EVIDENCE LIBRARY
Surgical-ICU Delirium Score Included Benzodiazepine Exposure but Was Not a Causation Calculator
A prospective surgical-ICU study identified benzodiazepine use among delirium predictors. Its score estimated group-level risk in the studied setting, rather than the cause of an individual episode.
A prospective surgical-ICU study identified benzodiazepine use among delirium predictors. Its score estimated group-level risk in the studied setting, rather than the cause of an individual episode.
The model’s performance
Among 250 patients, sixty-one developed postoperative delirium. The investigators combined age, organ-failure score, benzodiazepine exposure, diabetes, ventilation, and an informant cognitive measure into a prediction score.
The reported area under the curve was 0.84. At the selected cutoff, sensitivity was about seventy-two percent and specificity about eighty-one percent. These values describe classification performance, not certainty for a particular patient.
Important gaps
The study concerned general surgical ICU patients and did not establish performance in cardiac or neurosurgical populations. Some preoperative and intraoperative information, including previous delirium and medication details, was unavailable.
An exposure can predict an outcome because of illness severity, treatment selection, or a direct effect. Inclusion in a multivariable score does not distinguish all those pathways.
Interpreting an episode
Use the medication timeline alongside cognition, surgery, illness severity, pain, and ventilation. The score does not assign a percentage of responsibility to benzodiazepines or prove that withholding them would have prevented delirium. External performance and individual clinical reasoning remain separate questions.
Source paper
Original analysis of the supplied journal article, identified as PMC6425578.