EVIDENCE LIBRARY
ICU Computerized Sedation Orders Needed Revision for Age and Organ Function
A before-and-after study found unexpected prescribing changes after computerized order implementation. Later unadjusted improvements did not establish an independent mortality benefit.
A before-and-after study found unexpected prescribing changes after computerized order implementation. Later unadjusted improvements did not establish an independent mortality benefit.
Three phases of care
The prospective observational project included 279 ventilated ICU patients across no-protocol, computerized-protocol, and revised-protocol phases. The first computerized version was associated with higher fentanyl use.
A revision incorporating age and kidney and liver function was associated with lower fentanyl and lorazepam use and fewer recorded sedation-related complications compared with the first version. Unadjusted ICU mortality also differed.
The adjusted result matters
Neither protocol phase independently predicted ICU or hospital mortality in the multivariable analysis. The observational phases occurred at different times, so changes in patients or practice could confound comparisons.
The study did not measure protocol compliance or delirium, and clinicians could prescribe outside the order pathway. Merely installing a computerized protocol therefore did not show that all treatment followed it.
Practical implications
Review the actual order defaults, revision history, patient-specific adjustments, and administered medication. The paper supports scrutinizing decision-support design and implementation. It does not prove that computerized orders alone improve outcomes or that any unadjusted mortality difference was caused by lorazepam reduction.
Source paper
Original analysis of the supplied journal article, identified as PMC4672574.