A study in regards of the death rate of COVID-19 patients initiated by researchers at Friedrich-Alexander-Universität Erlangen-Nürnberg and University Hospital Erlangen concluded that the COVID-19 mortality declined over the course of the pandemic. Specifically, an analysis across 14 university hospitals showed a decline in the average mortality from an initial 20.7 percent (January through April) to 12.7 percent (May through September). In addition, the dependence of mortality on various risk factors and in combination with various comorbidities was investigated. These factors included ventilatory status, gender, age, and length of stay in the ICU. The risk factors found, such as male sex, older age, and comorbidity pattern, support the results published in previous publications.
As with the previous publication focusing on the lockdown impact on hospital admissions, this study benefited from the efforts of the German medical informatics initiative (MII). The basis for the analysis was the MII infrastructure, which is currently being rolled out as part of the NUM (Netzwerk Universitätsmedizin) project CODEX (COVID-10 Data Exchange Platform) as the basis for a research data platform at the participating partner sites.
For the study, data from the data integration centers were analyzed in a federated approach, meaning that analysis methods were applied to distributed data. The data used for this purpose originated from the five basic modules of the MII core data set (person, demographics, case data, diagnoses, and procedures) and were provided in the standardized format of the §21 data set. After integrating the data into i2b2 research databases and applying local preprocessing, the aggregated analysis data was centrally collected, analyzed and visualized using R at Erlangen University Hospital.
The complete manuscript “Letalität von Patienten mit COVID-19: Untersuchungen zu Ursachen und Dynamik an deutschen Universitätsklinika” is now available online (in German)!