DELIKT — DELIrium risK Tool

What this tool is and how to interpret its evidence

DELIKT is a points-based model for predicting incident delirium in older hospital patients outside intensive care. It was designed for potential automation from information recorded during the first 24 hours of admission. Candidate electronic variables were reduced statistically and converted into a score from 0 to 98. In the source cohort, a score above 20 was associated with later delirium and provided sensitivity of 79.7% and specificity of 62.3%. The authors note that a service could choose a different threshold depending on whether sensitivity or specificity is more important. Such a change requires local evaluation rather than assuming unchanged performance.

Schulthess-Lisibach and colleagues used retrospective data from more than 10,000 admissions at one Swiss hospital. Delirium outcomes drew on routine DOSS or CAM documentation, and the Clinical Risk Assessment Scale contributed to the model. The study included development and internal validation, but no independent hospital was used for external validation. Missing or inconsistent routine screening could therefore influence both model building and reported accuracy.

DELIKT is a promising implementation-oriented risk score, not a diagnostic test for delirium already present. Its exact electronic definitions and points should be taken from the original publication and supplementary material, not reconstructed from this commentary. External validation, recalibration and impact testing are needed before broad adoption. A high score should trigger prevention and systematic assessment; a low score cannot safely exclude a later acute change.