I95.8
Conditions
Interventions
Group 1: Retrospective analysis of routine clinical data from approximately 270,000 patients (= 18 years, non-cardiac surgery with anaesthesiological involvement = 30 min) at Charité – Universitätsmed
Sponsors
Charité - Universitätsmedizin Berlin, Klinik für Anästhesiologie und Intensivmedizin, CBF
Eligibility
Sex/Gender
All
Age
18 Years to No maximum
Inclusion criteria
Inclusion criteria: Non-cardiac surgical procedure with anaesthesiological involvement at Charité – Universitätsmedizin Berlin between 01.01.2010 and 31.01.2026; anaesthesia duration = 30 minutes. For the UI development component: clinical staff (physicians).
Exclusion criteria
Exclusion criteria: For the retrospective data analysis: no exclusion criteria defined. For the UI development component: refusal to participate.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Predictive performance (AUROC) of the algorithm with regard to intraoperative hypotension, determined by comparing AI-based predictions with actual events (gold standard: documented clinical outcomes). | — |
Secondary
| Measure | Time frame |
|---|---|
| - Occurrence of perioperative and postoperative complications within a defined observation period up to postoperative day 7 or until discharge (primary follow-up), including in-hospital mortality, acute kidney injury (AKI), myocardial injury, major adverse cardiovascular events (MACE), and strokes; supplementary data collection up to day 30 within the Charité care setting. - Postoperative delirium, recorded as clinically documented delirium based on routine documentation; analyses are stratified by time, taking into account established delirium screening practices. - Correlation between predictive model output and documented intraoperative clinical interventions to prevent hypotension, as an exploratory endpoint to assess the algorithm’s potential clinical utility (actionability) - Retrospectively determined lead time between the first positive predictive model output and the actual occurrence of intraoperative hypotension. - Determination of the frequency and duration of intraoperative hypotension/perfusion events during the respective duration of surgery or anesthesia, aggregated over the entire retrospective study period. - Comparison of postoperative vital sign trends up to hospital discharge or up to postoperative day 7, stratified by the occurrence and severity of intraoperative hypotension (MAD < 60 mmHg). - Requirements analysis and conceptual development of a prototype user interface (UI) for an AI-powered CDSS, based on the results of the retrospective model analyses | — |
Countries
Germany
Contacts
Public ContactSascha Treskatsch
Charité - Universitätsmedizin Berlin, Klinik für Anästhesiologie und Intensivmedizin, CBF
Outcome results
None listed