At all time-points identical data will be collected in the intervention and control groups (Fig.
2). In our study we will follow an intention to treat (ITT) approach. Therefore we will try to follow up all participants in the 12 months periods following hospital discharge irrespectively of whether they are still in treatment or on medication. We seek for patients’ permission to contact them directly as well as their actual physicians (if applicable) at 3/6/9/12 months. By applying this 2-way-approch we will also be able to approach patients who no longer receive any kind of psychiatric treatment.
Baseline parameters
For all patients enrolled socio-demographics, diagnosis, illness severity (CGI and GAF scores) and data on illness history (previous hospitalizations, duration of illness etc.) will be recorded at baseline (at study entry).
Since decision making and patients’ perception of decision making may be influenced by patients’ participation preferences, the Autonomy Preference Index (API [
25]) will be obtained. The API is a four item questionnaire assessing patients’ general desire to participate in decision making. API scores have been shown to be associated with treatment satisfaction and health care outcomes [
26]. Further, a recently developed questionnaire addressing patients’ competence for shared decision making (PatPart-19) will also be used. This 19 item questionnaire covers the subscales open communication, critical communication and adherence in therapy and was validated in two large samples of psychiatric inpatients (Kohl et al., in preparation). In addition, active patient decision making has been shown to be related to patients’ having less unmet needs [
27]. Therefore, the Camberwell Assessment of Need self-report questionnaire (CANSAS-P), a 22-item questionnaire addressing patients’ needs for support in various domains will be applied. As SDM-PLUS focusses on critical decision processes in which disagreement between patients and physicians may occur we aim to identify patients with potentially reduced insight, allowing for subgroup analyses in patients with or without insight. We will administer the Birchwood Insight Scale [
28], an eight-item questionnaire addressing insight, necessity of medication or treatment. Likewise patients’ perception of the current admission (i.e. the extent to which patients perceived the admission as involuntary) will be recorded using the MacArthur Admission Experience Survey as used by O’Donoghue et al. 2013 [
29].
Primary outcome
The primary outcome parameter is the patients’ perceived involvement in decision making using the SDM-Q-9 questionnaire at 3 weeks after enrollment in the study or discharge (whatever occurs first). Referring to Rodenburg-Vandenbussche et al. [
30] we see a 15 point difference as clinically meaningful (in the mentioned study this 15- point difference differentiated between shared decision making and physician dominated decision making as perceived by patients) and we expect a 15 point mean difference between the intervention and control group (Table
1). The SDM-Q-9 refers to a distinct medical decision which will be the patient’s medication regime, which in many cases is the result of several smaller decisions during inpatient stay and thereby reflecting the decision atmosphere between patients and their treating physicians. Therefore the SDM-Q-9 will serve as a proxy for inpatient decision making.
Table 1
Planned data collection
Secondary outcomes
As outlined above, a better involvement of patients in medical decisions and health care professionals better addressing “difficult decisions” may also result in a better therapeutic relationship, higher treatment satisfaction, fewer unmet needs, more thoroughly implemented therapies, improved adherence, and probable reduced relapse rates (see Fig.
1). These outcomes will be obtained at T1 (three weeks after study enrolment) and during a one year follow up with quarterly assessments (T2-T5).
Patients’ involvement in medical decisions from the patients’ point of view will be obtained as the primary outcome measure at T1 and as a secondary outcome during follow up (T2-T5). In addition, physicians will provide their view on patients’ involvement using the physician version of the SDM-Q-9 at T2-T5 (SDM-Q-Doc [
31]). As most patients will have changes in their treating psychiatrist when switching from inpatient to outpatient treatment, the SDM-Q-9 refers to the inpatient psychiatrist at T1 and to the outpatient psychiatrists at T2-T5.
Whether or not the intervention also has an influence on the therapeutic relationship will be determined using the Helping Alliance Scale at T2-T5, which has a patient (HAS-P) and a clinician (HAS-C) version [
32]. The HAS comprises five items rated on a visual analogue scale ranging from 0 (‘not at all’) to 10 (‘extremely well’). While HAS-P includes items on ‘right treatment’, ‘understood by therapist’, ‘criticized by therapist’, ‘committed therapist’ and ‘trust therapist, HAS-C items cover ‘getting along with patient’, ‘understand patient’, ‘look forward to meeting patient’, ‘feel actively involved’, and ‘feel I can help patient’.
Treatment satisfaction will be measured using the Questionnaire on Patients’ Treatment Satisfaction (ZUF8) at T1, an eight-item questionnaire addressing general satisfaction with hospital care [
33]. The prevalence of unmet needs on the patients’ side after intervention will be assessed using the Camberwell Assessment of Need self-report questionnaire (CANSAS-P) also at T1 and T2 [
34]. For the assessment of adherence, patients will fill out the Medication Adherence Rating Scale at T2–T5 (MARS [
35]). This measure also includes the 10 items of the Drug Attitude Inventory allowing for an analysis of patients’ drug attitudes.
In addition, aspects of patients’ well-being and quality of life will be addressed using the WHO-5 well-being index and the EUROHIS-QOL index at T3 and T5. The WHO-5 index is a five item self-report measure of general well-being and the EUROHIS-QOL is an 8-item self-assessment instrument of generic quality of life. Both instruments were validated in German versions and yielded good to excellent psychometric properties [
36].
Another aspect of interest is whether or not SDM-PLUS supports clinicians in addressing “difficult medical decisions”. This assumption is based on evidence that the discussion of certain decisions (e.g. long acting antipsychotic injectable) is often avoided by psychiatrists resulting in low implementation rates for these treatments [
37]. To document the discussion and implementation of “difficult medical decisions” we will perform a semi-structured interview with the psychiatrists in charge at T1. This interview will – for every patient participating in the study - cover four domains: “difficult compounds”, “difficult other therapies such as ECT”, “difficult psychosocial interventions such as legal guardianship”, and “difficult other decisions as specified by the psychiatrist”. For each domain it will be documented whether or not there was any need for discussing decisions, whether there was a discussion and whether any decisions were implemented. Data gathered in this interview will result in two “sum scores” of difficult decisions (number of decisions addressed, number of decisions implemented).
In addition, we will obtain at discharge, the length of the inpatient stay, the future psychiatrist after discharge and a brief description of discharge medication (antipsychotics prescribed: yes/no, LAI prescribed: yes/no, monotherapy with antipsychotics vs polypharmacy with antipsychotics: yes/no). Outpatient psychiatrists will be asked to document any rehospitalization of their patients, their estimate of the patient’s adherence (one item visual analogue scale), CGI and GAF scores at T2–T5.
Qualitative data
The rationale behind also obtaining qualitative data is twofold. First, we would like to get a better insight into the process and working mode of SDM-PLUS and, second, we aim to study the potential barriers and facilitators of SDM-PLUS.
Therefore, we will use the critical incident method to identify clinical situations of interest (e.g. dissatisfied/very satisfied patients, decisions with or without patient participation, patients for whom health staff judges SDM-PLUS as inappropriate or impossible etc.) and then perform qualitative interviews with 24 key informants involved in the trial (physicians, nurses, patients). In addition, we will perform focus groups with health care staff after the end of the study to collect their experiences with the intervention. The qualitative part of the study will take place only on intervention wards and will be performed by two trained researchers who are working for the project and are in permanent contact with patients/staff of the participating hospitals.
The qualitative data will be audio-recorded, transcribed and then analyzed using content analysis [
38].
Treatment integrity
To assess treatment adherence we will document: a) how many of the team members actually receive the SDM-PLUS training, b) whether there was a change in attitudes towards SDM (pre-/post-questionnaire), and c) to what extent single elements of the intervention are implemented for patients (i.e. use of a “fidelity scale” during weekly supervision with all physicians in charge). On the patients’ side, the number of patient intervention sessions per patient and the percentage of patients receiving question prompt sheets will be documented.
In addition, we will use a qualitative approach in the intervention group to assess various aspects of the process of SDM-PLUS as well as barriers and facilitators of SDM-PLUS (see “Qualitative data”).