Situation: A large healthcare provider wants to facilitate early prognosis of cardiovascular diseases through increased private data utilization.

Complication: Privacy laws such as HIPAA and confidentiality concerns prevent contributors from sharing data outside of organizational firewalls.

Resolution: XOR enables researchers to privately compute across organizational data sources in order to increase both sample size and patient attributes, leading to improved model performance and disease prognosis.

Outcomes:

• Increased sample size improves predictive model performance

• Sample size increase >100% and growing with additional data contributors

• Incremental feature set enables new predictive insights

Situation: Clinical trials need sufficiently large sample sizes to be viable, so researchers need access to large patient populations in order to recruit candidates that meet the study eligibility criteria. Additionally, EHR data have rich textual features that can be extracted for efficient patient matching.

Complication: EHR data contains sensitive information and must remain HIPAA & GDPR-compliant, making it difficult for researchers to access sufficient patient populations for their trials. Also, descriptive textual features often go unused during the selection process because of their qualitative, language-based nature.

Resolution: Leverage XOR to run advanced matching and analysis across distributed, private EHR repositories for patient selection while maintaining privacy and compliance.

Outcomes:

• HIPAA & data residency compliance is maintained during the computation

• Additional textual features are leveraged in patient selection, increasing overall patient eligibility

• Medical researchers fill trials more quickly, leading to more efficient drug development