Can Neo4j Replace PostgreSQL in Healthcare?

Area of Science:

• Health Informatics
• Database Management
• Big Data Analytics

Background:

• The increasing volume of big data necessitates advanced analytical skills and tools.
• SQL databases like PostgreSQL are traditionally used, while graph databases like Neo4j are often limited to specific data types.
• The MIMIC-III patient database presents a complex dataset for evaluating different database technologies.

Purpose of the Study:

• To compare the performance and complexity of PostgreSQL (SQL) and Neo4j (Cypher) for health data analysis.
• To assess the suitability of graph databases for storing and querying large-scale patient datasets.
• To determine the optimal database solution for the MIMIC-III patient database.

Main Methods:

• A side-by-side comparison of PostgreSQL and Neo4j was conducted.
• The MIMIC-III patient database was used as a case study for evaluation.
• Query complexity and runtime were measured for both database systems.

Main Results:

• Neo4j implementation is more time-intensive compared to PostgreSQL.
• Neo4j queries demonstrated reduced complexity and faster runtimes than comparable PostgreSQL queries.
• The performance difference suggests advantages for Neo4j in specific health data analysis scenarios.

Conclusions:

• While PostgreSQL is adequate for general database needs, Neo4j presents a strong alternative for health data.
• Neo4j should be considered a viable contender for efficient health data storage and analysis.
• The study highlights the potential of graph databases in managing and querying complex biomedical datasets.