Journal of the Louisiana State Medical Society
antipsychotic medication to the resolution of delirium. The distribution of the durations of delirium symptoms after medications were started was right-skewed and non-normally distributed. Patients who were discharged with delirium were censored. Kaplan-Meier survival curves and log-rank tests were used to test differences in time to resolution be- tween groups, with shorter times indicating better clinical outcome. Potential covariates were entered into the mainmodels and tested with Cox proportional hazards regression models.
RESULTS
Patients were divided empirically into two groups based on how they actually re- ceived antipsychotic medication (as opposed to how it was ordered). The Scheduled group received two or more doses of antipsychotic medication per 24 hours once treatment be- gan; the PRN group received on average one or fewer doses during each 24-hour period (which included some 24-hour periods in which the patient did not receive any doses). This method was based on the duration of action and conventional recommendations of using antipsychotic medications. There were 42 patients with complete data to analyze; 31 received scheduled doses, and 11 received PRN doses. Sixteen patients were dischargedwith delirium symptoms that had not resolved and their durations of delirium ranged from 13 to 1,400.5 hours. The demographics of the Scheduled and PRN groups are shown in Table 1. The Sched- uled group included significantlymore white patients (68% vs. 9%) and had more patients with delirium that resolved (74% vs. 27%) compared to the PRN group. The groups did not differ on age, sex, reason for psychiatry consultation, type of delirium, disposition status, or location of consultation. Table 2 shows means and medians for durations of delirium after initiation of anti- psychotic. The median duration of delirium
Figure 1: Kaplan-Meier survival estimates for duration of delirium symptoms in Scheduled (N=31) vs. PRN (N=11) groups.
Figure 2: Kaplan-Meier survival estimates for duration of delirium symptoms in Timely (N=20) vs. Delayed (N=22) groups.
Differences in demographic variables between the Scheduled and PRNgroups were comparedwith chi-square tests for categorical variables and t-test for the continuous age variable. Because some patients were discharged before resolution of their delirium, survival analysis was used, which can handle data when the endpoint of interest has not been reached. This method was used to test durations of delirium between the Scheduled vs. PRN groups and between the Timely vs. Delayed groups. The time to event variable was the number of hours from the first dose of
for the total sample was 129 hours (5 days and 9 hours). The range was 13 to 1,400 hours, with 37% of the sample (n=15) showing delirium symptoms for more than one week. The types of medication that patients received included halo- peridol (n=22), olanzapine (n=10), risperidone (n=4), and quetiapine (n=9). Fourteen percent of the sample showed hypoactive delirium, 67% showed hyperactive delirium, and 19% presented with mixed delirium.
244 J La State Med Soc VOL 166 November/December 2014
Made with FlippingBook - Online catalogs