Weekend Prescribing Practices and Subsequent Seclusion and Restraint in a Psychiatric Inpatient Setting

This study was a case-control study based on review of existing medical records. The records of all patients admitted to the adult and geriatric floors at Beth Israel Medical Center in New York City between 5 p.m. Friday and 8 a.m. Saturday from July 2001 to June 2006 were screened to identify patients who experienced a subsequent seclusion or restraint event during their hospital stay. These 39 patients made up the case group and were compared with an equal number of patients in the control group, who were admitted on the same nights but did not have a seclusion or restraint episode during their hospitalization.

Descriptive data (age and gender), clinical data (diagnosis and medications prescribed), and length of stay were extracted from medical records for all individuals. On the basis of past research, primary discharge diagnoses of bipolar disorder, mixed or manic episode, schizophrenia, and schizoaffective disorder were defined as high-risk diagnoses for seclusion and restraint events (5). For patients in the case group, seclusion and restraint characteristics (need for, number of episodes of, and time to event) and associated psychopharmacological interventions were recorded. Prescribing practices were classified according to whether "as-needed" medications (PRNs) were made available and whether standing medication was increased, decreased, or left unchanged during the first 48 hours of hospitalization. Early psychopharmacological intervention was defined as starting patients on a new standing medication or increasing the dosage of the standing medications during this period.

Analyses were conducted with SPSS software, version 16.0. The two groups were compared with the use of contingency tables and chi square tests for categorical variables, t tests and analyses of variance for continuous variables, Poisson regression models for count data, and Kaplan-Meier analyses for time-to-event data. The Beth Israel Medical Center Institutional Review Board declared this research exempt from review.

The sociodemographic and clinical characteristics of the two groups are presented in Table 1 . Patients in the case and control groups differed at a statistically significant level with regard to age (t=2.6, df=76, p=.01), sex ( χ ² =5.2, df=1, p=.023), and length of stay (t=–3.4, df=76, p=.001).

Patients in the case group were also more likely to have a high-risk primary discharge diagnosis ( χ ² =14.2, df=1, p<.001). Among patients at high risk of seclusion or restraint, those in the case group were more likely to have PRN orders written for agitation. PRN medications were prescribed for 100% of patients in this group but for only 78% of patients in the control group ( χ ² =8.8, df=1, p=.003). The patients who had PRN orders averaged 1.2 episodes of seclusion and restraint, whereas those without PRN orders averaged 0 (t=-2.2, df=76, p=.032).

A total of 80 seclusion or restraint events occurred in the study period. In the Poisson regression, the patients whose standing medication was left unchanged during the first 48 hours of hospitalization had 5.5 times greater predicted counts of subsequent restraints than those whose medication dosages were increased or who received new prescriptions for medications (exponentiated regression coefficient [Exp B]=5.5, 95% confidence interval [CI]=1.22–24.81, p=.027). A multivariate Poisson regression that adjusted for the high-risk diagnoses replicated this finding (Exp B=5.3, 95% CI=1.18–23.98, p=.030). Patients whose standing medication was decreased during the first 48 hours of hospitalization had a 1.2 times greater number of seclusions and 2.3 times greater number of restraints than those whose medication was increased or for whom new medications were prescribed. However, neither of these differences was statistically significant.

There were two main findings in this study. First, early initiation of standing medications intended to control agitation and treat psychopathology for newly admitted patients was associated with a greater than fivefold reduction in the use of future physical restraints. Nevertheless, many patients do not receive adequate medication treatment early in the course of their hospitalization. In a previous study, just 34% of all patients admitted to a large psychiatric hospital received the necessary medications at the time of admission (4).

Second, the admitting clinicians in this study were able to predict future seclusion and restraint incidents, as indicated by the greater number of PRN prescriptions for patients with a high-risk diagnosis who had subsequent seclusion or restraint events. Yet in many cases, these clinicians did not start new standing medications or increase the dosage of the patient's current medications. This might represent a "wait-and-see" approach of the weekend treatment teams, perhaps related to the diminished weekend coverage in hospitals recently noted by Shulkin (6). Future research needs to explore the barrier to earlier psychopharmacological treatment when the risks of future seclusion and restraint are clearly detected.

Future attempts at reducing seclusion and restraint occurrences should also focus on formalizing a set of guidelines for predicting and preventing these events in inpatient settings. In this study, young males with a diagnosis of bipolar disorder, mixed or manic episode, schizoaffective disorder, or schizophrenia were at greatest risk of seclusion and restraint. These findings are in keeping with those of previous studies suggesting a higher risk of seclusion and restraint among younger patients and patients with a primary diagnosis of bipolar disorder or a psychotic disorder (5,7).

This study's findings should be considered in the context of the study's limitations. The study focused on a limited number of factors associated with the occurrence of seclusion and restraint, namely demographic characteristics, diagnosis, and the use of PRN and standing medications. Alternative approaches to the control of aggression in inpatient settings, such as behavioral interventions, specialized staff training, and administrative initiatives (7–10), were not examined. Also not explored were patient-staff interactions, including staff attitudes and behaviors related to the application of seclusion and restraint. Future research also needs to further explore the effectiveness of PRN medication and comparative effectiveness across medication groups (that is, benzodiazepines, mood stabilizers, and first- and second-generation antipsychotics) in reducing the occurrence of seclusion and restraint without engendering significant drug-induced side effects.

Efforts to predict the use of seclusion and restraint would benefit from prospective study designs. This study may be perceived as too narrow because it was limited to weekends. However, weekends were specifically targeted because of the likelihood that treatment may be more conservative during this period and because of the greater tendency than one would encounter during the week to opt for observation rather than for increasing medications. A final limitation for discussion is that this study did not explore differences in adverse effects of medication between patients who received increased doses of medication over the weekend and those who did not. Higher doses of psychiatric medications are often associated with increased risk of sedation, confusion, abnormal movements, and metabolic disorders—adverse effects that can complicate the course of hospitalization and patient outcomes. The benefits of reduced seclusion and restraint as a result of earlier adjustment of standing medication should be balanced against the increased risk of these adverse effects, and treatment plans—including decisions to adjust medication—should be tailored to the patient's specific needs. Furthermore, future studies need to explore whether the benefits of medication adjustment early in the course of acute hospitalization justify the possibility of increased or earlier occurrence of adverse effects of such medication in this patient population.

Although admitting clinicians are able to predict future incidents of seclusion and restraint, at least among patients with high-risk diagnoses, on many occasions clinicians failed to appropriately increase the dosage of standing medications or start prescribing new ones—interventions that could reduce the need for seclusion and restraint during the inpatient admissions. Changing the practice styles of admitting teams and removing the barriers to effective early psychopharmacological prevention of agitation and aggressive behaviors in inpatient settings would reduce the incidence of unnecessary seclusion and restraint events. Future studies should consider the relative safety of this practice.

Dr. Mojtabai has received research funding and consulting fees from Bristol-Myers Squibb. The other authors report no competing interests.

Dr. Goldbloom and Dr. Serby are affiliated with the Department of Psychiatry, Beth Israel Medical Center, First Ave. and 16th St., New York, NY 10003 (e-mail: [email protected]). Dr. Mojtabai is with the Department of Mental Health, John Hopkins Bloomberg School of Public Health, Baltimore.