MACE (5-point MACE) encompasses all diagnoses and procedures related to major cardiovascular events, including myocardial infarction, stroke, cardiovascular death, rehospitalization for cardiovascular causes, and revascularization or other cardiovascular interventions. Myocardial infarction (ICD-10-GM): Acute myocardial infarction (I21.0–I21.9), recurrent myocardial infarction (I22.0–I22.9) Stroke (ICD-10-GM): Ischemic stroke (I63.0–I63.9), hemorrhagic stroke (I61.0–I61.9), subarachnoid hemorrha
Conditions
Interventions
Group 1: As part of the project, retrospective routine data from medical records at the designated Deutsches Herzzentrum der Charité (DHZC) will be analyzed in digital form from January 1, 2007, to De
Sponsors
Charité – Universitätsmedizin Berlin, Deutsches Herzzentrum der Charité
Eligibility
Sex/Gender
All
Age
18 Years to No maximum
Inclusion criteria
Inclusion criteria: - Patients with complete, electronically documented treatment data after admission to the emergency room/inpatient or intensive care unit in the cardiology ward, depending on the diagnosis, including routine data, vital sign monitor data, and laboratory data from the PDMS. - Complete, electronically documented treatment data for patients from the PDMS/hospital information system (HIS). - Data from cost and revenue accounting from Charité's controlling department is available, cf. HIS, §21 routine data records.
Exclusion criteria
Exclusion criteria: - Purely administrative records - Missing or invalid timestamps necessary for treatment history, temporally inconsistent sequences (e.g., death before encounter) - Insufficient temporal context to support outcome labeling (e.g., no recorded MACE)
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The occurrence of a medically documented major adverse cardiovascular event (MACE) or acute clinical exacerbation, a situation requiring immediate care (e.g., visit to the emergency room or hospitalization) within 1, 5, 10, and 30 days after the index contact (e.g., visit to the emergency room) or the index admission (e.g., hospitalization). Measurement period: January 1, 2007, to December 31, 2024 | — |
Secondary
| Measure | Time frame |
|---|---|
| The measurement period for all secondary end goals is also January 1, 2007, to December 31, 2024: 1) Dynamic inclusion and exclusion of the various machine learning algorithms (AML) in the dynamic knowledge graphs based on evaluation of their assessment metrics in order to limit computing capacity. 2) Inclusion of link predictions that reveal latent connections between exposure and health, such as previously unknown correlations between fine dust peaks and cardiac arrhythmias. Identification of additional variables for diagnostic confirmation (routinely collected laboratory values, clinical parameters, and vital monitor parameters, as well as other relevant health data) that are recorded and stored in the patient data management system (PDMS)/hospital information system (HIS) as part of routine care in order to improve MLA endpoints from secondary endpoint 1. 3) Use of MLA as defined in 1 and 2 to identify newly occurring or progressive cardiovascular diseases, time to diagnosis and therapy, patient outcomes, and comparative evaluation of case costs and revenues related to hospital stays. 4) Use of the MLA in the sense of 1 and 2 to identify performance indicators of routine clinical processes (doctor contacts, type of contact, diagnostics, change of medication, interventional or surgical procedures) of newly occurring or progressive serious cardiovascular events and comparative evaluation. The evaluation assesses how adequately and promptly the healthcare system has responded in order to identify gaps in care or potential for optimization in routine clinical processes. 5) Systematic evaluation of the explainability of the MLA and its predictions. Counterfactual simulation to consider alternative exposure pathways and implications for clinical guidelines for assessing the impact of public health measures (e.g., emission controls). This is to ensure that the predictions generated are not only technically correct, but also comprehensible and clinically inter | — |
Countries
Germany
Contacts
Public ContactLaura Spies
Charité – Universitätsmedizin Berlin, Friede Springer Cardiovascular Prevention Center
Outcome results
None listed