1Department of Psychiatry, Tai Po Hospital, Tai Po, New Territories, Hong Kong, China; kh.gro.ah@579ktl
2School of Nursing, the Hong Kong Polytechnic University, Kowloon, Hung Hom, Hong Kong, China
Find articles by Tsz-Kai Lee
Maritta Välimäki
3Xiangya Nursing School, Central South University, Changsha 410013, China
4Department of Nursing Science, University of Turku, 20014 Turku, Finland; if.utu@alejet
Find articles by Maritta Välimäki
Tella Lantta
4Department of Nursing Science, University of Turku, 20014 Turku, Finland; if.utu@alejet
Find articles by Tella Lantta
Paul B. Tchounwou, Academic Editor
Author information Article notes Copyright and License information Disclaimer
1Department of Psychiatry, Tai Po Hospital, Tai Po, New Territories, Hong Kong, China; kh.gro.ah@579ktl
2School of Nursing, the Hong Kong Polytechnic University, Kowloon, Hung Hom, Hong Kong, China
3Xiangya Nursing School, Central South University, Changsha 410013, China
4Department of Nursing Science, University of Turku, 20014 Turku, Finland; if.utu@alejet
*Correspondence: nc.ude.usc@ikeamileav.attiram
Received 2021 May 24; Accepted 2021 Jun 19.
Copyright © 2021 by the authors.
Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution [CC BY] license [//creativecommons.org/licenses/by/4.0/].
Associated Data
Data Availability StatementThe data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical and privacy reasons.
Abstract
There is a considerable amount of literature describing how nurses’ knowledge contributes to their attitudes and practices related to patient physical restraint. However, whether or not there have been any improvements in nurses’ knowledge levels, attitudes or practices regarding physical restraint during the past few years is unknown. A survey was conducted on nurses [n = 133] in one psychiatric hospital in Hong Kong [n = 98, response rate = 74%]. The data were analyzed using independent t-tests, ANOVA, a Mann–Whitney U test, a Kruskal–Wallis test and Spearman’s rho. In general, nurses had good restraint-related knowledge with satisfactory attitudes and practices, although their knowledge levels, attitudes, and practices regarding restraint varied. Having a higher age, seniority, and education level contributed to a higher restraint-related knowledge level. Male nurses demonstrated more desirable practices [i.e., care of restrained patients], while nurses with a higher education level were more likely to avoid restraint. Nurses’ restraint-related knowledge positively correlated with restraint practices. Although nurses’ knowledge levels, attitudes, and practices regarding restraint were found to be satisfactory, more training efforts should focus on young nurses working in psychiatric settings with less work experience and lower education levels. As some nurses seem to favor the use of restraint with limited reflection, more studies are needed to verify nurses’ emotions and how their emotions influence the use of restrictive practices.
Keywords: physical restraint, nursing knowledge, nursing practice, nursing attitudes
1. Introduction
Patient physical restraint is a controversial topic in healthcare services [1]. Physical restraint, such as magnetic restraint, belts or manual restraint, is used to protect patients from self-harm and to keep them from harming others [2]. However, various adverse effects for patients have been reported, such as psychological distress, affected physical health condition [3] and even death [4]. It has also been reported that the use of restraints exposes staff to occupational hazards [5] and psychological burden [6]. Further, a systematic review by [7] estimated that economic costs for using force, manual restraint and seclusion can total EUR 28,518 annually for one ward.
In a review conducted by [8], the authors revealed that restraint can be a form of abuse when it is inappropriately used, and can result in fear, neglect, and failure to use de-escalation techniques. Therefore, more critical attitudes towards these restrictions are needed [9]. Staff members’ abilities to reflect on their own attitudes, emotions and actions should also be promoted in order to reduce coercive practices [10], which should be used as a last resort when other less restrictive methods are exhausted [11]. Indeed, intervention approaches such as the Safewards model have been shown to increase meaningful activities on wards and thereby decrease the use of coercive practices [12]. Such approaches involve a package of agitation management strategies, including expanding de-escalation skills among ward staff; regular patient meetings to facilitate inter-patient support; having a structured, personal and innocuous information flow between staff and patients; and displaying positive messages about the ward from discharged patients [12]. Other good examples include patient involvement in aggression management [13] and recovery-oriented programs with anger management technique training for the staff and patients involved [14].
Despite the wide range of existing international studies regarding the use of physical restraint, the literature in this field is still not globally well-targeted. Still, the topic is important because restraint is currently a common practice in Asian psychiatric hospitals. In mainland China the prevalence of using mechanical restraints in psychiatric care has been reported to vary between 27.2% [n = 1364 patients] [15] and 51.3% [n = 160 patients] [16]. In one study carried out in Hong Kong, 39.7% [n = 335] of patients were restrained within the first week of admission [17]. In Taiwan, 29.5% [n = 59] of patients visiting psychiatric emergency care were restrained during their care episode [18]. A review by [19] including studies from Chinese language databases concluded that the frequency of restraint is higher in China compared to average numbers globally. From the patient’s perspective, the use of restraints was perceived negatively: over half [61.2%] of psychiatric inpatients surveyed in Hong Kong [n = 129] reported traumatic experiences due to witnessing another patient being taken down and 41.1% of that group reported traumatic experiences due to witnessing another patient being put in restraints of any kind [20].
Nurses are the primary decision makers who apply restraints in psychiatric care settings [21]. Although nurses were found to view the use of restraints as an inevitable intervention in psychiatric inpatient settings, necessary for calming patients, previous studies have reported that nurses have acknowledged that the intervention should be a “last resort” to use only after all the less coercive alternatives have been tried [6]. However, it has been assumed that the knowledge and attitudes among staff may directly or indirectly affect restraint practices and thus the likelihood for initiating restraints [22]. A systematic review by [23] concluded that, during the last two decades, staff’s attitudes towards coercion have changed from seeing it as part of therapeutic care towards viewing it as a matter of safety. Indeed, this may be the case in countries where significant efforts have been made to diminish the use of coercion, especially restraint [24,25]. Still, outside Western countries, studies report only moderate knowledge and attitudes towards coercion and the use of physical restraints [26,27].
To better understand the current situation in Hong Kong and to allow comparability to other countries and previous situations in Hong Kong [28], we conducted a survey using the Physical Restraint Questionnaire [PRQ] [29]. The PRQ is an instrument to measure nurses’ knowledge, practice, and attitudes regarding physical restraint. This instrument and its various language versions have been used in inpatient psychiatric settings in Hong Kong [30], Finland [31], India [32], Malaysia [33], Saudi Arabia [26], Sudan [11], and Turkey [34,35,36].
The PRQ has been previously used for both evaluating effectiveness of interventions and for cross-sectional study purposes. Results from two trials with continuing education programs offered contradictory results using the PRQ. An eLearning course for nurses did not have long or short-term effects on nurses’ knowledge, attitudes or practices regarding the use of physical restraints [31], while another study with group-based psychoeducation for nurses had positive short-term effects on nurses’ knowledge, attitudes, and practices [35]. Mixed results have also been found in cross-sectional studies. Levels of knowledge, attitudes and practices have varied between studies from moderate to excellent [33,34]. For instance, clear deficits have been detected in nurses’ knowledge about treatment complications [34] and alternatives [33] related to physical restraints. On the other hand, results regarding factors associated with nurses’ knowledge, attitudes, and practices were partially inconsistent between studies. Nurses’ gender, level of education, years of work experience and type of working unit seem to impact knowledge, attitudes and practices regarding the use of restraints, although findings appear to vary between countries. One previous study was conducted in local psychogeriatric wards in Hong Kong [30] with a small sample size [n = 42 nurses]. The study’s findings were inconsistent with several international studies [37,38] in terms of nurses’ emotions towards restraint; more than half [n = 15, 53%] of the nurses had no ethical dilemma regarding restraints [30]. In this study, we therefore describe the current state of the knowledge, practices and attitudes among nurses about physical restraint and the associated factors at a psychiatric inpatient hospital in Hong Kong, and compare them to previous local and international findings. We aim to identify potential gaps in current local practices in Hong Kong to identify where targeted interventions could possibly be offered to improve nurses’ know-how related to psychiatric inpatient care.
2. Materials and Methods
2.1. Study Design
A cross-sectional design was adopted for this study.
2.2. Setting
The study was conducted in one psychiatric department at a hospital in Hong Kong. The hospital provides psychiatric inpatient and outpatient services [day hospital and community psychiatric services] to people living in one cluster with a population of less than 2 million [39]. The study was conducted on 8 inpatient wards [4 male and 4 female wards], of which all used restraint intervention. The wards specialized in general adult psychiatry, learning disabilities, substance abuse, and psychogeriatrics. Patients were admitted on voluntarily or compulsorily bases. At the time of data collection in December 2016, the average patient capacity was 45 beds per ward, and the length of stay was approximately 36 days.
2.3. Study Population and Eligibility Criteria
Our target population was nursing staff working in inpatient psychiatric care. The current size of the psychiatric nursing staff population in Hong Kong is 3266 registered nurses [psychiatric] and 1637 enrolled nurses [psychiatric] [40]. The Hospital Authority manages 43 public hospitals and institutions, 49 specialist outpatient clinics and 73 general outpatient clinics in Hong Kong. These organizations are divided into seven hospital clusters based on their locations. The aim of the hospital clusters is to ensure that patients receive a continuum of high-quality care within the same geographical setting and throughout their episode of illness from the acute phase through convalescence, rehabilitation, and community after-care [41].
In our study of wards in December 2016, there were a total of 157 nurses [1 department operational manager, 7 ward managers] and other included nurses were advanced practices nurses, registered nurses and enrolled nurses] with a high school diploma or below, a bachelor’s degree, a master’s degree or a PhD.
The inclusion criterion for the participants was that they were qualified nurses [i.e., nurses enlisted in the Nursing Council of Hong Kong] [42] who provided direct patient care and who worked in inpatient settings during the time of data collection. The exclusion criteria were if they were nursing students, nurses on any kind of leave during the data collection period, or nurses working in outpatient settings.
2.4. Instruments and Variables
Background information was collected on gender [male/female], years of experience as a qualified nurse, educational level [high school diploma or below/bachelor’s/master’s degree or above] and exposure to management of violence [MOV] training [yes/no]. In Hong Kong, in-service training for management of violence in all psychiatric hospitals consists of three levels of training: the first and second levels involve basic workplace violence awareness with de-escalation techniques and break-away techniques respectively, and they are mandatory for all qualified psychiatric nurses, whereas the third level deals with control and restraint techniques for patients with agitation/aggression, using a team approach [43].
The Physical Restraint Questionnaire [PRQ] [29] was used to measure the nurses’ knowledge, practices and attitudes regarding physical restraint. A physical restraint is defined by [29] as “any article, device, or garment that interferes with a person’s free movement and secures him or her to a bed or chair” [p. 345]. The PRQ was selected because it allows international comparisons, its content is suitable in the Hong Kong practices, and its validity and reliability have been reported to be good [44].
The PRQ includes three sections. First, the Knowledge section consists of 18 statements using a 3-point Likert scale [“true,” “false,” “not sure”]. Every correct answer gets 1 point, while indicating “not sure” or providing a false answer merits 0 points. The total score can range from 0 to 18; a higher score indicates a higher number of correctly answered items [45]. The intraclass correlation coefficient has been established to be 0.85 [44] and the Cronbach’s alpha to be 0.65 [46]. Second, the Practice section consists of 18 statements measured with a 3-point Likert scale [“Always,” “Sometimes,” “Never”] [45]. This section concerns the nursing care prior to and during the use of restraints. In this section, 4 statements are negatively phrased, and the ratings are reverse-scored [“Never,” “Sometimes,” “Always”]. A total score of 18 represents the least desirable practice and 54 the most desirable practice. The intraclass correlation coefficient has been established to be 0.99 [44] and the Cronbach’s alpha to be 0.94 [46]. Third, the Attitudes section consists of 12 items measured with a 3-point Likert scale [“Agree,” “Neutral,” “Disagree”] [45]. Again, negatively phrased items are reverse-scored. The total score ranges from 12 to 36. The lower the score, the less desirable the attitudes among nurses towards restraint. The intraclass correlation coefficient was established to be 0.84 [44] and the Cronbach’s alpha to be 0.61 [46].
2.5. Data Collection and Sample Size
The data were collected between 12 December 2016 and 2 January 2017. First, ward managers of the 8 wards were contacted one week before the data collection period to confirm the number of nurses for which to prepare the questionnaires. They were provided with an explanation letter for the study and asked to encourage staff participation and facilitate opportunities for them to complete the questionnaires, so as to boost the response rate. Second, on the first day of data collection, the questionnaires were delivered by the researcher to the clerical staff of each ward, and subsequently distributed to the participants in their personal mailboxes. A temporary drop-box was placed on each ward to facilitate the collection. Third, the questionnaires were collected by the researcher in person on the final day of data collection.
The sample size required to represent the population was calculated to be 96 with a margin of error of 10% and a confidence level of 95% [47]. To achieve the sample size needed, convenience sampling was used. All nurses fulfilling the eligibility criteria were invited to participate in the study.
To manage missing data, list-wise deletion [excluding any questionnaires returned with missing values] [48] was used, since less than 5% of the questionnaires had missing data [49]. Answers to negatively phrased statements were reversed to be congruent with the answers of the other items.
Due to the small sample size, personal data of the respondents were recategorized for a more relevant statistical analysis, and the data were presented using descriptive statistics [standard deviation, frequencies, percentages, mean, median, frequencies and percentages]. To generate the mean total score of a particular section [i.e., Knowledge, Practices, Attitudes], the ratings of a participant were first summed up to calculate a total score. After that, the calculation was repeated to generate an array of total scores for all the participants regarding each section. All of these total scores were summed up to generate a mean total score [50].
Nurses’ responses in the Practice and Attitudes sections were also recategorized. In the Practice section, the 3-point scale [“Always,” “Sometimes,” “Never”] was recategorized into 2 categories [“Always,” “Sometimes/Never”]: responses in the negatively phrased statements [e.g., “All disoriented acute care patients should be restrained”] were recategorized from “Never,” “Sometimes,” “Always” to be “Never,” “Sometimes/Always.” Similar recategorizations were done in the Attitudes section: responses were recategorized from a 3-point scale to be “Agree” or “Neutral/Disagree,” while responses in the negatively phrased statements were grouped as either “Disagree” or “Neutral/Agree.”
2.6. Statistical Methods
Independent t-tests [49] or ANOVA were used for the total score[s] that were normally distributed. In cases of non-normal distribution of the scales, a Mann–Whitney U-test [median test] or Kruskal–Wallis test was used to compare possible differences in scores between the sub-groups [e.g., males vs. females] [49,51]. An ANOVA or Kruskal–Wallis test was used to describe any group differences, and a post-hoc analysis was used to identify in which specific groups a difference could be found [49] by using the Bonferroni test for ANOVA [49] and the Dunn’s test for Kruskal–Wallis [52]. In addition, Spearman’s rho was used [50] to describe correlation between the knowledge, practice and attitudes scores.
All analyses were performed using SPSS 23 [IBM Corp. IBM SPSS Statistics for Windows, Version 23.0; Armonk, NY, USA] [53], and the level of statistical significance was set to p < 0.05 [51].
2.7. Ethical Considerations
Ethical approvals were granted by the local university human ethics committee [HSEARS20160923001] and at the hospital ethics committee [CRE-2016.536]. An information letter of the study was given to the participants as the cover of the questionnaire to explain the study’s purpose, procedure, voluntary participation as informed consent [returned and completed questionnaire was implied to be consent to participate], and privacy issues. Anonymity and confidentiality were ensured by coding all completed questionnaires [54].
3. Results
3.1. Participants
Out of our total study population [n = 157], 133 were eligible to participate in the study. Those who were excluded in the target population worked in outpatient settings. We received 98 completed questionnaires [response rate 74%].
3.2. Descriptive Data
Out of 98 nurses, males were in a slight majority [53.1% vs. 46.9%]. Ages ranged from 25 to 55 years [Mdn [median] = 30, inter-quartile range = 11]. About three-quarters [n = 70, 71.4%] had a nursing education level higher than a bachelor’s degree, and half [n = 49, 50%] had 5 or more years of nursing experience. All of them [100%] had completed mandatory in-service training for the management of violence [see Table 1].
Table 1
Demographic characteristics of nurses [n = 98].
AttributesnPercentageGenderMale5253.1Female4646.9Age [Divided by Median]