International comparison of mortality following AMI using hospital administrative data – the example of Germany and the USA
Epidemiology and Medical Biometry/Statistics
Final Report Abstract
Mortality for myocardial infarction is widely used in international comparisons as an indicator of health system performance that allows conclusions to be drawn about the effectiveness of care processes. However, different health care structures, reimbursement rules, and other non-qualityrelated factors can affect the international comparability of such indicators. One possible reason for discrepancies could be different recording of treatment cases with a short length of stay due to early death after arrival at the hospital. This seems relevant regarding acute myocardial infarction, where the risk of death within the first 24 hours after onset of symptoms is high. Therefore, this project investigated, by using the case of Germany and the United States, whether mortality of acute myocardial infarction measured in United States administrative hospital data could be biased by underreporting of early deaths compared to mortality in German administrative hospital data. For Germany, the DRG statistics were used to analyze data on all acute inpatient cases in hospitals billing under the DRG system. For the United States, weighted samples of acute inpatient cases (National Inpatient Sample, NIS) and cases treated in hospital emergency departments (Nationwide Emergency Department Sample, NEDS) were extrapolated to the total United States population. The years considered were 2014 to 2019. In-hospital mortality was calculated according to the OECD indicator definition (indicator AC2: 30-day hospital mortality for acute myocardial infarction [in the same hospital] using unlinked data [admission-based]) based on inpatient data first. Then, hospital mortality in the United States was calculated again, considering emergency department deaths that were not included in the inpatient data. In both, Germany and the United States, approximately one-quarter of acute myocardial infarction deaths occurred within 1 day of hospital admission or emergency department arrival. Age- and sexstandardized hospital mortality was 1.6 times higher in Germany than in the United States (2019: 7.3% vs. 4.6%). After including deaths from U.S. emergency department data, this ratio decreased to 1.4 (2019: 7.3% vs. 5.2%). While short-duration treatments due to early death are generally recorded in German inpatient data, in United States inpatient data those cases are partially missing. Potential differences in the recording of early deaths, which may also exist in other countries, should be considered in international comparisons of acute myocardial infarction mortality.
Publications
-
Analyse der Zeitangaben in Krankenhausabrechnungsdaten bei ST-Streckenhebungs-Herzinfarkt und Linksherzkatheterintervention. Das Gesundheitswesen, 83(S 02), S122-S129.
Nimptsch, Ulrike & Busse, Reinhard
-
Möglichkeiten und methodische Grenzen der Nutzung von Krankenhausabrechnungsdaten in der Herzinfarktforschung. Vortrag beim Interdisziplinären Symposium des Berlin-Brandenburger Herzinfarktregisters (B2HIR). Berlin, November 2021.
Nimptsch U.
-
Accounting for early death recording in acute myocardial infarction mortality indicators based on administrative data – Observational study comparing Germany and the United States. Cold Spring Harbor Laboratory.
Nimptsch, Ulrike; Mansky, Thomas & Busse, Reinhard
-
Herzinfarkt-Mortalität in Deutschland – Was ergeben die Routinedaten? Vortrag bei der 89. Jahrestagung der Deutschen Gesellschaft für Kardiologie. Mannheim, April 2023.
Nimptsch U.
-
Recording early deaths following emergency department visits in inpatient data: An observational study using data of 16 German hospitals. Zeitschrift für Evidenz, Fortbildung und Qualität im Gesundheitswesen, 177, 35-40.
Nimptsch, Ulrike; Busse, Reinhard; Möckel, Martin; Fischer-Rosinský, Antje; Slagman, Anna; Keil, Thomas; King, Ryan; Reinhold, Thomas; Roll, Stephanie; Baier, Natalie & Henschke, Cornelia
