Where should electronic records for patients be stored?

Area of Science:

• Health Informatics
• Computer Science
• Health Services Research

Background:

• A nationwide health information infrastructure (NHII) is essential for modern healthcare.
• Two primary architectural models exist: distributed (institution-centric) and centralized (patient-centric).
• Limited data exists on the performance implications of these architectural choices.

Purpose of the Study:

• To identify the optimal architecture for efficient and comprehensive nationwide health information exchange.
• To evaluate large-scale operational characteristics of distributed versus centralized models.

Main Methods:

• Simulation modeling was employed to assess architectural performance.
• Key performance markers included transaction volume and data retrieval failure rates.
• Factors studied were data fragmentation and probability of patient encounter.

Main Results:

• Data fragmentation and encounter probability directly correlate with transaction volume.
• The distributed model exhibited significantly higher transaction volumes with >4 data nodes (p<0.0001).
• Data retrieval failures increased with fragmentation, significantly higher in the distributed model with ≥2 data nodes (p<0.0059).

Conclusions:

• Simulation indicates the distributed model scales poorly regarding data availability and integrity.
• The centralized model demonstrated a lower failure rate compared to the distributed model.
• Architectural choice significantly impacts the efficiency, usability, and effectiveness of the NHII at the point of care.