting of our methodology, validation status,
and clinical evidence across disease domains.
Methodology
Our approach to building and validating virtual twin models.
Data Collection
Multi-modal data integration from lab results, wearables, questionnaires, and EHR systems.
- Standardized biomarker panels with 500+ analytes
- Continuous biometric monitoring via validated devices
- Validated patient-reported outcome measures
- HL7 FHIR integration for EHR data
Model Architecture
Ensemble of domain-specific models trained on longitudinal health data.
- Disease-specific risk models with interpretable outputs
- Trajectory prediction using temporal deep learning
- Causal inference for intervention modeling
- Uncertainty quantification for clinical decision support
Longitudinal Design
Repeated measurements over time to capture health dynamics and intervention response.
- Minimum 3-month baseline period recommended
- Quarterly reassessment for stable populations
- Adaptive sampling for high-risk individuals
- Retention optimization through engagement modeling
Validation by domain
Honest assessment of where we are in the validation journey for each disease area.
Validation status reflects current evidence level. We commit to transparent reporting and will update this table as new evidence emerges. Results from research and pilot studies may not generalize to all populations.
Peer-Reviewed Research
Multi-modal virtual twins for cardiovascular risk prediction: A prospective cohort study
Longitudinal biomarker integration for metabolic disease trajectory modeling
Privacy-preserving federated learning for multi-institutional health modeling
Patient-reported outcomes as virtual twin inputs: Validation framework
Full text available on request for research partners. Contact research@biotwin.ai for access.
Questions about our evidence?
Our research team is available to discuss methodology and validation details.
Contact research team