Evaluation of Machine Learning-Based Risk Prediction Models for Diverse (Critical) Medical Events in Real-World Clinical Settings

Status

Active, not recruiting

Phases

Unknown

Study type

Observational

Source

DRKS

Registry ID

DRKS00036685

Enrollment

30000

Registered

2025-04-16

Start date

2025-05-08

Completion date

Unknown

Last updated

2026-02-02

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Blood transfusion N17 F05 E16 L89 R29.6 A40 A41

Interventions

Group 1: Risk predictions are generated by MAIA 1.0 in the application setting for the medical diagnoses/events of sepsis, acute kidney injury, delirium, severe hypoglycemia, falls, pressure ulcers, a

Eligibility

Sex/Gender

All

Age

18 Years to No maximum

Inclusion criteria

Inclusion criteria: 1. Inpatient care 2. Admitted as inpatients after the software installation date

Exclusion criteria

Exclusion criteria: 1. Patients without Tiplu risk predictions during the inpatient stay 2. Incomplete case

Design outcomes

Primary

Measure	Time frame
Predictive performance of the Tiplu risk prediction models measured by AUROC for the edicmal events considered in MAIA 1.0	—

Secondary

Measure	Time frame
Sensitivity and specificity of the Tiplu risk alerts, Proportion of cases with occurrence of the medical event (in-hospital prevalence of the medical event), Proportion of cases with a Tiplu risk alert (in-hospital prevalence of risk alert), Temporal interval between Tiplu risk alert and onset time of the medical event, Comparative analysis of established sepsis scoring systems and the Tiplu sepsis alert with respect to discriminative performance and temporal interval	—

Measure

Time frame

Sensitivity and specificity of the Tiplu risk alerts, Proportion of cases with occurrence of the medical event (in-hospital prevalence of the medical event), Proportion of cases with a Tiplu risk alert (in-hospital prevalence of risk alert), Temporal interval between Tiplu risk alert and onset time of the medical event, Comparative analysis of established sepsis scoring systems and the Tiplu sepsis alert with respect to discriminative performance and temporal interval

—

Countries

Germany

Contacts

Public ContactDirk Schädler

Universitätsklinikum Schleswig-Holstein, Campus Kiel

Dirk.Schaedler@uksh.de0431 500-20801

Outcome results

None listed

Source: DRKS (via WHO ICTRP) · Data processed: Feb 7, 2026