@article{4820,
  author = {Koodichimma Ibe-Ariwa},
  title = {A Normalized Relational Schema for Hospital Management Systems: Design Principles, Network Centralization, and Utilization Phenotypes},
  journal = {Journal of Information & Systems Management},
  year = {2026},
  volume = {16},
  number = {3},
  doi = {https://doi.org/10.6025/jism/2026/16/3/93-109},
  url = {https://www.dline.info/jism/fulltext/v16n3/jismv16n3_1.pdf},
  abstract = {Background: Healthcare organizations increasingly require information systems that support not only clinical
data storage but also efficient retrieval, integration, and advanced analytics. Poorly designed database
schemas introduce redundancy, update anomalies, and integrity risks that compromise both operational
efficiency and data quality.
Objective: This study presents a normalized relational database schema for hospital management and
demonstrates its capacity to support diverse operational and patient level analyses.
Methods: A fourteen table relational schema was designed applying first, second, and third normal forms
(1NF-3NF) with primary and foreign key constraints to ensure referential integrity across patient
demographics, staff management, departmental organization, physical resources, appointments, clinical
records, surgical events, shift scheduling, and billing. The schema was then evaluated through mediation
analysis, moderation analysis, bipartite patient doctor network analysis, four class Gaussian mixture latent
profile analysis, group comparisons with effect sizes, and robustness/sensitivity checks.
Results: Mediation of payment through appointment completion was not statistically supported (bootstrap
95% CI:- 53.45 to +6.08). Moderation by age and department was non-significant. The patient doctor network
exhibited strong centralization, with one physician accounting for 505 weighted links and 426 unique patients
among 726 patients and 1,000 appointment edges. Four distinct utilization phenotypes emerged.
Appointment modality showed negligible practical effects on payment or completion ( 2 = 0.0012; Cramér's
V = 0.056). Findings were robust across alternative specifications.
Conclusion: The normalized schema functions as an integrated analytical foundation bridging data
management and data services, supporting network analytics, patient phenotyping, and reusable clinical
phenotype development for future hospital information infrastructures.},
}