Introduction and rationale:

Unhealthy eating habits and lack of physical activity are risk factors for many diseases (including metabolic syndrome) and contribute to a shortened lifespan of 15-30 years in people with severe mental illness (SMI). Literature, mainly including short-term hospitalized or outpatients, show strong positive effects of activation on both physical and mental health. However, studies in long-term care are limited. In recent years, implementation of a lifestyle enhancing treatment intervention in clinical settings in "GGz Centraal" has demonstrated to be effective. The question is whether this kind of lifestyle intervention in sheltered housing is applicable and effective.

Objectives:

This research aims to develop an appropriate lifestyle intervention for patients living in sheltered housing services of GGz Centraal, based on input of patients and directly involved. Does applying this lifestyle treatment result in a positive effect in health and quality of life of patients and what is the influence of contextual factors, personal- and disease characteristics?

Study design:

In this intervention study, we use an experimental design. Municipal locations are paired based on the number of participants to generate equal cluster sizes. These paired clusters are randomly allocated to the control or intervention arm by means of a random number generator by an independent person (not involved in this project). At the start of the lifestyle treatment patients in the experimental and control group are invited to participate in the baseline screening. After twelve months, following a post-test on all outcome measures.

Treatment intervention:

The intervention in this study consists of formulating a lifestyle intervention, by patients and directly involved, aimed at enhancing a more active and healthier lifestyle . After formulation of the plan (based on psycho education, nutrition and physical activity), it wil be executed for a twelve month period.

Hypothesis:

Hypotheses is that lifestyle enhancing treatment is associated with improved metabolic health, quality of life and reduced use of medication. Furthermore we expect that movement disorders and disease severity will be negative related in becoming more active.

Sample size calculation:

To calculate the sample size we use the effect size on decrease in waist circumference in the previous intervention study (d =0.51) and the same analysis (multiple regression with correction for baseline value's on age, diagnosis and disease severity). To be able to detect the same effect in the current study with a minimum 80% power as a benchmark for a fair test and a significant level of 95% (α = 0.05), a minimum sample of 124 people is required (2 groups of 62). Taking into account a response rate of 73% from the first measurement a sample size of 168 patients is required.

Analyses:

We use multilevel regression to analyze the intervention effect. Possible clustering of data within the residential locations (and thus care teams) is taken into account by using a two-level structure with the first level residential location and the patients as the second. The intervention variable is set as an independent variable against difference scores of outcome variables (T2 minus T1) and corrected for the baseline value of the outcome to prevent regression to the centre. As we are unable to randomize patients individually in the current configuration on personal and disease characteristics (gender, age, diagnosis, disease severity at the start of intervention), these factors will be corrected for in the analyses if they differ significantly(p<0.05) between intervention and control group, analysed using independent t-tests and chi-squared tests. Characteristics that differ significantly will be included as covariates in the analysis described above. Multicollinearity will be checked with correlation coefficients and collinearity statistics (tolerance and Variance Inflation Factors (VIF) values).

Missing data:

Patients who are hospitalized for more than two months will be excluded from analyses.

If baseline or follow-up data are missing for two or more measures on physical or psychiatric health, patients are excluded from the analysis as insufficient difference scores can be calculated. Patients lacking difference scores on one outcome variable, are excluded from the analysis of that particular variable.

Inclusion Criteria:

• Severe Mentally Ill patients,
• living in Sheltered Housing facilities

Exclusion Criteria:

• Incapacitated patients,
• without informed consent from their legal representative