Mental health professionals ’ feelings and attitudes towards coercion

Background: Despite absence of clear evidence to assert that the use of coercion in psychiatry is practically and clinically helpful or effective, coercive measures are widely used. Current practices seem to be based on institutional cultures and decision-makers ’ attitudes towards coercion rather than led by recommendations issued from the scientific literature. Therefore, the main goal of our study was to describe mental health professionals ’ feelings and attitudes towards coercion and the professionals ’ characteristics associated with them. Method: Mental health professionals working in the Department of Psychiatry of Lausanne University Hospital, Switzerland, were invited to participate to an online survey. A questionnaire explored participants ’ socio- demographic characteristics, professional background and current working context, and their feelings and attitudes towards coercion. Exploratory Structural Equation Modelling (ESEM) was used to determine the structure of mental health professionals ’ feelings and attitudes towards coercion and to estimate to which extent sociodemographic and professional characteristics could predict their underlying dimensions. Results: 130 mental health professionals completed the survey. Even if a large number considered coercion a violation of fundamental rights, an important percentage of them agreed that coercion was nevertheless indis- pensable in psychiatry and beneficial to the patients. ESEM revealed that professionals ’ feelings and attitudes towards coercion could be described by four main dimensions labelled “ Internal pressure ” , “ Emotional impact ” , “ External pressure ” and “ Relational involvement ” . The personal as well as the professional proximity with people suffering from mental disorders influences professionals ’ feeling and attitudes towards coercion. Conclusions: As voices recommend the end of coercion in psychiatry and despite the lack of scientific evidence, many mental health professionals remain convinced that it is a requisite tool beneficial to the patients. Clinical approaches that enhance shared decision making and give the opportunity to patients and professionals to share their experience and feelings towards coercion and thus alleviate stress among them should be fostered and developed.


Introduction
In 2019, Dunja Mijatovic, the Council of Europe Commissioner for Human Rights, in her speech in front of the Parliamentary Assembly of the Council of Europe, recommended the end of coercion in mental health (Council of Europe Commissioner for Human Rights, 2019). The argument was based on the UN Convention on the Rights of Persons with Disabilities (the CRDP) (UN General Assembly, 2006) which pros (Newton-Howes & Gordon, 2020;Zinkler & von Peter, 2019) and cons (Appelbaum, 2019) were strongly discussed among academics. More recently, the World Psychiatric Association issued a statement on the necessity to develop alternatives to coercion in psychiatry in order to improve mental health care (Rodrigues, Herrman, Galderisi, & Allan, 2020). Several practices, policies and interventions seem promising to achieve this goal (Barbui et al., 2020;Gooding, McSherry, & Roper, 2020). The use of coercion in psychiatry has been debated for centuries, because it counterpoises the fundamental biomedical ethics principles of respect of patients' autonomy, beneficence and non-maleficence, and justice. Medical literature in psychiatry tends to justify the use of coercion based on the ethical assumption that coercive measures are implemented in the best interest of the patient (beneficence), their relatives, and the society (Monahan et al., 1995;Szasz, 2009). However, no definitive evidence allows asserting that the use of coercion is practically and clinically helpful or effective (Huber et al., 2016;Kisely, Campbell, & O'Reilly, 2017;Muralidharan & Fenton, 2006;Sailas & Fenton, 2000). Some studies have shown its negative impact on patients (De Haan, Van Amelsvoort, Dingemans, & Linszen, 2007;Jaeger et al., 2013;Kalisova et al., 2014;Lay, Kawohl, & Rossler, 2018;Nyttingnes, Ruud, & Rugkåsa, 2016;Rusch et al., 2014). Nevertheless, coercive measures are still widely adopted almost everywhere (Hotzy et al., 2018;Raboch et al., 2010;Rains et al., 2019).
The goal of our study was to explore mental health professionals' feelings and attitudes towards coercion and the sociodemographic and professional characteristics associated with them.

Participants
Eligible participants were all the mental health professionals working in the General Psychiatry (including hospital and ambulatory units), the Community Psychiatry (including mobile teams and the rehabilitation unit), the Emergency Psychiatry, the Child and Adolescent Psychiatry and the Psychogeriatric Services of the Department of Psychiatry of Lausanne University Hospital. A presentation of the study and an invitation to participate with a link to an online survey was sent by email to all the participants. In 2018, the canton of Vaud, where Lausanne is located, had a rate of psychiatric beds per 1 ′ 000 inhabitants of 0.68 (Infosan.vd, 2020). The rate of involuntary hospitalisations per 1 ′ 000 inhabitants was 1.87 (Obsan, 2020), one of the highest in Switzerland (Silva, Golay, & Morandi, 2018).

Measures
The online survey was a structured questionnaire specifically elaborated for the study. The first part of the survey consisted of nine questions about participants' sociodemographic characteristics, professional background and current working context. In the second part, participants were asked if they had faced different forms of formal and informal coercion at least once during the previous six months. Formal coercion was defined as involuntary hospitalisation, community treatment orders, forced medication, seclusion, closed ward and restraint.
Informal coercion was defined as by Szmukler and Appelbaum (Szmukler &Appelbaum, 2008): persuasion, interpersonal leverage, inducements andtreats (Szmukler &Appelbaum, 2008). The final part of the survey was a set of questions that explored participant's feelings and attitudes towards coercion in mental health care. It contained 37 questions answered on a 5-point Likert scale. The questions were elaborated by the research team and based on its members' clinical experience with coercion, then discussed and improved during two focus groups. The focus groups were conducted with other mental health professionals from different services and institutions, people with mental health problems and peer practitioners. During the focus group discussions, participants discarded, rephrased, or suggested new questions.

Statistical analysis
All statistical analyses were performed using Mplus 8 and IBM SPSS 26. All statistical tests were two-tailed and significance level was set at p < .05. Descriptive statistics were used to summarise participants' sociodemographic and professional characteristics, including their experience of formal and informal coercion during the last six months, as well as their current feelings and attitudes towards coercion. While the questions were not designed as a psychometric scale, we wanted to verify whether underlying dimensions influencing the professionals' answers could be revealed. This could allow us to verify the multivariate relationship between the questions and participants' sociodemographic characteristics, professional background and current working context. Thus, Exploratory Structural Equation Modelling (ESEM) with oblique quartimax rotation was first performed on the 37 questions. Cattell's scree test (Cattell, 1966), model comparison, model fit and factor interpretability were used to determine the most adequate number of factors. ESEM was then performed to estimate the relationship between sociodemographic and professional characteristics and the factor scores. Each variable was first tested independently. Then, a multivariate parsimonious model was performed, including all the variables reaching a p < .05 level of significance in the univariate analyses.

Participants
Out of the 730 mental health professionals (medical doctors, nurses, psychologists, social workers, occupational therapists) working in the services contacted for this study, 130 (17.8%) accepted to participate and completed the survey. The sample description is detailed in Table 1.

Mental health professionals' feelings and attitudes towards coercion
Answers to the 37 questions exploring the professionals' feelings and attitudes towards coercion are detailed in Table 2.
About a third of professionals agreed or completely agreed that coercion was a violation of the fundamental rights of the patients (question 4; 34.6%) and was not used fairly (questions 5; 37.6%). Informal coercion was taken into account as well as formal coercion (question 8). However, and despite the lack of scientific evidence, an important proportion of professionals agreed or completely agreed that coercion was indispensable in psychiatry (question 1; 42.3%), beneficial to patients (question 2; 35.4%), that it had a therapeutic role (questions 7; 47.7%) and that it did not hinder patients' recovery (question 3; 59.2%). Moreover, professionals were mostly convinced that they were doing what was best for the patients when they had to resort to coercion (question 13; 56.2%). Thus, only a minority of participants experienced, often or very often, a moral dilemma (question 21; 26.9%) or a dissonance between what they would have liked to do and what they were actually doing (question 22; 15.4%) when resorting to coercion. Nevertheless, using coercion was frequently stressful for professionals (question 9; 37.7%) and left a deep impression on them (questions 10; 53.8%). A majority had to debrief with colleagues afterwards (question 27; 61.5%). Professionals never or rarely experienced loneliness when they had to decide on or to apply coercive measures (questions 14 and 15; 76.1% and 76.9%). Professionals' answers did not clarify whether or not coercion altered the therapeutic relationship (question 18).
Patients' aggressive behaviours (questions 11 and 35; 63.9% and 88.5%) and the disruption of community life were never or rarely identified as reasons to resort to coercion (question 37; 65.4%). On the other hand, coercion could sometimes lead the patients to aggressiveness (question 12; 53.8%). If the lack of alternatives was sometimes recognised as a justification to use coercion (question 19; 38.5%), other external factors such as lack of time (question 20; 93.1%), the fear of legal consequences (questions 23 and 24; 82.3% and 96.2%), the pressure of third party and colleagues (questions 25, 28 and 29; 70.7%, 96.9% and 84.6%), and the administrative burden (question 26; 94.6%) were never or rarely identified as such.
Empathy towards the person placed under coercion was often expressed by professionals (question 16; 84.6%), who were only sometimes expecting patients' gratitude afterwards (question 17; 58.5%). The majority of the professionals estimated to have often enough time to discuss different therapeutic options and to take into account patients' opinion before using coercion (question 31; 70.7%), to consider patients' preferences during coercive measures (question 32; 50.0%) and to talk with them afterwards to know what they went through (question 30; 82.3%). Furthermore, professionals also found it helpful to have access to the patient's advance directives or joint crisis plans (question 33; 52.3%). Most of them did not find it difficult to define therapeutic objectives when care took place under coercion (question 34; 52.3%) and to decide when coercion should end (question 36; 57.7%).

Structure of mental health professionals' feelings and attitudes towards coercion
The scree plot suggested a four factors solution. Model fit could also be considered good starting from four factors (Root Mean Square Error of Approximation (RMSEA) = 0.045; Comparative Fit Index (CFI) = 0.919; Tucker Lewis Index (TLI) = 0.898). Model comparison between the three-and four-factor solutions suggested that four factors were needed in comparison with a simpler model (χ 2 (34) = 105.035, p < .001). Finally, factor interpretability was in line with all these statistical criteria and the four-factor solution appeared as the most satisfactory. ESEM factor loadings revealed that professionals' feelings and attitudes towards coercion could be described by four underlying dimensions: 1."Internal pressure", 2."Emotional impact", 3."External pressure" and 4."Relational involvement" (Table 3).
Firstly, the "Internal pressure" factor described the ethical conflict professionals felt when confronted with coercion. Nine questions loaded above 0.04 on this factor, all of them questioning professionals about the usefulness of coercion (question 1), its benefit for patients (question 2,7 and 13), its negative impact on them (questions 3 and 4) and the moral unease they might experience when they faced it (questions 21 and 22). Secondly, "Emotional impact" was mainly defined by questions focusing on the feelings professionals may experience when confronted with coercion, such as stress (question 2), long lasting memories (question 10), loneliness (questions 14 and 15) and cognitive dissonance (questions 21 and 22). Thirdly, the "External pressure" factor referred to the main issues that might drive professionals to resort to coercive measures. These included patients' behaviours (questions 11, 12 and 35), organizational matters, such as lack of valid alternatives (question 19) or lack of time to implement them (question 20), the administrative burden associated with the use of coercion (question 26) or the fear of legal consequences (question 24). Mental health professionals could also be put under pressure by third parties such as colleagues, relatives, police, legal guardians and other health or social services (questions 25 and 28). Finally, six questions loaded above 0.04 on the "Relational involvement" factor. These questions explored if professionals shared their patients feelings (questions 16 and 17), debriefed with them (question 30), took time to explain them the measure (question 31) and could take into account their preferences (question 32 and 33) when using coercion.

Sociodemographic and professional characteristics associated with participants' feelings and attitudes towards coercion
The ESEM results are presented in Table 4. Results with statistical significance showed that internal pressure was stronger among senior executive clinicians (with two or more subordinate hierarchical levels) and it was reduced by being exposed to any form of coercion during the last six months. Emotional impact was significantly greater among psychiatrists and lower among nurses. It was also weaker for people working in hospital compared to people working in other services. Having a relative with a mental disorder also increased the emotional impact of coercion on professionals. External pressures were significantly higher on psychiatrists than on psychologists and social workers. The need of relational involvement was stronger among psychiatrists, professionals working in services other than the hospital and junior executive clinicians (with at the most one subordinate hierarchical level). On the contrary, it was significantly lower among nurses and non-executive clinicians.
The standardized results for the multivariate parsimonious ESEM model showed that internal pressure was significantly lower among participants who had been confronted with coercion during the last six months. The emotional impact when facing coercion was stronger among participants with a relative with mental disorder. External pressure was lower among social workers. No professionals' characteristic was specifically associated with the need of relational involvement.

Discussion
Even if a large number of mental health professionals considered coercion a violation of fundamental rights and despite the lack of scientific evidence, an important percentage of them agreed that coercion was nevertheless indispensable in psychiatry and beneficial to the patients. Few experimented external pressures that could lead them to use coercion, except for lack of alternatives. Most of them reported empathizing with coerced patients, sharing their experience and taking their preferences into account. The professionals' confidence in their practice should be discussed in light of patients' experience of coercion in mental health settings to see whether they would agree or not. These exchanges could possibly modify professionals' feelings and attitudes. Unfortunately, there was no record of patients' satisfaction with mental health care during the last years in Lausanne. However, a local study has shown that not all the patients would agree that the use of coercion is necessary and useful (Golay, Morandi, Silva, Devas, & Bonsack, 2019). This was also demonstrated elsewhere (Nyttingnes, 2018). Despite the fact that our questions were not designed as a psychometric scale, our study allowed to described professionals' feelings and attitudes towards coercion by four underlying general dimensions labelled "Internal pressure", "Emotional impact", "External pressure" and "Relational involvement", each of them with a specific relationship with participants' sociodemographic characteristics, professional background and current working context.
Previous research has shown that the more professionals were involved in situations where coercion was used the more they believed in it (van Doeselaar, Sleegers, & Hutschemaekers, 2008) and thought of it as a mean to achieve care and security (Dahan et al., 2018). Another study has highlighted that the more professionals experienced general moral doubt, the more they thought that coercion was offending (Molewijk et al., 2017). Moreover, respondents with decisional and leadership responsibilities saw coercion less as treatment. Our results confirmed these findings as senior executive clinician felt more internal pressure when facing coercion, which was instead reduced by being exposed to at least one coercive measure during the previous six months.
Emotional impact of coercion was stronger among psychiatrists, who were the only mental health professionals in the survey authorized to decide on coercive measures. Other professionals such as social workers who are not involved in deciding coercive measures were less affected. Professionals working in hospital, who more frequently face patients in crisis and have to apply coercive measures more regularly, were less emotionally affected by coercion than people working in other services. The use of coercion in these situations could be more easily justified by patients' acute clinical state and the need of protection, security and order. On the other hand, professionals with relatives who suffered mental disorders experienced strongest feelings when they were confronted with coercion. These situations likely reminded them of their relatives or caused emotional resonance, especially if the patients were struggling.
party, such as family, other social and medical services or judicial bodies. The need of relational involvement was greater among psychiatrists, people working in other services than hospital and in junior executive clinicians. One explanation could be that all these groups have more regular contacts and longer therapeutic relationships with patients. In an institution such as the Department of Psychiatry of Lausanne University Hospital, psychiatrists are mainly responsible of the psychiatric ambulatory care and in the front line for the long-term follow up of patients. Moreover, junior executive clinicians are more often in charge of the most difficult patients, who could more easily require the use of coercive measures.
In the multivariate model, internal pressures were significantly lower among participants with at least one experience of coercion during the last six months, emotional impact was stronger among those with a relative with mental disorders and external pressures were lower among social workers. No professionals' characteristics were instead specifically associated with the need of relational involvement.
Our study did replicate only partially the results of previous researches showing that professionals' feelings towards coercion were related to their gender and age. Raveesh et al. (2016) found that psychiatrists had a more negative view of coercion and saw it as more "offensive" and violating the integrity of the patient than caregivers who were much more likely to support the use of coercion as part of the patient's treatment (Raveesh et al., 2016). Moreover, older psychiatrists (above 46 years), with more experience, felt more often that coercion could harm the therapeutic relationship and might represent a failure on the part of the mental health services. Finally, older caregivers, male and more experienced psychiatrists (more than 10 years) believed more than the other participants that coercion should not be used in treatment. These results are to be taken with reserve knowing that other studies have shown that paradoxically, the feeling that coercion was an offence against patients decreased as the total work experience increased (Husum et al., 2011). In another study, Bregar et al. (2018) showed that, among nurses, women were more prone to resort to coercion than men (Bregar et al., 2018). Finally, Steinert, Lepping, Baranyai, Hoffmann, and Leherr (2005) found that to have a relative suffering from mental disorders slightly reduced the use of coercion which was on the contrary supported more frequently by women (T. Steinert et al., 2005).
One strength and novelty of this research project stems from the fact that it was developed in collaboration with people with mental health problems and peer practitioners. It was designed to answer their questions about how professionals feel and behave when they faced coercion. The use of an anonymous online survey proved to be both a strength and a weakness for our study. On the one hand, thanks to this method we did not have missing data. On the other hand, only a minority of the approached mental health professionals invited to participate completed the survey (one out of five) and we were unable to identify those who did not, nor could we make any comparison between the responding and non-responding groups. This could have led to selection bias with the sole participation of professionals with strong polarised feelings towards coercion. However, the answers were relatively heterogeneous and showed a broad spectrum of perception among participants. Finally, as our study was limited to sole Department of Psychiatry of Lausanne University Hospital, any generalization of our results must be considered with caution. Another issue regarding the generalization of our results is the fact that professionals may have different feelings and attitudes depending on the type of coercive measures examined (Kinner et al., 2017). For the sake of clarity, however, we decided to ask participants about coercion in general. Therefore, a questionnaire on a specific measure such as seclusion could have led to different results. Note: The questions with factor loading ≥0.4 are in grey. Rotation method: Oblique Quartimax. * p < .05.

Conclusions
As more and more voices recommend less coercion in psychiatry, many mental health professionals still remain convinced that it is an indispensable tool which could benefit the patients. Our study highlights this ambiguity. It further shows that the personal as well as the professional proximity with people suffering from mental disorders influence professionals' feeling and attitudes towards coercion. On the one hand, to have a relative with a mental disorder may induce stronger emotions among professionals when they face coercion. On the other hand, the more professionals use coercion the more they seem to become accustomed to it. In the absence of rational evidence about the benefits of coercion in psychiatry, further research should investigate professionals' feelings and attitudes leading to its use. Clinical approaches that enhance shared decision making and thus alleviate stress among patients and professionals should be fostered and developed. Moreover, further studies should give the opportunity to patients and professionals to share their experience and feelings towards coercion in order to develop joint models of care taking into account all points of view and expectations.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement
The data that support the findings of this study are available from the corresponding author, Stéphane Morandi, upon reasonable request.

Declaration of Competing Interest
None.