Photo Credit: Monsitj

A machine learning algorithm can identify patients with common variable immunodeficiency disease (CVID) from their electronic health records, according to a study published in the May 1 issue of Science Translational Medicine.

Due to the low prevalence and extensive heterogeneity in CVID phenotypes, resulting in delayed diagnoses and treatments, Ruth Johnson, Ph.D., from the University of California in Los Angeles, and colleagues presented a machine learning algorithm (PheNet) to identify patients with CVID from their electronic health records.

The researchers note that PheNet learns phenotypic patterns from patients with CVID and uses this information to rank patients according to their likelihood of having CVID. More than half of the patients with CVID could have been diagnosed one or more years earlier than they were using PheNet. When applied to a large electronic health record database, 74 percent of the top 100 patients ranked by PheNet were highly probable to have CVID. PheNet was validated externally using more than 6 million records from disparate medical systems in California and Tennessee.

“Artificial intelligence approaches like PheNet can be used to expedite the referral of undiagnosed patients to immunologists,” the authors write. “In the future, we will recruit patients identified by this algorithm to our immunology clinics. The impact of our work will benefit the rare disease community as there is an urgent need to identify patients early and efficiently.”

Several authors disclosed ties to the biotechnology industry.

Abstract/Full Text (subscription or payment may be required)

Copyright © 2024 HealthDay. All rights reserved.