Pennsylvania State Hospital System's Seclusion and Restraint Reduction
Total Page:16
File Type:pdf, Size:1020Kb
smi.qxd 8/22/2005 9:42 AM Page 1115 Special Section on Seclusion and Restraint Pennsylvania State Hospital System’s Seclusion and Restraint Reduction Program Gregory M. Smith, M.S. Robert H. Davis, M.D. Edward O. Bixler, Ph.D. Hung-Mo Lin, Ph.D. Aidan Altenor, Ph.D. Roberta J. Altenor, M.S.N. Bonnie D. Hardentstine, B.S. George A. Kopchick, M.S. Objectives: This study examined the use of seclusion and mechanical re- he use of seclusion and re- straint from 1990 to 2000 and the rate of staff injuries from patient as- straint in psychiatric hospitals saults from 1998 to 2000 in a state hospital system. Methods: Records of Tto control people with dis- patients older than 18 years who were civilly committed to one of the nine turbed or violent behaviors has been state hospitals in Pennsylvania were included in the analyses. Two data- occurring for centuries. The litera- bases were used in each of the nine hospitals: one identified date, time, ture is replete with examples of how duration, and justification for each episode of seclusion or restraint and clinicians have struggled with the eth- the other identified when a patient was hospitalized and the demograph- ical issues involved with the use of ic characteristics and the diagnosis of the patient. Rate and duration of these interventions (1–4). The litera- seclusion and restraint were calculated. Reports from compensation ture has also provided the rationale claims were used to determine staff injuries from patient assaults. Results: for the continued use of physical con- The rate and duration of seclusion and mechanical restraint decreased trols in modern psychiatric settings, dramatically during this period. From 1990 to 2000, the rate of seclusion much like it did in the past for asy- decreased from 4.2 to .3 episodes per 1,000 patient-days. The average lums (4–6). More recent research has duration of seclusion decreased from 10.8 to 1.3 hours. The rate of re- questioned the need to use these straint decreased from 3.5 to 1.2 episodes per 1,000 patient-days. The av- practices and pointed to studies erage duration of restraint decreased from 11.9 to 1.9 hours. Patients showing inconsistency in the applica- from racial or ethnic minority groups had a higher rate and longer dura- tion of seclusion and restraint (6,7). tion of seclusion than whites. Seclusion tended to be less likely, but An international review of these pro- longer, during the night shift. Patients were restrained less often during cedures found the same inconsisten- the night shift, but for a longer duration. The rate of restraint was high- cies in the use of seclusion and re- er during the week than during weekends and holidays. Younger patients straint with psychiatric patients and were more likely to be secluded and restrained, but older patients re- the same struggles with the ethics of mained secluded and restrained longer. No significant changes were seen these practices (8). Seclusion and re- in rates of staff injuries from 1998 to 2000. Conclusions: Many factors straint has prevailed internationally as contributed to the success of this effort, including advocacy efforts, state the primary approach to managing vi- policy change, improved patient-staff ratios, response teams, and second- olent behaviors, although the degree generation antipsychotics. (Psychiatric Services 56:1115–1122, 2005) of use depends on the country or ge- ographical location (9–12). Since 1990 the Pennsylvania state hospital system has experienced con- Mr. Smith, Dr. Davis, Dr. Altenor, Ms. Altenor, Ms. Hardentstine, and Mr. Kopchick are affiliated with the Office of Mental Health and Substance Abuse Services of the Com- stant changes that have resulted in re- monwealth of Pennsylvania in Harrisburg. Dr. Bixler and Dr. Lin are with the Pennsyl- markable outcomes in reducing the vania State University School of Medicine in Hershey. Send correspondence to Mr. Smith use of seclusion and restraint. The at Allentown State Hospital, 1600 Hanover Avenue, Allentown, Pennsylvania 18109-2498 Pennsylvania state hospital system, (e-mail, [email protected]). This article is part of a special section on the use of seclu- one of the oldest in the country, sion and restraint in psychiatric treatment settings. serves adults within a civil and foren- PSYCHIATRIC SERVICES ♦ http://ps.psychiatryonline.org ♦ September 2005 Vol. 56 No. 9 1115 smi.qxd 8/22/2005 9:42 AM Page 1116 sic inpatient setting who have severe were averaged across patients who and older. Race and ethnicity were and persistent mental illnesses. Dur- had at least one event during that considered in terms of whites and all ing most of its history the hospital sys- year. racial and ethnic minority groups. tem provided direct admission servic- The Jonckheere-Terpstra test (13) The events were categorized in terms es. However, during the study period, was used to compare the ordinal dif- of hospital work shift (day shift, 7 a.m. the admission of civilly committed pa- ferences of the seclusion and restraint to 3 p.m.; evening shift, 3 p.m. to 11 tients was limited to referrals from lo- rates over time or among different p.m.; and night shift, 11 p.m. to 7 cal psychiatric acute care settings for classes. The generalized estimating a.m.) and by weekday compared with individuals who were unable to be equations (GEE) method (14) was weekend or holiday. The Jonckheere stabilized within a 30-day acute care employed to perform various hypoth- Terpstra test was used to evaluate stay. esis tests, including testing whether data derived from the annual census The purpose of this study was to the overall duration or rate of one fac- data from the hospital system. All systematically report changes in pat- tor depends on the level of the other other analyses were evaluated by us- terns of seclusion and mechanical re- factor. The GEE method is a useful ing the GEE method (14). straint over an 11-year period (1990 extension of the generalized linear Data on staff injuries from patient to 2000) that took place within Penn- models for continuous and categori- assaults were available only for 1998 sylvania’s nine state hospitals. Be- cal data that does not require the ex- to 2000. We examined the number of cause so many changes were imple- events in which employees lost time mented during this period it was not from work as a result of injuries from possible to isolate and analyze the ef- assaults and events in which first aid fects of a single variable. Changes only was needed as a result of an as- that influenced the reduction of sault. Each hospital’s compensation seclusion and restraint are discussed. claims were used for this analysis. Since 1990 The counts were not duplicated with- Methods in these measures. To determine the The data used for this study were tak- the Pennsylvania rate of change each year, these events en from two sources maintained by were calculated per 1,000 patient- each of the nine hospitals. The first state hospital system days of care. database identified date, time of day, duration, and justification for each has experienced constant Results event of seclusion or restraint. The The annual census for the hospital second database identified when a changes that have resulted system during this 11-year period de- patient was present in the hospital creased 56 percent, from about 6,300 and supplied basic demographic and in remarkable outcomes to about 2,800. During this time the diagnostic information. All data were gender ratio remained constant (53 merged and reformatted by using an in reducing the use percent men and 47 percent women), anonymous patient identifier for and representation of racial or ethnic analysis. Only adults 18 years and old- of seclusion and minority groups increased slightly er who were served on civil units of (p=.03). This change was consistent the hospitals were included in this restraint. with changes in the size of the popu- study. In addition, only the use of lation of racial or ethnic minority seclusion and mechanical restraint groups in Pennsylvania. was evaluated. As shown in Table 1, the rate of An individual annual rate was cal- seclusion decreased from a high of culated as the total number of events 7.2 episodes per 1,000 patient-days in divided by the total number of days act knowledge of variance-covariance 1991 to .3 episodes per 1,000 patient- in the hospital that year multiplied structure of the repeated outcomes. days in 2000. The average rate of by 1,000 (referred to as rate per 1000 The GEE is favored over multivariate seclusion was similar for men and patient-days). If a patient was not re- analysis of variance (MANOVA) be- women, and this rate significantly de- strained during that year’s stay in the cause our study consists of double re- creased during the study period hospital, they were assigned a rate of peated measurements from hospitals (p<.01). The rate of seclusion was zero. For the year-specific average and patients nested under hospitals greater for persons in racial or ethnic rate, the individual rates were aver- over the 11-year period. Estimation minority groups than for whites (4.3 aged across all patients in the hospi- of the covariance structure using episodes per 1,000 patient-days com- tal that year. Duration was defined as MANOVA proves to be problematic pared with 2.5 episodes per 1,000 pa- hours per event. An individual annu- sometimes. tient-days, p<.01). Over the study pe- al duration was calculated if the pa- For the purpose of analysis, the riod, the rate of seclusion decreased tient had one or more events that sample was divided into four age to a greater degree among persons year.