A decision rule algorithm for the detection of patients with hypertension using claims data

Area of Science:

• Health Informatics
• Epidemiology
• Health Services Research

Background:

• Claims data is valuable for population health studies but often lacks diagnostic information.
• Identifying specific conditions like hypertension within claims data is challenging due to missing diagnosis codes.

Purpose of the Study:

• To develop and validate a decision rule algorithm for detecting hypertension patients using prescription data in claims.
• To leverage large-scale claims data for epidemiological surveillance of hypertension.

Main Methods:

• Retrospective analysis of Iran Health Insurance Organization (IHIO) data (2012-2016).
• Development of a 13-rule algorithm incorporating medication, age, sex, and physician specialty.
• Validation using sensitivity, specificity, PPV, NPV, and accuracy metrics.

Main Results:

• The algorithm identified 76.89% of patients receiving antihypertensive medication as having hypertension.
• High sensitivity (100%) and negative predictive value (100%) were observed.
• Overall accuracy was 76.50%, with a specificity of 48.91%.

Conclusions:

• The developed algorithm demonstrates robust performance for hypertension detection in claims data.
• This approach facilitates large-scale hypertension surveillance and can inform public health strategies.
• The algorithm aids policymakers and researchers in improving personalized care quality.