January 5, 2017

Heart failure “dashboard”

A new computer algorithm developed at Vanderbilt could save billions of dollars in health care costs by identifying patients at risk for readmission after being hospitalized for heart failure.

by Sanjay Mishra

Millions of patients every year are hospitalized for heart failure, and a fifth of them are readmitted within a month of discharge. Identifying patients at risk for readmission could save billions of dollars annually in health care costs.

In this month’s American Heart Journal, Zachary Cox, PharmD, and colleagues describe the first fully automated, electronic heart failure identification algorithm. It identified more than 3,000 heart-failure admissions to Vanderbilt University Medical Center over 26 months with high accuracy and good specificity.

The algorithm requires three of four identifiers: heart failure diagnosis on admission, International Classification of Disease (ICD-9) codes for heart failure, intravenous diuretic administration, and elevated BNP (a heart failure biomarker).

The researchers are working with VUMC to ensure all patients identified with decompensated (worsening) heart failure by this “dashboard” receive further evaluation and outpatient follow-up to reduce the risk for readmission. The algorithm was less accurate for patients whose initial hospital care began elsewhere and who were missing standard identification data.

Send suggestions for articles to highlight in Aliquots and any other feedback about the column to aliquots@vanderbilt.edu