MISSED DIAGNOSES OF ACUTE MYOCARDIAL INFARCTION IN THE EMERGENCY DEPARTMENT: AN EXPLORATION USING HCUP DATA

Background:  Inaccurate and missed diagnoses, important aspects of patient safety and adverse medical events, are gaining increased prominence.  Both can lead to treatment delays with no, mild, moderate, or serious consequences.  The literature suggests that 2-5% of patients with acute myocardial infarction (AMI) are mistakenly released from the emergency department without having been admitted as inpatients.  Failure to hospitalize may be related to race, gender, and the absence of typical cardiac symptoms.  We were interested to learn how estimates of missed AMI diagnoses from administrative data compare to the literature and how these rates vary across subgroups and hospitals.

Methods:  This cross-sectional analysis of hospital discharge records used data from the 2007 Healthcare Cost and Utilization Project’s (HCUP) State Inpatient Databases (SID) and the State Emergency Department Databases (SEDD) for 9 states (AZ, FL, MA, MO, NH, NY, SC, TN, UT) with reliable person linkages and race/ethnicity data.  These states encompass 23% of the U.S. resident population, 9.6 million inpatient stays, and 23.4 million ED visits.  For study inclusion, patients were 18 years or older and had experienced a first AMI admission between February and December 2007; this criteria captured 115,429 AMI admissions. The key measure was the percentage of patients with an AMI admission who had been seen in the ED 2 to 7 days prior for a cardiac-related symptom or abdominal pain. 

Results:  The 7-day missed diagnosis rate for AMI in this study was 1.85%, with rates across the study states ranging from 0.9% to 3.2% - slightly lower than estimates in published studies.  Rates of missed diagnoses were higher among African-Americans, the uninsured, and low-income communities; higher in hospitals in micropolitan or rural areas, in hospitals with less than 100 beds, in hospitals without cardiac catheterization labs, and in public or for-profit hospitals; and higher on weekends and on “slow” ED days. Rates of missed AMI diagnoses were lower among Hispanics and patients with Medicare; lower in teaching hospitals, hospitals with ≥300 beds, and hospitals with moderate to high occupancy rates; and lowest on “crowded” ED days and during January through February.

Conclusion: Hospital administrative data are a reasonable source for estimating rates of missed AMI diagnoses and show considerable variation across types of hospital and patients.   These administrative data yielded lower estimates of missed diagnoses than rates from the literature.  The 9 states may not be representative of the entire U.S., and these preliminary analyses do not account for correlations among attributes of hospitals.