EHR Data Persistence for a Nationwide Shared Medical Record: An Overview

Abstract

Electronic health records (EHR) are considered the backbone of modern health system. When they are implemented effectively at scale, they hold the promise of enhancing healthcare services' quality and efficiency. However, EHR implementations vary based on factors such as EHR usage scenarios and technological capacities, resulting in varying volumes, complexity, and types of healthcare data. Consequently, the data storage model and query capabilities must align with these variations to achieve desired outcomes. This paper reviews the standardized EHR data persistence based on existing EHR literature and discusses its appropriateness for current healthcare usage scenarios including nationwide access. Insights into relevant literature reveal that NoSQL databases are more appropriate for current EHR usage scenarios. However, choosing among NoSQL databases’ persistence depends on the use case it is intended for in addition to the standards used for data integration between systems. Accordingly, clinical usage scenarios are best served with document stores for standardized EHR following the dual model. For population-based queries, JSON stores provide a good balance between storage capacity, query performance, and flexibility. The graph-based model presents a good solution for nationwide access considering the workload and the interconnected healthcare data nature and structure. One way to respond to the conflicting data persistence requirements is the multi-model database (MMDB) concept in both its forms: polyglot and multi-model databases.