‘Triadic’ shared decision making in mental health: Experiences and expectations of service users, caregivers and clinicians in Germany

Abstract Background Shared decision making (SDM) in mental health may contribute to greater patient satisfaction and is sometimes associated with better health outcomes. Here, SDM should not only involve service users and clinicians but also involve the service users' caregivers. Aim This study aimed to achieve better insight into the current SDM patterns of triads of service users, caregivers and clinicians in inpatient mental health care and the three parties' expectations towards the prospects of triadic SDM. Design The current research uses data from a representative cross‐sectional study on caregivers in psychiatric inpatient treatment. We analysed data on n = 94 triads of service users, their caregivers and their clinicians. Results All three parties acknowledge caregivers to be of great support to monitor the progress with mental disease. The caregiver's role during consultations is most often described as being an expert, receiving or providing information and supporting service users. However, caregivers at times try to seek support for themselves during caregiver‐clinician interaction, or their behaviour was described as unhelpful. The potential prospects of caregiver involvement are clearly acknowledged despite the low implementation of caregiver involvement in this sample (only in one‐third of the cases). Conclusion Triadic SDM rarely takes place in routine inpatient care. First, there should be a focus on interventions aiming at inviting caregivers to consultations. Only in the second step should a better conceptualisation of triadic SDM be undertaken. Public Contribution Early results were discussed with a local peer support group for caregivers of individuals living with mental illness.


| INTRODUC TI ON
Clinicians and patient representatives increasingly witness shared decision making (SDM) with service users experiencing mental health conditions as an ethical imperative, 1 which may contribute to a greater service user satisfaction and is sometimes associated with better health outcomes (eg improved adherence to treatment and fewer relapses). 2,3 Shared decision making in general medicine 4 and even more in mental health care should-on top of service users and clinicians-involve the service users' caregivers. 5 In this regard, family and non-family (eg friends and neighbours) caregivers are often referred to as 'informal carers' (also referred to as 'caregivers'). Caregivers take on responsibility in areas that are insufficiently covered by health-care professionals (eg monitoring medication and improving compliance) 6,7 and play a decisive role in coping with everyday life (eg finances, housing and social contact). 8,9 Consequently, caregivers are at a high risk of experiencing health, emotional and financial burdens themselves. 10,11 It is encouraging that caregiver involvement in psychiatric treatment does not only positively influence the service user's course of the illness but also improve the caregiver's health and well-being. [12][13][14][15][16] However, most SDM approaches do not explicitly involve the service user's caregivers. 17 There are various reasons for this pattern, such as a lack of co-operation between clinicians and caregivers in general 18 and the worry that an inclusion of a third-party may make SDM even more difficult. 17 Moreover, there are insufficient data on how caregivers may be integrated into SDM. Therefore, the existing frameworks for the implementation of SDM (eg Ref. 19) do not yet offer concrete guidance for including third parties into SDM.
This study aimed to achieve better insight into the current SDM patterns of triads of service users, caregivers and clinicians in inpatient mental health care and the three parties' expectations towards the prospects of triadic SDM.

| ME THODS
Data of this study stem from a large, representative cross-sectional study on caregivers in psychiatric inpatient treatment in psychiatric hospitals in Germany. 18 In this study, n = 247 inpatients and their clinicians were interviewed. We also aim to address all service users' caregivers, which was possible in 94 cases. This study aimed to gain deeper insight into triadic SDM patterns of service users, caregivers and clinicians in inpatient mental health care and the three parties' expectations towards the prospects of triadic SDM.

| Recruitment of participants and data acquisition
Data were collected for 10 months (October 2018 to August 2019).
Recruitment took place on all wards of the participating hospital, except from wards with an emphasis on elderly psychiatry (65+ years) or alcohol/drug dependency. In each ward, clinicians were first asked for their next two to three service users pending discharge. These service users were then invited to be interviewed.
Subsequently, all treating clinicians and, when possible, the service users' caregivers were interviewed. The caregivers were named by the service users and could be of any relation to the service users (informal carer). All structured interviews were performed face-to-face asking closed, semi-open and open-ended questions. In a first analysis of the data, service users and clinicians consistently reported that contact between the caregiver and the clinician in charge took place in only one-third of the cases. The most important predictors for clinician-caregiver contact included the service user's diagnosis (eg schizophrenia) and the treating hospital. 18 We used data on n = 94 triads of service users, their caregivers and their clinicians for the present analysis. In particular, participants' answers to the following questions (displayed here is the service users' version of the questionnaire) were studied: • 'What role does your caregiver play, when it is about your illness?' • 'Has there been contact between your caregiver and the clinician in charge during your inpatient stay?' (If yes) 'What role did your caregiver play during this contact?' • 'Did you take action to make contact between your clinician and your caregiver happen?' • 'Were there any serious disagreements between your caregiver and your clinician during your inpatient stay?' • 'What is (or could be) the benefit of an involvement of your caregiver?' • 'What should this involvement of the caregiver ideally look like?' All answers to these open-ended questions were written down and later categorized to allow a descriptive analysis. For all participants, sociodemographic and clinical data were available.

| Statistical analysis
In a first step, we inductively created categories based on the answers to the open-ended questions. Two judges independently rated 30 answers as to whether these categories were present or not for each open question. Cohen's kappa coefficients (κ) ranged from 0.89 to 0.91 and thus indicated an excellent inter-rater reliability. 20 In the second step, these categories were then coded for all participants, and descriptive statistics was performed via SPSS Statistics 25 (eg frequencies, mean values).

| Ethics
An ethical and legal review of the study was done by the local ethics committee. The ethics committee raised no objections to the conduct of the study.

| Participants
We included 94 triads of service users, caregivers and clinicians in this study. The original sample of 247 service users and the subgroup of 94 service users comprised in the triads did not differ substantially in age, gender, diagnosis and severity of illness. Table 1 shows the sociodemographic data of service users.

| What role do caregivers play in the context of psychiatric disorders?
The most often cited category here was 'support', which was reported by service users, caregivers and clinicians in 98.4%, 92.6% and 66.7% of the cases, respectively. This category was further subdivided into 'emotional support', 'support in everyday life' and 'support in coping with the illness'. Service users, caregivers and, to a lesser extent, clinicians stated that 'emotional support' is the most relevant aspect of support in general. However, clinicians attributed a negative or detrimental role to 36% of the caregivers. They described the caregivers' influence as either causing or sustaining the mental illness, or perceive the caregiver to be mentally ill themselves. An overview of the different roles caregivers played according to the participants' replies is provided in Figure 1

| What role do caregivers play during service user-caregiver-clinician contact?
Many service users could not provide an answer (35.7%) regarding the caregivers' role during service user-caregiver-clinician communication. The most frequently reported answer from the service users' view though was that caregivers were seen as experts in the disease (25.0%) and seeking information (25.0%). Caregivers themselves most often reported that they wanted to acquire information (46.4%), followed by the self-perception as an expert in the disease (42.9%). The clinicians' perspective differed substantially. They most frequently described the caregivers' role as giving the service user a feeling of security and offering support (30.0%), followed by the caregiver providing important information (25.0%). Clinicians mentioned inappropriate behaviour of the caregiver in 17.5% of the cases ( Figure 2). Two more sets of categories were derived from the participants' answers. First, we categorized whether the caregivers primarily wanted to express their own opinion during caregiver-clinician interaction or if they wanted to make the service user's voice heard.
Here, all three parties more often reported that the caregivers wanted to express their own opinion rather than the service user's opinion (28.6% service users, 50.0% caregivers and 32.5% clinicians). Second, we assessed how frequently caregivers were seeking help for themselves during meetings with clinicians. This topic was mentioned by clinicians and caregivers in approximately 20% of the cases ( Figure 2).

| What is (or could be) the benefit of an involvement of the caregiver?
All three parties (service users, caregivers and clinicians) most often stated that caregiver involvement is beneficial and improves therapy

| What should involvement of the caregiver ideally look like?
Although respondents gave a wide range of answers, some major categories regarding initiation, extent, timing and setting of involvement could be identified. for caregiver-clinician interaction through telephone or writing. Figure 4 shows the details concerning expectations towards the ideal caregiver involvement.

| D ISCUSS I ON
Caregivers are viewed as an important part in supporting their relatives dealing with mental illness. Their role is often described as being an expert, receiving or providing information and supporting service users during consultations with clinicians, which seldom take place. However, caregivers were sometimes seen as seeking support themselves or as acting unhelpfully. The potential prospects of caregiver involvement are clearly acknowledged despite the low implementation of caregiver involvement in this sample. Some guidance as to how involvement could best take place is given.

| Strengths and limitations
This investigation is one of the few looking into service user-car- We ended up with 94 triads (out of 247 service user cases) in our study despite our best efforts. It is conceivable that the caregivers interviewed represent a subgroup having better relationships with service users or showing more interest in the service users' course of disease than average. We had to limit our selection to the 94 cases in which a complete triadic data set was available (service users, caregivers and clinicians) because we wanted to explore the role of the caregiver from the perspectives of all three parties involved. This approach could lead to biased results.
Participants generally had difficulties to answer our questions or remained rather vague with their answers during the investigation. This is probably due to the fact that many participants have not given sufficient thought about the caregivers' role in matters of SDM or are strongly influenced by the existing hospital routines.
Another limitation is that we were not able to observe conversations between service users, caregivers and clinicians but only received reports about them.

| The findings in the context of shared decision making
What should be made based on preferences. In our opinion, the findings of the current study can be attributed best to the first phase of the three-talk model. Here, it should be discussed who is involved in the process of SDM. All three parties (service users, caregivers and clinicians) agree that the initiative to involve a caregiver must come from the clinician at this stage according to our investigation. This is especially important as other research shows that caregivers want to participate in SDM and typically feel excluded. 21 Hamann and Heres 17 further proposed that the caregiver's role should be clarified in the first phase. Our investigation shows that caregivers not only want to acquire information but also want to be perceived as an expert on mental disease. Service users who were able to answer this question also mentioned these two categories most frequently.
Furthermore, the first phase of the three-talk model should be used to clarify whether the caregiver wants to be involved in person, by telephone or via the Internet. Quotes from participants did not significantly contribute information to phases 2 (option talk) and 3 (decision talk) of the three-talk model, so no new conclusions for them could be drawn.

| CON CLUS ION
It was the aim of the current investigation to gain a deeper insight into SDM patterns of triads of service users, caregivers and clinicians in inpatient mental health care and the three parties' expectations towards the prospects of triadic SDM. It is hardly possible to derive a concrete concept of triadic SDM from there because the participants' answers remained rather vague and the contact between service user, caregiver and clinician rarely took place. In our view, the first step should be to focus on interventions that aim at inviting caregivers to consultations and only in the second step should a better conceptualization of triadic SDM in mental health be undertaken.

ACK N OWLED G EM ENTS
We would like to thank all participating clinics: kbo-Lech-Mangfall- Access funding enabled and organized by ProjektDEAL.

CO N FLI C T O F I NTE R E S T
Dr Hamann reports grants from Janssen-Cilag, Germany, during the conduct of the study, and personal fees from Janssen-Cilag, Germany, Otsuka and Lundbeck, outside the submitted work. Dr Heres reports grants from Janssen-Cilag, Germany, during the conduct of the study, and personal fees from Janssen-Cilag, Germany, Otsuka and Lundbeck, outside the submitted work. Dr Holzhüter and Mr Schuster have nothing to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.