Approach

In the case of rare diseases or exposures, real-world evidence studies may require collaboration across multiple health systems and data sources. To support this work, we spearheaded use of a distributed data network approach, enabling large-scale research while maintaining data privacy and institutional control.

Our approach, developed for the FDA Sentinel Initiative, allows healthcare organizations to maintain physical and operational control over their data within their own secure environments. Rather than pooling patient-level data in a central repository, standardized analytic programs are sent to participating partners, where the programs are executed locally.

To support consistent analysis across diverse data environments, data partners transform their healthcare data into a standardized structure. This harmonization allows analytic programs to be applied consistently across multiple datasets while preserving the local governance of each partner.

The distributed approach enables researchers to study large and diverse populations across healthcare systems while protecting patient privacy and respecting institutional data stewardship. It has become a foundational model for modern pharmacoepidemiology and real-world evidence research and supports the large multi-site studies conducted by our Real-World Evidence Practice.

In addition to collaborating with external data partners through distributed data networks, we have access to important internal and Harvard-affiliated healthcare data resources that support real-world evidence generation.

HPHCI’s parent organization, Point32Health, is a collaboration between two regional health insurers, Harvard Pilgrim Health Care and Tufts Health Plan. These health plans provide access to large administrative claims datasets covering insured populations across the Northeastern United States. Point32Health is one of the largest nonprofit health insurers in the region, covering millions of members across commercial, Medicare, and Medicaid programs. These data provide detailed information on healthcare utilization, diagnoses, procedures, and prescription medications and have supported pharmacoepidemiologic research for decades.

In addition to health plan data, investigators collaborate with clinical partners across Harvard-affiliated healthcare systems. These collaborations enable studies using electronic medical record data from major academic health systems, such as Mass General Brigham and other Harvard-affiliated hospitals and clinical networks. This relationship is strengthened by the many clinicians with joint appointments at Harvard-affiliated hospitals and our Department of Population Medicine.

Together, these resources complement our distributed network collaborations by enabling studies that combine claims data, electronic health records, and multi-site research networks to address complex questions about the safety, effectiveness, and use of medical products in real-world populations.

The HPHCI Real-World Evidence Practice has extensive experience conducting studies using national Medicare and Medicaid data through the Centers for Medicare & Medicaid Services Virtual Research Data Center (VRDC). The VRDC provides secure access to comprehensive Medicare and Medicaid claims data, enabling researchers to study healthcare utilization, medical procedures, and prescription drug use among older adults and other Medicare and Medicaid beneficiaries across the United States.

These data support large-scale pharmacoepidemiologic studies evaluating the safety, effectiveness, and utilization of medical products in populations that are often underrepresented in clinical trials, including older adults and individuals with complex health conditions.