Key result
ICD-9 codes miss ~70% of adjudicated MIs, driven largely by uncaptured type 2 MI.
Why the study?
The ICD system does not currently classify type 2 myocardial infarction according to the Universal Definition, leading to potential diagnostic discordance.
Cohort (n=1,927)
Effect estimate: Kappa index 0.386 to 0.439
p-value: p=<0.001
Administrative codes undercapture MIs; leaves open the reliability of ICD-based endpoints for research and surveillance.
BACKGROUND: International Classification of Diseases (ICD) coding is the standard diagnostic tool for healthcare management. At present, type 2 myocardial infarction (T2MI) classification by the Universal Definition of Myocardial Infarction (MI) remains ignored in the ICD system. We determined the concordance for the diagnosis of MI using ICD-9 coding vs the Universal Definition. METHODS: Cardiac troponin I (cTnI) was measured by both contemporary (cTnI) and high-sensitivity (hs-cTnI) assays in 1927 consecutive emergency department (ED) patients [Use of TROPonin In Acute coronary syndromes (UTROPIA) cohort] who had cTnI ordered on clinical indication. All patients were adjudicated using both contemporary and hs-cTnI assays. The Kappa index and McNemar test were used to assess concordance between ICD-9 code 410 and type 1 MI (T1MI) and type 2 MI (T2MI). RESULTS: Among the 249 adjudicated MIs using the contemporary cTnI, only 69 (28%) were ICD-coded MIs. Of 180 patients not ICD coded as MI, 34 (19%) were T1MI and 146 (81%) were T2MI. For the ICD-coded MIs, 79% were T1MI and 21% were T2MI. A fair Kappa index, 0.386, and a McNemar difference of 0.0892 (P < 0.001) were found. Among the 207 adjudicated MIs using the hs-cTnI assay, 67 (32%) were ICD coded as MI. Of the 140 patients not ICD coded as MI, 27 (19%) were T1MI and 113 (81%) were T2MI. For the ICD-coded MIs, 85% were T1MI and 15% T2MI. A moderate Kappa index, 0.439, and a McNemar difference of 0.0674 (P < 0.001) were found. CONCLUSIONS: ICD-9-coded MIs captured only a small proportion of adjudicated MIs, primarily from not coding T2MI. Our findings emphasize the need for an ICD code for T2MI.
No takes yet. Share an insight, caveat, or question.
Díaz–Garzón et al. (2016) conducted a cohort in Myocardial Infarction (n=1,927). ICD-9 coding vs. Universal Definition of Myocardial Infarction was evaluated on Concordance for the diagnosis of MI using ICD-9 coding vs the Universal Definition (Kappa index 0.386 to 0.439, p=<0.001). ICD-9 coding captured only 28% to 32% of adjudicated myocardial infarctions, showing fair to moderate concordance (Kappa 0.386-0.439, P<0.001) primarily due to not coding type 2 MI.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: