Comorbid insomnia represents a frequent health problem in patients with severe mental disorders, and cognitive behavioral therapy for insomnia (CBT-I) has been identified as the first line treatment. However, CBT-I has not sufficiently been implemented in acute psychiatry settings. Rather, patients are often overtreated with benzodiazepines or benzodiazepine receptor agonists, related to adverse effects and the risk of tolerance and dependency. This work aims to empower patients with severe mental disorders to take care of their own sleep health based on a pragmatic behavioral treatment program ("Become your own SLEEPexpert"). Implementation research strategies in collaboration with patients and health care providers were used to adaptat CBT-I components to the needs of psychiatric inpatients. Evidence for feasibility in an acute hospital setting and preliminary evidence for efficacy has been shown. ln the proposed project, the investigators aim to compare treatment as usual (TAU) + SLEEPexpert to TAU + sleep monitoring in a pilot randomized controlled trial. The objective is to target sleep to improve mental health and to investigate the efficacy of the SLEEPexpert programme for the improvement of sleep and mental health. Given the high burden of comorbid insomnia in psychiatry, the investigators believe that the presented work is of interest to basic scientists and clinicians and, potentially, of heightened public health relevance.

- Standard operating procedures: patient recruitment

During the enrolment phase, all patients admitted to the participating wards of the UPD Bern will be screened in person. The screening will consist of a few questions regarding sleep, implemented into the routine clinical interview at enrolment, and filling in the insomnia severity index. Eligible patients will be contacted by study staff, informed about the study, and invited to participate. For further information about the study, a flyer containing the most important details will be given to the participants. The screening and enrolment process will be standardized across the wards to minimize the risk of selection bias.

- Standard operating procedures: data collection & data management Source data verification Quality assurance: data validation & registry procedures

Data is recorded and collected in paper CRFs as source documents. ECRFs will be given as an option for direct anonymized data entry from participants if feasible. The eCRFs will be created and stored on the database RedCap which is a database recommended by the Clinical Trial Unit Bern. Data entry from the source paper CRFs to eCRFs should only be performed by authorized personnel and data should be consistent with the available source documents. When this is not the case, a detailed explanation of the discrepancies should be provided. The principal investigator should ensure the accuracy, completeness, and timeliness of the data reported and sign off the completed eCRFs. In order to reflect the current patient status throughout the study, eCRFs will be kept up to date. This means that eCRFs should be completed no later than two weeks after a participant's visit and that data entry not should be delayed unnecessarily.

- Data checks

Data edit checks will be implemented into the EDC (electronic data capture) system (RedCap), limiting entries to appropriate, realistic values (e.g. not allowing future dates, missing values etc.). Furthermore, selected data points are cross-checked for plausibility with previously entered data for each individual participant. Before database lock, each PI will validate the collected data from his site with his signature.

- Quality assurance: Site monitoring & auditing

On-site monitoring will be part of the quality control activities implemented for this study. For the purposes of monitoring, the PI at the site will provide the monitor with access to study documentation, patient records, facilities and any other resources necessary. No site audits by the sponsor are planned for this study, but the sponsor reserves the right to perform such an audit should it be considered necessary.

- Standard operating procedures: reporting of adverse events

AEs will be categorized concerning the type of the reaction, severity, relatedness to the intervention, action taken, and outcome. Adverse events will be asked at every study timepoint.

The frequency of occurrence of adverse events of different categories will be analysed using descriptive statistics and compared between the two groups using Chi square tests.

The sponsor investigator will report to the Ethics Committee via BASEC device deficiencies that could have led to serious adverse events if suitable action had not been taken, intervention had not been made, or circumstances had been less fortunate within 7 days

Sample size assessment A power analysis was conducted with G-Power. Assuming an alpha error level of 0.05 (two-sided) and a power of 0.8, a total sample size of n=60 will be sufficient to detect a medium effect (Cohen's d=0.65). This is a pilot trial with the primary aim of collecting qualitative experience and preliminary efficacy data. A large confirmatory trial will be conducted if the results of this pilot trial are encouraging.

Plan for missing data If a patient withdraws consent for further study participation, the data collected up to the time point of withdrawal will be included in the study and no further data will be collected. The data of withdrawn participants will not be anonymized but will remain coded after analysis of the data set. For participants who discontinue the study intervention but do not withdraw from participation in the study, study visits and data collection will continue in the same way as for non-withdrawn participants. Dropouts before start of treatment (i.e. non-starters) will be replaced by recruitment of new participants. Multiple imputation will be used in case data are not available for more than 5-10% of the sample.

- Standard operating procedures: Data analysis & statistical analysis plan Primary Analysis

The primary outcome, i.e. the four overall scores of the ISI of all follow-up time-points will be analyzed with a mixed effects linear model. The model will take into account the hierarchical structure of the data: multiple ISI scores per patient and clusters (cross-over). Because of the complexity of the data and the small number of clusters the investigators assume that they will not be able to introduce random effects for patients and clusters. However, the first approach will still be a linear mixed model with the following co-variables: fixed intervention effect (TAU plus SLEEPexpert versus TAU plus sleep monitoring), fixed baseline ISI score effect, random patient effect (multiple ISI scores during follow-up), random cluster effect. Depending on model fit, the investigators may need to simplify the model and modify the random effects. This model is pre-specified and the rationale for switching will be made explicit. The analysis will be made after all data has been collected by one of the principal investigators.

Secondary Analysis

Continuous outcomes (Brief Symptom Inventory, Beck Depression Inventory) will also be analyzed by a mixed effects linear model using the same procedure used to estimate the primary outcome. Binary outcome data will be analysed using mixed effect logistic regression with odds-ratio as the effect measure. The number of hospitalizations will be also analyzed by a mixed effects model.

Data dictionary The detailed data dictionary can be found in the study protocol in the document section.

Inclusion Criteria:

• Age 18 years or older
• Documented diagnosis of acute insomnia, i.e. insomnia criteria according to ICD-10 for at least 2 weeks
• Insomnia Severity Index (ISI) total score ≥ 8, equivalent to relevant insomnia
• Inpatient in one of the two participating psychiatric wards in the UPD Bern
• Ability to understand the aims and procedures of the study
• Willingness to participate and ability to provide written informed consent

Exclusion Criteria:

• Incabability of judgement
• Inability to participate in a low-threshold behavioral treatment program, e.g. due to severe cognitive impairment or high symptom severity precluding participation (e.g., severe catatonic symptoms, massive hallucinations, acute endangerment to self or others, involuntary commitment). Note that the named symptoms will lead to exclusion only if they lead to inability to give informed consent or an inability to participate in the program. Excluded patients will be re-assessed regularly and included if symptoms improve and participation is possible at a later point in time.

Study participants and care providers cannot be blinded because they will notice whether they receive / deliver an actual treatment or a monitoring control. Due to the crossover cluster design (randomization of wards to either TAU plus sleep monitoring or TAU plus SLEEPexpert), patients who are on the same ward at the same time will not get in contact with patients who are in the other arm. Thus, contamination due to exchange between patients will be reduced as far as possible.

Outcome assessors (interviewers) and data analysts will be blinded. In the respective documents and database entries, patients will be identified with a code. The key for assigning the codes to patients will not be known to the interviewers and data analysts.