There is considerable geographical variation in the rates of compulsion in psychiatric services within as well as between countries. Reducing the use of compulsion of patients with severe mental illness is an expressed policy aim, and also a demand from service user organisations. In Norway, municipalities hold responsibility for primary care and are therefore central to the delivery of services to people with severe mental illness. This indicates a potential for intervening at the municipal level to reduce the use of compulsion where it is high. The Reducing Coercion in Norway study (RECON) will, in collaboration with municipalities with high compulsion rates, develop a municipal-level intervention (Stage 1) that will be implemented in a cluster-RCT (Stage 2) to test if it has effect on compulsion rates.pulsion rates.

BACKGROUND:

In Norway, as in other parts of the Western world, there have been concerns about increase in the use of compulsory mental health care for patients with severe mental illness. There is growing public concern, however, despite the explicit instruction from the Norwegian government that health services should reduce the use of compulsion, rates have not decreased. The Reducing Coercion in Norway study (RECON) is designed to develop and test an intervention, at the municipal level, to reduce the use of compulsion.

Most referrals to compulsory treatment come from primary care. In Norway, it is the 400+ municipalities who hold responsibility for primary care. This includes multi-disciplinary municipal mental health teams, substance abuse services, social care and GP services. At the Specialist level, four Regional Health Authorities Commission a number of Health Trusts to deliver specialist mental health care through acute hospitals and District Psychiatric Centres. These centres often collaborate, in different ways, with municipal services.

Municipal service providers are often those in most frequent contact with patients with severe mental illness, and will commonly be those detecting early signs of relapse. Importantly, municipal services (usually GPs or out of hours emergency services) will serve as gatekeepers for involuntary admissions, as they refer to specialist services who then may invoke the Mental Health Act that sanctions compulsion. This suggests a potential for developing strategies within municipal services to intervene early and possibly prevent some episodes of compulsion. Very little research has investigated the role municipal services play in the compulsion or on relevant cross-sectoral collaborations. There is also limited research on out-patient interventions designed to reduce compulsion. A systematic review could only include 15 experimental studies internationally. Joint crisis plans, risk assessments and counselling of staff were identified as promising interventions. However, most studies had insufficient samples or were of too poor quality to draw conclusions. Moreover, all included studies were based in specialist out-patient services.

The 'Six Core Strategies' intervention, developed to reduce seclusion and restraint in psychiatric hospitals, has shown positive effects. The six strategies involved are: structured efforts to improve leadership and organisational change; use of data to inform practice; workforce development; use of seclusion and restraint reduction tools; involvement of peers, and; rigorous debriefing techniques. It has been implemented in six countries, with positive results. No equivalent has been developed for municipal services. Service users and clinicians in Norway have, however, suggested potential strategies for municipal services to impact on compulsion rates. These include crisis plans; structured meetings of patients' networks, spaces for informal contact; key contact persons for patients; improved skills for early detection of relapse and; improved local collaboration. Given the potential role of municipalities for reducing the use of compulsion, it seems clear that further research is needed. RECON takes the Six Core Strategies model as a starting point for developing and testing a municipality-based intervention for reducing the use of compulsion.

AIM To develop, implement and test the effectiveness of a municipality-based intervention to reduce the use of compulsion in mental health services, in municipalities with rates of involuntary treatment episodes above the national average.

The aim will be reached through a two-staged approach.

OBJECTIVES AND METHODS STAGE 1:

In order to facilitate stakeholder input and to ground the research in a thorough understanding of the structures and processes of municipal service delivery, much of the data to address Objectives 1-3 will be collected through a mix of qualitative methods.

1. To identify and secure the participation of mid-sized municipalities with above average levels of involuntary treatment episodes Mid-sized municipalities (20-50K inhabitants), with above average rates of compulsion and motivated to change service delivery to tackle compulsion, will be recruited to take part in the study. This will take place in the period June- September 2019.

2. To map the mechanics of current practice together with local health and social services, service users, family caregivers and other stakeholders Local current practice will be mapped regarding what services are available in the intervention sites and how they work individually and collaboratively with patients with severe mental illness. Particular focus will be on the processes in municipal services that lead to involvement of GPs and possible referral for assessment for compulsion, and how patients are cared for after discharge (some of them subject to outpatient compulsion). Across the five sites semi-structured interviews and/or focus groups will be conducted with service leads in municipal mental health and addiction services (outpatient and housing services), GPs, social services and corresponding services in the local DPCs and focus groups with service users and family caregivers. These data will be subject to thematic analysis. Content analyses will be conducted of relevant service plans and other documents outlining leadership and workforce development, service collaboration, and any relevant guidelines. This will take place in the period July- December 2019.

3. To develop a municipality-based intervention for reducing compulsion in collaboration with municipalities and service users.

   The data collected will be used in the development of the intervention through mediated dialogue seminars with services in the intervention sites. Based on the Six Core Strategies it is likely that themes will include: leadership and organizational change; using data to inform practice; workforce development; use of seclusion and restraint reduction tools (in the municipal setting this might include crisis plans, risk assessment and de-escalation tools); involve peers, and; rigorous debriefing techniques (most likely after a referral for assessment for detention and/or after discharge from hospital). This will take place in the period January- March 2020.

   OBJECTIVES AND METHODS, STAGE 2:

4. To implement and study the intervention in five municipalities The intervention developed in Stage 1 will be implemented by services in five municipalities, supported by the research team. Regular site visits and contacts will be conducted to monitor progress and fidelity to the intervention in the intervention sites. This will feed into a process evaluation that will identify potential facilitating factors and barriers for successful implementation. This will take place in the period April 2020- April 2021.

5. To test the effect of the intervention in a cluster-randomised controlled trial The effectiveness of the intervention will be measured by comparing the change in outcomes between the intervention and the control group (difference in difference). Outcome data will be obtained aggregated at municipality level, from the National Patient Registry and be measured over three time periods.

The main hypotheses are that:

• There will be a larger reduction in prevalence rate of involuntary treatment episodes in the intervention arm as compared with the control arm during the intervention period (i.e., Index vs T1) (primary analysis)
• There will be a larger reduction in prevalence rate of involuntary treatment episodes in the intervention arm as compared with the control arm in the longer term (i.e., Index vs T2).

The investigators similarly hypothesise that the change in secondary outcomes will be larger in the intervention arm than the control arm during the intervention period (i.e., Index vs T1) and in the longer term (i.e., Index vs T2)

SAMPLE SIZE CONSIDERATIONS Asking municipalities to take part in a research programme by which they are required to change current practices is likely to be experienced as demanding. As explained above, research staff will be supporting the municipalities throughout. It is therefore necessary to keep the number of clusters at a manageable size.

Because outcome data will be aggregated at municipality level, power calculation at patient level is precluded. Based on current statistics, a mid-sized municipality has 40-100 involuntary admissions and 12-30 episodes of outpatient compulsion each year. Municipalities with above average rates of compulsory treatment per 10K population will be targeted, with the aim to include five municipalities in the intervention arm and five in the control arm. This should be sufficient to give stable estimates of the primary and secondary outcome measures.

STATISTICAL ANALYSES Prevalence rates in the intervention and the control arms during the Index, T1 and T2 periods will be calculated and presented together with 95% confidence intervals. Differences between the intervention and the control arm in the change in prevalence rate from the Index to the T1 period (as primary analysis) and the Index to the T2 period will be assessed by z-test for proportions. Change in overall duration of involuntary treatment episodes will be compared between the intervention and the control arm by Independent samples t-test. Even though the available data will be aggregated at municipality level, knowing the size of each municipality, the number of episodes, mean length of episodes and standard deviation for those length of episodes will enable the use of the aforementioned tests. Exploratory analyses comparing outcomes between T1 and T2 periods will be considered.

CONSENT

• In each cluster, consent to participate will be obtained from the person in charge of municipal mental health services
• All professionals, patients and family carers who participate in qualitative research activities in Stage 1will provide individual informed consent prior to enrolment
• All primary and secondary outcome measures will be collected from the Norwegian Patient Records and will be aggregated at municipal level. As no research activity will take place in control sites, no consent will be sought.

RANDOMISATION:

Randomisation of eligible municipalities took place prior to inviting interventions municipalities to take part as follows. All municipalities that met the inclusion criteria (n=28) were ranked in descending order of the rates of involuntary admission per 10K population. In the cases where two or more neighbouring municipalities were on that list (n=7), the ones with the lowest rate were excluded. This left a list of 21municipalities. Paired stratification was applied by which paired the first and second municipality on the list were paired, number 3 and 4, number 5 and 6 etc. Within each pair, one municipality was drawn at random to take part in the intervention arm and the other was allocated to the control arm. If a municipality allocated to the intervention arm declines to take part, the control municipality will not be included. It was important to avoid a large spread in terms of the rate of involuntary admissions. Therefore, if a municipality allocated to the intervention arm declines participation and their neighbouring municipality had been excluded, that neighbour municipality will no longer be ineligible and will be included in the intervention arm, if willing to participate.

As we are developing a new intervention, we will target municipalities in the higher 50 percent of rates of involuntary treatment episodes, and of a size that makes it possible to detect effect. To avoid the effect of the national reorganisation of Norwegian municipalities that is currently underway, those municipalities involved in a merging process of their services will not be targeted.

Eligible municipalities are those that:

• Have a population size between 20-50K inhabitants (figures from 2018)
• Are in the upper half of all municipalities in terms of the rate of involuntary treatment (figures from 2016-17)
• Are not affected by a merging process

Exclusion criteria:

• Municipalities neighbouring another eligible municipality will not be included to prevent contamination should one of them be randomised to the intervention group. The municipality with the lowest rate of involuntary treatment episodes will be removed

Not applicable. No individual patient date will be collected.

The control group will be masked, but it is not possible to mask the intervention group.