Common laboratory tests predict imminent death in ward patients.

Area of Science:

• Clinical Medicine
• Medical Informatics
• Biostatistics

Background:

• Predicting imminent death in ward patients is crucial for timely interventions and resource allocation.
• Existing prediction models may not fully utilize routinely collected laboratory data.

Purpose of the Study:

• To evaluate the predictive capability of common laboratory variables for imminent death in hospitalized patients.
• To assess the added value of laboratory tests and patient age in predicting short-term mortality.

Main Methods:

• Retrospective observational study involving two university-affiliated hospitals.
• Analysis of laboratory test results and patient demographics from a large cohort (42,701 patients) and an external validation cohort (13,137 patients).
• Multivariable logistic regression modeling to predict imminent death (within the same or next calendar day).

Main Results:

• A multivariable model incorporating laboratory variables and age achieved an Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.87 in the inception cohort.
• External validation demonstrated consistent performance with an AUC-ROC of 0.88.
• The model effectively predicted imminent death in ward patients.

Conclusions:

• Routinely measured laboratory tests are valuable predictors of imminent death in hospital inpatients.
• The developed prediction model shows significant potential for clinical application.
• Further prospective studies are warranted to confirm the clinical utility of these predictive findings.