Changes to International Classification of Diseases, Tenth Revision (ICD-10) coding to introduce myocardial infarction (MI) subtypes in 2017 have resulted in a decrease in ICD-10 codes attributed to type 1 MI and changes in the characteristics and treatment patterns of these patients, say researchers.
Published online Monday and in the Sept. 21 issue of the Journal of the American College of Cardiology, the study reported that the introduction of ICD-10 codes for type 2 MI (T2MI) and types 3-5 MI was associated with fewer hospitalizations for codes now attributed to type 1 MI (T1MI) and changes in the demographics, characteristics and management of these patients.
“We found that the introduction of the new ICD-10 codes was associated with a 5% reduction in primary diagnosis hospitalizations for patients with ICD-10 codes now designated as T1MI among Medicare beneficiaries,” said the researchers, led by Cian P. McCarthy, MB, BCh, BAO, from Massachusetts General Hospital, Boston.
McCarthy and colleagues said the differentiation of the subtypes of MI and myocardial injury in administrative claims data is important for several reasons – most importantly because T1MI has established evidence-based therapies from randomized clinical trials and international guidelines to guide treatment, while the management of T2MI and myocardial injury is less certain.
“Consequently, patients with T1MI (but not T2MI) are subject to clinical performance and quality measures,” they noted.
“Furthermore, owing to concerns about unsustainable costs and mediocre outcomes in the United States, several value-based programs have been established to improve healthcare delivery by linking Medicare payment to the quality of care provided. Acute MI has been prominent among value-based programs, and T1MI has been the intended target.”
The authors set out to determine whether the introduction of ICD-10 codes for type 2 and types 3-5 MI was associated with changes in hospitalizations for ICD-10 codes attributed to T1MI – using the Nationwide Readmissions Database to identify patients with ICD-10 codes now attributed to T1MI between January 2016 and December 2018. Patients were then stratified according to the timing of their event in relation to the introduction of the type 2 and types 3-5 MI codes on October 1, 2017.
McCarthy and colleagues reported that there were 2,680,323 hospitalizations for ICD-10 codes attributed to T1MI – noting that after adjustment for seasonality, there was a 13.7% decline in monthly T1MI hospitalizations after the introduction of the new subtype codes.
The analysis showed that patients with ICD-10 codes now attributed to T1MI after the coding change were less likely to be female and had lower prevalence of several comorbidities including heart failure, valvular heart disease, hypertension, diabetes mellitus, sepsis, anemia, obesity, depression and cancer, while they had higher rates of coronary angiography and revascularization.
Furthermore, after introduction of the new codes, there was a positive deflection in the slope of risk-adjusted in-hospital mortality (0.007%; P < 0.001) and a negative deflection in risk-adjusted 30-day readmission (-0.002%; P = 0.05) for patients with ICD-10 codes now attributed to T1MI, they said.
The team said the shifts in ICD-10 coding for acute MI could have important implications for whether hospitals are penalized or rewarded based on their performance, adding that the decline in hospitalizations for T1MI ICD-10 codes associated with the introduction of type 2 and types 3-5 MI subtypes should also be considered when conducting observational research on national trends in T1MI.
“Accordingly, it is possible that the introduction of the T2MI and types 3-5 MI codes is refining and more appropriately identifying other patients with T1MI using administrative codes,” they added.
“Our results suggest at least the possibility of better accuracy of administrative data through refinements and additions of new codes.”
McCarthy and colleagues said they, therefore, advocate for the continued refinement of cardiovascular ICD-10 codes.
“However, when new codes are introduced or refined, the potential impact of these changes on other coded conditions and the downstream effects need to be considered,” they noted.
“In this case, the new codes for T2MI and types 3-5 MI changed the meaning of codes now attributed to T1MI, causing difficulties in interpreting the longitudinal analyses of MI through claims data.”
Writing in an accompanying editorial, Andrew P. DeFilippis, MD, MSc, from Vanderbilt University Medical Center, Nashville, and Michael E. Hall, MD, MSc, from the University of Mississippi Medical Center, Jackson, said that the correct identification and differentiation of T1MI from all other types of myocardial injury, specifically T2MI, is of tremendous clinical significance – adding that while T1MI is highly prevalent and potentially deadly, it has effective, time-sensitive, evidence- based treatments, while no therapeutic guidelines are established for T2MI.
They noted that the findings from the new study underscore the need for precise coding of MI subtypes and the impact that “lumping” of MI subtypes can have on databases that use ICD codes to identify patient populations.
“These data suggest that how MI is coded (eg, T1MI vs T2MI) can have a significant impact on these performance and quality measures, and they raise the question of whether these measures are more reflective of differences in coding practices than of the quality of care delivered,” they said, noting that the study reported small differences in hospital length of stay, in-hospital mortality, and 30-day readmission for patients who received an ICD code for acute MI before versus after ICD code changes in 2017.
The expert commentators said such a finding that significant implications for clinical performance and quality measures and several value-based programs, including those that link Medicare payment to the quality of care for T1MI.
Furthermore, they proposed that regulatory agencies like the U.S. Food and Drug Administration and the U.S. Centers for Disease Control and Prevention follow the lead from the ICD and require such detailed MI characterization where MI reporting is required or disseminated, noting that such reporting will “undoubtedly” advance the understanding of the epidemiology of MI events, including associated risk factors, and allow for better prediction, prevention and management.
McCarthy CP, Kolte D, Kennedy KF, et al. Hospitalizations and Outcomes of T1MI Observed Before and After the Introduction of MI Subtype Codes. J Am Coll Cardiol 2021;78:1245-1253.
DeFilippis AP, Hall ME. Impact of New ICD Codes on Acute MI Characteristics and Outcomes: What You Call It Matters. J Am Coll Cardiol 2021;78:1254-1256.
Image Credit: Paolese – stock.adobe.com