The Claims Data Learning & Enhancing for Algorithm Refinement (CLEAR) Study: Overview of the Study Design and...

Area of Science:

• Health Informatics
• Epidemiology
• Data Science

Background:

• Medical claims data are crucial for epidemiological studies but often lack sufficient validation for disease identification accuracy.
• Existing validation methods for claims data can be costly and time-consuming.

Purpose of the Study:

• To introduce the Claims data Learning & Enhancing for Algorithm Refinement (CLEAR) Study, a novel database platform in Japan.
• To enable systematic, low-cost validation of algorithms used for identifying diseases in medical claims data.

Main Methods:

• The CLEAR Study platform links patient-level medical claims data with diagnostic data (e.g., lab results, imaging reports) serving as the gold standard.
• Data linkage is achieved using pseudonymized medical record numbers, with personal information protected by research identification numbers.
• The platform's feasibility was demonstrated using data from respiratory syncytial virus (RSV) infections and intussusception cases.

Main Results:

• Eight hospitals joined the CLEAR Study, with three already providing claims data.
• Data were collected for 5,022 RSV infection cases and 1,450 intussusception cases.
• Initial analysis of diagnostic data for tens of thousands of cases confirmed the database's utility for algorithm validation.

Conclusions:

• The CLEAR Study platform facilitates crucial validation of medical claims data in Japan.
• By enhancing the reliability of claims-based research, the platform supports more accurate epidemiological studies and drug risk assessments.