Participants

HPs including GPs, GP assistants, nurses, community mental health team workers and psychotherapists were recruited from seven primary care regions and two NHS Trusts providing community mental health services in the South of England between January and May 2017. GP practices and individuals were recruited via email and were invited to return a reply slip. HPs who expressed an interest were invited to take part in one of four focus groups taking place in the South of England between March and May 2017. Twenty-one sites returned a reply slip, with 38 participants taking part in either a focus group or interview (22 females, 12 males, and four participants for whom demographic information was not provided). The reported range of years since qualified was 8–34. Focus groups were chosen over individual interviews to allow participants to exchange views on the topic thereby providing topic rich data as well as an insight into group and individual views, including important areas of consensus and disagreement. Individual interviews were offered to psychotherapists as this group was under-represented in the focus group sample (n=2) and to one GP in order to pilot the topic guide. Every participant was taken through the informed consent process and given the opportunity to read the information leaflet and ask questions prior to data collection. Each focus group had between 7 and 10 participants, and the length of each ranged between 43 and 59 min.

Patient and public involvement

Patient and public members of the REDUCE team were involved in discussions about the design and recruitment for this study, and were invited to comment on initial drafts of the topic guide.

Focus groups

A topic guide was developed based on the main aims of the study (online supplement 1). This guide was developed based on a review of existing literature and discussion within a team of academics, GPs, psychiatrists and patient contributors. Topics explored long-term antidepressant use and knowing when discontinuation may be appropriate, negotiating the decision to discontinue antidepressants with patients, HP roles in supporting and negotiating appropriateness of discontinuation, optimising discussions about possible discontinuation and optimising implementation of a discontinuation intervention in routine practice. NPT24 informed the topic guide so that the questions addressed the processes involved in antidepressant discontinuation with regard to the four NPT constructs (coherence, cognitive participation, collective action and reflexive monitoring). For example, to address cognitive participation (ie, who does the work), participants were asked ‘What do you see as your role in negotiating medication discontinuation’. The topic guide was not limited to discussing depressive disorders and therefore was open to discussion about antidepressant use in other conditions (eg, anxiety and chronic pain).

The focus groups were conducted face-to-face and were organised pragmatically across different geographical locations (in GP practices and a community-based health centre) in the South of England. Two groups were held with mixed primary care HPs and two groups with GPs only (see table 1). To acknowledge potential ‘group’ effects (ie, participants being unaware of the degree to which other group members’ views represent their own experience), free participation was encouraged by the facilitators by avoiding censorship and conformity.25 Focus groups were facilitated by two experienced female qualitative researchers (SJW and WOB) and were audio recorded. A debriefing was conducted by the two facilitators following each focus group to identify issues that may affect analysis (eg, domineering or quiet members) and suggest possible modifications to the topic guide. No repeat interviews or focus groups were conducted.

Interviews

Three semistructured face-to-face qualitative interviews were conducted with two psychotherapists and a GP. The same topic guide that had been developed for the focus groups was used in the interviews to ensure consistency. As with the focus groups, the interviewer explored additional topics when brought up by the interviewee. Interviews were carried out by an experienced qualitative researcher (SJW) and were audio recorded.

Analysis

All focus groups and interviews were transcribed verbatim. Transcripts were read and re-read by SJW both during and after the data collection period. While the focus groups and interviews were taking place, the REDUCE Study team met regularly to discuss topics raised by participants, and the topic guide was refined as the focus groups and interviews progressed through debriefing with the two facilitators and through meetings with the wider research team. These discussions resulted in only minor changes regarding the order and wording of questions.

A thematic analysis approach was used to analyse data drawing on methods of constant comparison.26–29 SJW independently coded the seven transcripts using NVivo, and a secondary analysis team (SJW, AWAG, HMB, GL and TK) met to agree a preliminary coding frame, which was then agreed by the whole team. HMB independently coded two transcripts using the coding frame; discrepancies were minor and changes were made following discussion with the team. Codes were grouped into themes by SJW and HMB, where both within and between-participant variation was considered. Theme labelling and interpretation were continually discussed in regular team meetings. Data were assessed for saturation by SJW individually and across-group.30 Data saturation was determined when no new codes were emerging.

Theme 1: who is responsible for broaching the subject of discontinuation?

There were differing views about who is responsible for raising the topic of discontinuation in a consultation. A small number of HPs suggested it was the patient’s responsibility to broach the subject and that this expectation should be set when antidepressants are first prescribed.

I tend to say to people, ’Look, when you start it, I’d like you to continue for at least six months after you’ve felt well', and then right at the outset, I put the responsibility over to them and say, ’Look, one of the things about depression is that you lose control and the worst thing is to come to see the doctor and the doctor takes over control. So, as far as I’m concerned, you’re in control of these tablets and it’s your choice as to when you want to stop it but usually the recommendation is six months after you’ve been well'. I think most people—I haven’t audited it—at that stage, do come back round about six months-ish and are keen to stop and usually that works okay. (GP/09/0002)

One GP argued in favour of telling the patient at the initial prescription that they have the responsibility to initiate stopping and the choice to discontinue is up to them. By setting this expectation, it opens up the possibility of them taking control by broaching the subject with their GP when they are ready.

However, HPs highlighted there are problems with relying on the patient to broach the issue. Two nurses and a psychotherapist acknowledged that many patients may not instigate these conversations. One psychotherapist explained that patients may be reluctant to broach the subject due to expectations of how the doctor may respond, or perceiving the doctor to be more knowledgeable about the situation.

Even if they do get an appointment, I’ve met a lot of people who are really hesitant about asking about changes in medication, because of the response from the doctor perhaps or their perceived response… I think there’s often a worry, you know, the kind of, ’Doctor knows best, and they put me on this medication. So I don’t want to offend or I don’t want to question’. (PT/14/0002)

One GP suggested that when patients do not raise the idea of discontinuation, practitioners may assume that the patient wants to continue treatment. This mutual assumption that the HP wants the patient to continue, and that the patient him/herself wants to continue, may result in a form of collusion to maintain the status quo.

HPs appeared to be aware that relying on the patient to initiate discussion may be problematic, as evidenced through the way some HPs discussed the problem. It may therefore follow that the responsibility to initiate discussions about discontinuation should lie with the GP.

I think I’m guilty of this, it’s very easy just to keep kicking the can down the road and the patient keeps taking the medication because they feel they should and you keep prescribing it because you assume they still want it and there’s this kind of collusion that unless you actively intervene and say, ’Come and talk to me', or whatever. (GP/11/0001)

Some participants (including GPs) thought that it was the GP’s responsibility to broach the subject with patients; arguing that the person who prescribes the medication should be the person to initiate discussion of discontinuation. This was especially the case if a patient has been on the medication long term and may not have considered stopping. However, taking a proactive approach was not always considered feasible in practice; continuity may facilitate discussions of antidepressant withdrawal, although it is not always possible in primary care.

Some of the HPs referred to the discussion of discontinuation as a shared decision process. They talked about the need to assess the patient’s capacity in making decisions about withdrawal and negotiating with the patient to come to a shared decision about whether to discontinue. One GP also suggested it must be a shared decision as GPs are currently unable to manage the amount of work involved due to organisational factors which make having these conversations more challenging.

We’d say that because these patients are working and living in the community and are not sectioned and have capacity, that there is definitely a shared responsibility with the patient because it is their medicine and their mental health that we’re looking after. So I’m quite happy to say it’s a shared responsibility but it definitely can’t be just a primary care clinician’s because we’ll not manage to cope. (GP/19/0002)

The role of other HPs was also discussed with regard to conversations around discontinuing antidepressants. Although nurses have been considered to play a role in these discussions, there is acknowledgement that there are limitations regarding their authority and experience in managing medications, and often patients are signposted to their GP by their nurse. Social workers, pharmacists, care coordinators and psychiatrists were also mentioned as potential sources of additional support in stopping antidepressants, in some cases.

Theme 2: risk of destabilising current situation

Some HPs described it being easier to continue prescribing rather than raising discontinuation with patients and acknowledged a need to initiate more discussions about discontinuation with patients who may be eligible. There were concerns about instigating discussions with patients who are currently well as they did not want to risk destabilising the current situation. It was considered less risky to continue prescribing.

I think about—not just to patients but also to healthcare professionals or GPs—to reduce the medication is the concern that they might be working and reducing might destabilise a current stable situation, especially if the patient has been very, very difficult to control in the past and hasn’t got the support network perhaps. (GP/12/0001)

There was an assumption that patients also do not want to risk upsetting the current situation if they are feeling well.

I think for a lot that are on them, there is a massive fear factor about stopping, because they remember how awful they felt. They don’t want to feel like that. They feel well again and they just think, well, you know, I’d rather just keep the status quo. (GP/03/0004)

Theme 3: continuity and knowing the patient makes it easier to discuss discontinuation

Involvement in the initial prescription was perceived to place responsibility on the prescriber to prompt discontinuation later, and an opportunity to discuss and set patient expectations around withdrawal. Explaining to a patient at the initial prescription that discontinuation will be discussed at a later date was seen to be a facilitator in broaching the subject of discontinuation.

The very first consultation if you’re actually selling the idea of some medication being helpful, that it’s for a specific time period, expecting someone to be able to be able to come off it at about six months, so suggest your timescale of appointments and then say, ‘Oh, see you in about five months from the initiation of treatment. At that point, we can actually make a plan for withdrawal and I would be planning to withdraw it slowly if everything was going well in your life’. (GP/11/0004)

There were a number of facilitators to discussing discontinuation with patients. These included knowledge of the patient’s experience with antidepressants, their triggers for depression, why they started their medication and how things have changed since the initial prescription. This again suggests that continuity is beneficial, especially in terms of reducing risk.

Theme 4: a HP’s confidence in their skills and knowledge

Some of the HPs reported a lack of confidence, knowledge and skill with regards to antidepressant discontinuation which could act as a barrier to broaching the subject of stopping with patients. There was an awareness that discussing discontinuation with patients is something that could be improved on.

As a GP, I think for GPs, I think we’re very good at starting patients on it. We are good at titrating the dose up. Pretty good at picking the right medications suitable for the patients, because they have different side effects over spectrums. But what we’re probably not good enough, at the moment, is sort of the long-term managing and the coming-off part. (GP/12/0001)

HPs discussed a need for more support and information for themselves as well as for patients. They spoke about NICE guidance on antidepressant discontinuation, with many being unfamiliar with the guidance or not using them. They described being dissatisfied and, in one case, irritated by the current guidance. They highlighted that it is unclear (especially regarding tapering regimes), limited, not accessible and at times not applicable to real patients.

I don’t think there’s a lot of resources out there to kind of what to say and how to do it. I’m sure I’ve looked at the guidelines before and I thought a bit pants. (NP/12/0001)

Theme 5: organisational barriers and enablers to discussing discontinuation

The above processes are shaped by the context surrounding them, with environmental work contributing to decision-making around discontinuation. Some aspects of the healthcare system were described as further barriers to antidepressant discontinuation. A lack of continuity was reported with patients seeing different practitioners each time, and these practitioners were at times providing inconsistent recommendations. This may act as a barrier to discussing discontinuation due to the perceived need to be familiar with a patient to discuss withdrawal, and the idea that the responsibility for raising the topic of discontinuation lies with the HP who initially prescribed the antidepressant.

HPs repeatedly noted the challenge of time constraints in practice and how this is often a barrier to both initiating and managing discontinuation due to ten minute consultations not being long enough, and not having the time for review appointments.

Things are ticking along relatively okay, you know it’s not going to be necessarily a straightforward consultation and it might be time consuming, it might delay you and you haven’t got enough appointments anyway and da, da, da, da, you can see how that, as a clinician, restrains you from perhaps rocking the boat. (GP/11/0002)

HPs also mentioned the role of computer systems, explaining that patients can get lost in the system and that systems which adequately prompt medication reviews would be useful in broaching discontinuation with patients.

Strengths and limitations

This study is the first to explore HP’s perspectives of antidepressant discontinuation in UK primary care, with its larger sample consisting of a range of HP roles (including GPs, GP assistants, nurses, community mental health team workers and psychotherapists) which were lacking in previous research,20 21 33 40 41 and data reached saturation. GPs were the largest group among our interviewees, compared with the other professionals, which aligns with the current prescribing activity with the large majority of long-term antidepressants prescribed and monitored by GPs. This fits with our finding that GPs are often considered responsible for initiating conversations around withdrawal. However, we also identified that there are a number of professionals who may be involved in discontinuation (eg, pharmacists, social workers and care coordinators) and further research may be needed to explore these perspectives. For example, none of the practices in the current study managed discontinuation using practice pharmacists, who may play an important role in antidepressant withdrawal. In particular, it may be of interest to explore differences between professions.

The use of focus groups facilitated discussion and provided candid responses from participants. However, it is possible for discussions to become polarised or influenced by dominant members of the group. For example, in a focus group of nine GPs, there were two more dominant members and two members who spoke less frequently. As such, some participants’ views may be less well represented in a group setting. Giving participants an opportunity to provide feedback on the study’s findings might have helped provide greater representation.