Study design, research team, and reflexivity

The study operated within a critical realist perspective, in that participants’ comments were assumed to reflect an underlying reality that is mediated by perceptions, beliefs and social processes, and consideration is given to how broader cultural processes shape knowledge and practice, both in the clinical field of study, and in the research processes we undertook [35,36]. The study was part of the ‘RADAR’ programme (Research into Antipsychotic Discontinuation and Reduction, National Institute for Health Research (NIHR) Programme Grant, see: https://www.ucl.ac.uk/psychiatry/antipsychotic-discontinuation-and-reduction). It was conducted as preparation for a clinical trial investigating whether reducing and potentially stopping antipsychotics with the support of prescribers improves social functioning compared with maintenance treatment for people with schizophrenia spectrum disorders. All authors are involved in the RADAR study and are academic researchers, with backgrounds in psychology, psychiatry and social science. JM is also a consultant psychiatrist and prescriber of antipsychotics. We aimed to work reflexively in acknowledging the epistemological and personal positions that we brought to the research process and our interpretations of the data.

Setting and participants

Data were collected in mental health clinics in two London-based NHS Mental Health Trusts. These trusts serve socio-demographically mixed populations that include pockets of high deprivation and ethnic diversity in both urban and suburban areas. They provide inpatient and community-based outpatient mental health services for people with a range of diagnoses including psychosis, bipolar disorder, personality disorder, depression, and anxiety. These mental health services are multidisciplinary including psychiatrists, mental health nurses, social workers, support workers, and psychologists. In the community, psychiatrists will often have 4–6 monthly consultations with patients where medication will be discussed. Other members of the team such as nurses or social workers may see patients more regularly if necessary, providing emotional and more practical support, such as help with housing or employment.

Seven focus groups were conducted between July and October 2016 with each group consisting of 3–8 participants. Participants were recruited from community-based mental health services and consisted of prescribing and non-prescribing mental health practitioners who were psychiatrists, mental health nurses, social workers or clinical team managers. All participants had experience of working with people with psychosis. We recruited from mental health teams that we aimed to involve in the RADAR antipsychotic reduction trial. Participants were invited to participate in the focus groups via email invitations and groups were conducted with all who took up the invitation. It was not possible to ascertain who declined participation, as some people did not reply to these requests which may have been due to time constraints or missing the email rather than actively declining to take part.

Data collection

A topic guide was developed which explored four main areas (see S1 Appendix for the full topic guide):

• Experiences and views of reducing and discontinuing antipsychotics.
• The process of antipsychotic reduction and discontinuation, including perceived facilitators and barriers, and service user suitability.
• Detection and definition of relapse.
• Forms and sources of support for antipsychotic reduction.

Groups were facilitated by REC & ÉH and their involvement in RADAR was known to participants. Some participants were also known to the researchers due to participation in other research projects or clinical work. Focus groups consisted of participants and facilitators only. Where possible, separate groups were conducted for psychiatrists and other clinical team members, in order to facilitate open discussion from the perspectives of prescribing and non-prescribing practitioners. Participants were also asked to complete a short demographics questionnaire. Written informed consent was obtained from all participants before data collection, and discussions were audio-recorded, transcribed and anonymised. Each group lasted approximately 60 minutes. Field notes were made by facilitators during the focus groups.

Data analysis

Data were analysed using thematic analysis with NVivo software (Version 11) [37] by the researchers who facilitated the focus groups (REC & ÉH). In order to enhance validity, regular meetings were held with the two other authors (NM & JM), in which emerging codes, themes and patterns of similarity and difference across the data were critically discussed. The first three transcripts were independently coded by each researcher in order to identify key themes. These were then combined and compared to develop an initial coding framework, which was then used to code further transcripts. Codes were progressively refined and organised into emerging themes and sub-themes, and analysis continued until a coherent account of the variations in the data were met. Reflections and field notes from the researchers who facilitated the focus groups were part of the collaborative analysis process, and informed our understandings of the data. We were aware of the busy working lives of our participants, so did not attempt to do member checking or obtain feedback on our interpretations from participants.

1. Attitudes to antipsychotics and antipsychotic reduction

In all the focus group discussions, participants acknowledged that antipsychotics can have severe adverse effects, and some criticised what they viewed as overreliance on medication in services. Only a minority of psychiatrists had a completely positive attitude towards long-term antipsychotic treatment, although everyone acknowledged that they had some benefits.

2. Reduction and discontinuation of antipsychotics in practice

3. Barriers to reducing antipsychotics

Alongside experiences of adverse events meaning that reduction was approached with caution, attempts to reduce antipsychotics were often described by participants as hampered by significant organisational and knowledge barriers.

4. Communication and negotiation

The importance of open communication between clinicians and service users, particularly in regard to medications, was stressed in all focus groups. Clinicians described how treatment decisions around reduction or discontinuation were made through a process of negotiation between the service user and clinician, often after the service user had made a request to reduce. Related to the views described earlier about long-term antipsychotic use and risks associated with reduction and discontinuation, some psychiatrists reported that they never suggest reduction or discontinuation of antipsychotics, and that this was only ever raised by service users. The importance of respecting and engaging with these requests to reduce was acknowledged by most participants. Some emphasised how they engaged with such requests. Some participants articulated a process of weighing up potential costs and benefits. In supporting requests to reduce medication, they described being motivated by a desire to sustain their relationship with the service user, or to avoid disengagement from services, and/or the risks of unsupported discontinuation.

“I think so um, some people are susceptible to quite marked side-effects, especially weight-gain and you know my thinking is it’s better for somebody to take a smaller dose rather than take a higher dose and, and stop it–not take it–because of the side-effects. [Agreement from other group members] So sometimes it’s worth considering is it better to have a smaller dose and that person will better [several agree] adhere to the medication rather than having a too high, or a high dose and the side-effects will just like make them stop, you know?” P24, Clinical Team Manager, Group 5.

“I think it’s good for your relationship with people as well. If you just sit there going no, no, no, NO, NO, no [h] we’re not reducing, no[h]. You know, if you actually, people feel that you know you’ve said to them, you said well look, we probably can’t stop this, but let’s have a look–let’s try a small reduction and see what happens, if that goes okay then we can, you know, and if you can feel you’re working with them, it improves the relationship as well.” P21, Psychiatrist, Group 5.

A small number of psychiatrists felt that their professional knowledge gave them greater decision-making authority than the service user. These psychiatrists described adopting a more paternalistic stance, and in some cases, using a range of strategies to discourage requests to reduce or stop taking antipsychotics. These included relaying the statistical probability of relapse after discontinuation, and drawing parallels between psychosis and chronic physical health conditions such as diabetes.

“But that’s the group of people I think who, who are pressing [to reduce]…well maybe I’m wrong 'cos that’s all I see, but you know, they’re the ones that haven’t yet learnt their lesson of if they stop their medication they relapse” P17 Psychiatrist, Group 4.

“we go to great lengths to persuade patients to keep on taking them [antipsychotics] despite side-effects, and possible risks because we think that that’s in their best interests of staying well and having a chance of recovery…” P19, Psychiatrist, Group 4.

“…So I start talking to them about diabetes, COPD [Chronic Obstructive Pulmonary Disease] and epilepsy and said if you had epilepsy we wouldn’t be stopping your medication. You would accept that psychosis is a long-term condition like COPD and diabetes and epilepsy and actually you can’t stop your medication.” P34, Consultant Psychiatrist, Group 7.

Across all focus groups, relationships with families and carers were acknowledged as often providing vital information or support during the reduction process. For many clinicians, families were crucial in spotting early signs of relapse, and were often tasked with managing any initial return of symptoms. However, because of the caring role that families took, participants felt that some families were apprehensive about antipsychotic reduction due to fear of relapse. This could cause conflict if the family and service user had different views on treatment, and clinicians were tasked with managing families’ anxieties.

“Yeah, and they [the family] were quite adamant that they didn’t want this to happen [reduction], that, you know, they were even mentioning things: oh, you're gonna have to section her and, you know, that kind of line.” P27, Mental Health Nurse, Group 6.

Clinical and research implications

In the current study, antipsychotic reduction was perceived to be hindered by significant organisational barriers including lack of resources, drive to discharge service users to primary care, and poor continuity of care. These systemic barriers are concerning as they may be hindering good practice. For example, pressure to discharge means that service users may be stabilised quickly through high dose antipsychotics and then discharged to GPs (primary care physicians), who are unlikely to alter this medication [38–40]. This situation reduces patient choice and may result in the long-term use of potentially unnecessarily high doses. A lack of resources, such as clinician time for the additional monitoring and support that safe reduction may require, and limited access to alternative psychosocial treatments means that achieving patient stability through medication is crucial for services to operate effectively. This may mean that maintenance treatment is favoured over reduction for organisational rather than clinical reasons, and that service users may miss out on opportunities to minimise the adverse effects of long-term treatment, or to manage their illness in other ways.

Despite advocating antipsychotic reduction to the minimum effective dose as good clinical practice, clinicians were often reluctant to reduce, meaning that, in line with prior research [30], initiation of reduction processes were often driven by requests from service users. However, research shows that service users are highly aware of unequal power dynamics in medical consultations [25,40,41], and may be reluctant to ask for antipsychotic reduction even when they want this, preferring the suggestion to come from the psychiatrist first [19]. They may expect clinicians to make decisions for them, and be afraid to question the authority of the doctor in the desire to be a “good” patient [41]. These dynamic processes in the medical encounter may go some way to explaining why service users perceive a lack of choice about mental health treatments [19,42,43]. Reluctance of professionals to initiate or consider the possibility of antipsychotic reduction or discontinuation is likely to further deter service users from meaningful participation in decisions about their treatment. Service users may become or remain passive or compliant partners in the decision-making process, unable to explore the choices they would prefer in a meaningful way. This situation may also reinforce chronicity and hinder recovery by reducing service-users’ options for managing their mental health problems [26,44]. Although individual clinical encounters take place in the context of a general structural framework of power imbalances within mental health services, encouraging service users to take a more active role in consultations both helps individuals to feel more confident in putting forward their preferences, and starts to address the wider culture that impedes this. Factors that could help to enhance power in individual consultations include a reduction in coercive practices (discussed below), offering service users both drug and non-drug treatment choices such as psychosocial treatments [45], and having frank discussions about the pros and cons of these treatments and expected outcomes [46].

Clinicians’ reluctance to reduce antipsychotics appeared to stem from risk aversion or sometimes scepticism of the feasibility of reduction. Risk aversion stemmed from prior experiences of often serious adverse events and relapse, following reduction or discontinuation. These included suicide, criminal offences, loss of housing, relationships, and jobs. Clinicians’ feared they would be held professionally accountable for such outcomes, especially given the lack of official guidance for antipsychotic reduction. These concerns about risk are in line with clinician attitudes reported in other studies of antipsychotic reduction [31,32]. A small number of psychiatrists were more sceptical of the feasibility of reduction, adopting a more traditional paternalistic stance and believing there was never good justification to try and discontinue antipsychotics and reporting a range of strategies, including relaying statistics, to persuade people to stay on antipsychotics.

These sceptical attitudes may relate to issues of risk aversion previously discussed, but they may also stem from the widely accepted medical model of psychosis itself—that suggests that antipsychotics correct a ‘chemical imbalance’ and are therefore necessary for recovery [47]. Even though this model is the subject of considerable debate [48,49], for many psychiatrists it is regarded as established fact. Previous research has found that over-reliance on a scientific model of treatment along with potentially coercive practices, can undermine the trust, choice and power service-users have in their mental health treatment [46]. It has been suggested that coercion in mental health practice could be seen as a form of epistemic injustice, where clinicians do not see the service-users’ knowledge and decision making ability as legitimate due to the beliefs they hold about their diagnosis [50]. For example people with schizophrenia may be viewed as out of touch with reality, diminishing the legitimacy of their knowledge and requests, and increasing the possibility of more coercive or controlling treatment [50,51]. These attitudes could form the context of individual clinical decisions. In our data, for example, service user preferences often appeared to be trumped by scepticism and risk aversion which favoured the stability associated with maintenance treatment. Against this background, increasing the involvement of service users in treatment decisions, as recommended by national guidelines [5], is challenging, and is likely to require changes in wider cultural attitudes within services. Genuine shared-decision making in this context will require an openness to discussing antipsychotic reduction [25–28]. One way of enabling this would be through the development of formal guidelines for reduction and discontinuation, to help psychiatrists feel safer and more supported when reducing antipsychotics.

It was clear from our focus groups that mental health professionals, particularly psychiatrists, felt they needed more evidence-based guidance regarding whether, how and with whom antipsychotics can be reduced or stopped safely. Research to date has failed to provide any consistent predictors of successful reduction or discontinuation [52,53]. This absence of evidence and guidelines may discourage clinicians from considering reduction of antipsychotic treatment as standard, as our participants’ descriptions of treatment decisions and discussions with service users suggest.

Current guidelines recommend family involvement as best practice, however service-user—family/carers views on antipsychotic reduction were often described as conflicting, hindering the implementation of this recommendation and also shared decision making. Families and carers have an important role in supporting service-users but were described as sometimes being concerned or resistant to antipsychotic reduction, with psychiatrists and clinical teams having to negotiate reduction and manage the families concerns. This may be due to the responsibility family and carers feel for antipsychotic compliance, leading them to take a more controlling role in medication decisions [42]. The identified organisational barriers could be linked to these issues. Under-resourced services may mean more responsibility is placed on family and carers, with clinicians finding it difficult to provide reassurance that the appropriate increased support will be available when reducing. Rapid discharge from secondary services and lack of continuity of care may make it difficult to build trusting relationships and good communication with families and carers which would help to allay concerns. Clinical guidelines should provide information on how to manage conflicting demands from patients and their families or carers and how best to address the latter’s legitimate concerns.

Strengths and limitations

To our knowledge this is the first study to specifically investigate the views of a range of mental health professionals working in secondary care on reducing and stopping antipsychotics. Psychiatrists, psychiatric nurses and other professionals from a range of mental health services that support people with both short-term and longer-term mental health problems were included, maximising our ability to explore a diversity of views and experiences from different perspectives.

Limitations include the sample consisting primarily of consultant psychiatrists and nurses, and that several focus groups were made up of clinicians who worked on teams together, or knew each other through peer networks. While this may have put participants at ease, they may also have been reluctant to divulge service-related concerns or differences of opinion in front of colleagues or managers. Participants’ were aware of the facilitators’ positionality, i.e. their involvement in the RADAR antipsychotic reduction study. Although no further details regarding researchers’ views on antipsychotics were disclosed, we recognise that this awareness and potential interpretation that facilitators were supportive of antipsychotic reduction could have influenced some views. Participants may have self-censored critical comments, or conversely emphasised contrary views more strongly. We collected data in areas of London with rapid demographic expansion and consequent high pressure on psychiatric beds. This may reduce the applicability of findings to other areas.

Conclusion

While long-term antipsychotic treatment may successfully reduce psychosis symptoms and prevent relapse for some [2–4], it is known to be associated with significant health consequences, and can be subjectively and objectively impairing [8]. Relapse following reduction or discontinuation is not inevitable [23], and some service users reasonably wish to reduce their side-effect burden, or consider options other than continuous maintenance treatment. However, when service users stop antipsychotics covertly and abruptly (which may be more likely if support or options to reduce are not offered) their risk of relapse and negative or coercive service experiences may increase. More research is needed into the outcomes of antipsychotic reduction and discontinuation that is clinically supported and conducted systematically, gradually and responsively. This could inform evidence-based guidelines on reduction and discontinuation processes that are currently not available. Our findings suggest that practitioners would greatly value such guidelines, which could enable them to feel more confident in responding to service users’ requests or preferences to alter or reduce medication. In turn, this would enable service users to be more involved in managing their mental health. In addition, it is important to consider how organisational or systemic factors could enable rather than impede clinicians in these processes. Further support for engaging patients meaningfully and constructively in decision making would be useful, and financial constraints that encourage the prioritisation of short-term stability to enable rapid discharge, with long term high-dose antipsychotic prescribing need to be debated.