Laboratory Results in Ouest Data Hub - Standardization and Data Quality Overview

Area of Science:

• Health Informatics
• Clinical Data Management
• Laboratory Data Interoperability

Background:

• Clinical Data Warehouses (CDWs) aggregate extensive laboratory data, necessitating interoperability between institutions.
• The Logical Observation Identifiers Names and Codes (LOINC) standard aims to unify laboratory result identifiers, but hospital-level variations persist.
• Interoperability challenges hinder the seamless exchange and analysis of clinical data across healthcare networks.

Purpose of the Study:

• To assess the heterogeneity of laboratory data within Clinical Data Warehouses (CDWs) across multiple hospitals.
• To evaluate the effectiveness of LOINC standardization in ensuring data consistency for clinical decision support.
• To identify specific areas of data variation, such as creatinine measurements, impacting renal function monitoring.

Main Methods:

• Analysis of laboratory data from six hospitals in western France.
• Identification and comparison of LOINC codes used across participating centers.
• Examination of creatinine measurement data to detect phenotypic variations and patient profile similarities/differences.

Main Results:

• A significant heterogeneity was observed, with 910 distinct LOINC codes identified across the six hospitals, only 51 of which were shared universally.
• Phenotypic differences in creatinine measurements were noted, with distinct patterns observed in patients from ICO compared to larger centers (Tours, Rennes, Nantes) and smaller centers (Brest, Angers).
• Larger hospitals demonstrated more consistent creatinine values, while smaller hospitals exhibited similar patient profiles.

Conclusions:

• Despite LOINC standardization, substantial data heterogeneity exists within and between hospitals.
• Quality control and value standardization are critical for ensuring the reliability of clinical decision models.
• Addressing data inconsistencies is paramount for effective multi-center clinical data analysis and application.