Date of Award

January 2014

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Shamsuddin Akhtar

Subject Area(s)



Intensive Care Unit (ICU) delirium is associated with higher mortality rates, longer hospital stays, higher rates of cognitive impairment on discharge, and greater hospital cost. Whereas the 2002 Society of Critical Care Medicine (SCCM) Guidelines recommend haloperidol for the treatment of delirium in ICU patients (Recommendation C), the 2013 SCCM guidelines state that atypical antipsychotics might improve the duration of delirium (Recommendation C), without recommending a particular drug. Without specific guidelines for the treatment of ICU delirium or conclusive data in the literature, it is important to know what treatments are used in clinical practice. Prior studies that have tried to determine practice pattern in the management of ICU delirium have predominantly used surveys, which were not designed to determine the patient characteristics impacting the treatment choice. This study aims to describe the current treatment of ICU delirium at various critical care units at Yale-New Haven Hospital and determine independent predictors of receiving a particular atypical antipsychotic for the treatment of ICU delirium.

We hypothesized that the lack of specific guidelines and scarce data regarding treatment of ICU delirium result in unit-specific preferences for delirium treatment with atypical antipsychotics. This single center, retrospective chart review compares the number of new prescriptions for olanzapine, quetiapine, ziprasidone, aripiprazole, and risperidone in the Medical ICU (MICU), Surgical ICU (SICU), and Cardiothoracic ICU (CT-ICU). For each patient receiving a newly prescribed antipsychotic in the mentioned ICUs, the medical record was studied for patient age, gender, race APACHE-II score, and ICU location, and independent predictors of receiving a particular atypical antipsychotic were determined.

The analysis showed that 83.7% of new orders for an atypical antipsychotic in the SICU were for risperidone; 72.2% of new orders for an atypical antipsychotic in the MICU were for olanzapine, and 73.9% of new orders for an atypical antipsychotic in the CT-ICU were for quetiapine (Fisher-Exact Test p < 0.0001). The odds ratio of being a patient in each ICU given that the patient received a particular antipsychotic was calculated after adjusting for severity of illness (via APACHE-II score), age, gender, and race using multivariate logistic regression models. Patients who received risperidone were 42.5 (95% CI 14.9-121.1) times more likely to be in the SICU (p < 0.0001). Patients who received quetiapine were 3.9 (95% CI 1.3-11.3) times more likely to be a patient in the CT-ICU. Patients who received olanzapine were less likely to be in the CT-ICU or SICU (OR: 0.1; 95% CI 0.1-0.4; and OR: 0.1; 95% CI 0.0-0.2, respectively; p < 0.0001).

In conclusion, this study has confirmed previous survey data stating that physician choice for the treatment of ICU delirium is variable and has shown that at this center, the unit to which the patient is admitted has a strong association with the atypical antipsychotic used to treat ICU delirium after adjusting for intrinsic patient differences.


This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.