This randomized clinical trial compares the influence of joint crisis plans (JCP) or crisis cards to reduce psychiatric coercion for people with severe and often recurring mental illnesses like schizophrenia, bipolar disorder or schizoaffective disorder. Both interventions will be carried out as an integrated part of otherwise standard psychiatric in-patient and out-patient care in psychiatric units specializing in the acute or non-acute treatment of mentioned mental illnesses.

Single open-label randomised clinical trial with parallel control groups.

The intervention group "joint crisis plan" benefits from a joint crisis plan in addition to the usual in-patient care. The active control group "crisis card" benefits from the crisis card in addition to the usual in-patient or out-patient care.

OBJECTIVES:

The study aims at evaluating the effectiveness of a joint crisis plan compared with crisis cards regarding the reduction of coercive measure in the following hospital admission.

PROCEDURES:

A weekly screening of all service users regarding eligibility will be carried out. Each eligible person will be approached by a trained staff member of the unit and invited to participate in the study, but not before the patient is capable of consenting to treatment. After informed consent has been obtained, a staff member conducts a baseline assessment using the Global Assessment of Functioning (GAF), the Clinical Global Impression (CGI) and the Brief Psychiatric Rating Scale (BPRS). Then participants will be randomly allocated to either intervention or control group and specific intervention will be carried out. Joint crisis plan: Due to the duration of the process, the facilitated preparations for the negotiation session should start immediately. For those treated as in-patients, intervention should be finish before their release from the hospital. Crisis card: The in-patient and the attending physician fill out a crisis card together when the patient is to be released from the clinic. Out-patient prepare a crisis card together with their clinic physician during a routine visit to the clinic doctor's office. After the 12-month-follow-up-period participants will be assessed again regarding received psychiatric in-patient care and experienced coercive measures in any clinic during this period. Additionally, the patient's case files are examined for documented coercive measures.

HYPOTHESES:

The primary hypothesis to be tested is whether JCPs significantly reduce the average of involuntarily spent days of hospitalization during the 12 month follow-up period, compared with the control group receiving crisis cards.

Secondary hypotheses will be to determine whether compared with the control condition, JCP use will result in improvements regarding other common coercive incidents: involuntary medication, isolation, and physical restraint.

SAMPLE SIZE CALCULATION:

The necessary sample size regarding the primary outcome criterion "days in accommodation" was calculated utilizing routine data of a random sample (N = 20) of in-patients. The following inclusion criteria were used for the population of patients:

• relevant period: 1st of January to 31st of December 2016
• at least one acute psychiatric inpatient hospital stay with admission and discharge in 2016 (UKE: station PEAG)
• minimum length of in-patient treatment 3 days
• documented ICD-10 diagnosis: F2*/F3*

A zero-inflated negative binomial regression was calculated to obtain the estimator for the average number of days in accommodation (15.69) and the estimator for the overdispersion (0.49). The sample size calculation was carried out with a type I error of 5% (two-tailed) and a power of 80%.

The calculation was carried out with PASS 2008 (NCSS, LLC Kaysville, Utah). It led to a sample size of 151 patients per group (a total of 302 patients). Taking withdrawal and follow-up loss of combined approx. 20% into account an initial sample size of 374 patients (187 per group) will be necessary.

STATISTICAL ANALYSIS:

In order to address primary and secondary objectives, an intention-to-treat analysis (ITT) will be calculated as a zero-inflated negative binomial regression. The group variable (JCP vs. crisis card) is included as a fixed effect into the model. Additionally, a per-protocol analysis (PP) will be performed to examine the actual interventional effect.

Inclusion Criteria:

• mental capacity to consent the participation in the study
• principal diagnosis according to criteria in ICD-10 main groups F20-F29 or F30-F39
• patient is residing in the care area of the clinic OR currently at least 2nd inpatient hospital stay in this clinic OR currently treated in a psychiatric outpatients department of the participating clinic
• at least one psychiatric hospital stay within the last 24 months in any clinic including the current stay OR patient has a history of coercion (involuntary hospitalization, mechanical restraint, isolation, forced medication) in any clinic including the current stay

Exclusion Criteria:

• clinical impression of insufficient cognitive capacity
• insufficient understanding of the German language
• planned transfer to involuntary long-term inpatient care or forensic facilities
• serious somatic or organic brain disorder (dementia e.g.) or more than slight mental retardation