Retrospective analysis of data from routine clinical practice
Conditions
Interventions
Group 1: The study retrospectively combines anonymized laboratory data (test results) with anonymized discharge diagnoses (ICD-10 codes) for all patients at Klinikum Oldenburg over a 10-year period. T
Sponsors
Universitätsinstitut für klinische Chemie und Laboratoriumsmedizin des Klinikums Oldenburg
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: Patients who underwent at least one laboratory test at Oldenburg University Hospital between 2015 and 2024 and who were diagnosed with at least one condition according to ICD-10.
Exclusion criteria
Exclusion criteria: Patients who, during the relevant period, had either a diagnosis without laboratory findings or laboratory findings without a diagnosis.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The aim of the Study is to develope machine learning (ML) models for the recognition of patterns in laboratory test results that are predictive of specific diagnoses. Furthermore, the performance of the trained ML models will be evaluated by calculating the accuracy of the predictions made by the models per diagnosis (AUC, sensitivity, specificity, calibration). | — |
Secondary
| Measure | Time frame |
|---|---|
| Comparisons of Different AI Models Optimizing Model Performance | — |
Countries
Germany
Contacts
Public ContactGunnar Brandhorst
Universitätsinstitut für klinische Chemie und Laboratoriumsmedizin des Klinikums Oldenburg
Outcome results
None listed