Coding for Hospitalized Patients with Acute Decompensated Heart Failure May Miss the Diagnosis

Background:

The Joint Commission core measures for hospitalized patients with acute decompensated heart failure (ADHF) include: use of ACE inhibitors for left ventricular systolic dysfunction, discharge instructions with specific information, left ventricle ejection fraction assessment, and tobacco cessation counseling, Evaluation of compliance is based on patients coded with ICD‐9 code 428 as the primary diagnosis. With increasing focus on “value‐based purchasing” or “pay for performance,” accuracy of coding may be essential for optimal future reimbursement. Yet little is known about whether the coding system appropriately identifies all patients admitted with ADHF. We evaluated the accuracy of the ICD‐9 coding system in identifying patients with ADHF both retrospectively and prospectively.

Methods:

We analyzed the electronic records of 79 patients hospitalized with ADHF at an urban academic medical center. First, we randomly selected 20 individual patient charts with the primary ICD‐9 code 428, ADHF, in 2008. A hospitalist, an emergency physician, and a heart failure cardiologist reviewed these charts independently and applied the revised Framingham criteria (RFC) for the diagnosis of heart failure. The RFC have been used since 2003 to identify heart failure events in large epidemiological studies but have not been used to identify ADHF. Second, we identified 59 consecutive patients who were seen by the heart failure cardiologist from August to September 2009 and met the RFC criteria for ADHF. ICD‐9 codes in the patients' electronic records were reviewed after patient discharge.

Results:

For the initial 20 reviewed charts, the interrater agreement for diagnosis between the hospitalist, emergency physician, and cardiologist was 100%. All 3 confirmed that 18 of 20 charts were property coded with ICD‐9 code 428 and that the 2 remaining charts were patients with asymptomatic left ventricular systolic dysfunction and diastolic heart failure. Of the 59 patients prospectively identified as having ADHF by the heart failure cardiologist using RFC, only 30 of 59 were assigned the ICD‐9 code 428 as the primary diagnosis. The 3 most common primary diagnoses for patients not coded primarily as ADHF were “complications peculiar to certain specified procedures” (6 of 29), “cardiac dysrhythmias” (3), “candiomyopathy” (2), “other diseases of endocardium” (2), and “other forms of chronic ischemic heart disease (2).”

Conclusions:

This study demonstrates that the use of ICD‐9 coding may lead to errors in identification of patients with ADHF. Although 3 physicians had 100% agreement on retrospective assessment of ADHF, correct prospective coding occurred for only 50% of patients with a primary diagnosis of ADHF based on physician review. Of note, patients admitted with diastolic heart failure fall outside the Joint Commission guidelines. Attention to coding accuracy will be essential for both quality measurement and future efforts at value‐based purchasing.

Author Disclosure:

J. Barsuk, none; D. Malkenson, none; L. Klein, none; P. Page, none; M. Williams, none.