Aim of the study and hypotheses
The aim of the study is to evaluate the effectiveness and cost-effectiveness of a health promotion intervention targeting PA and healthy eating in people with MD living in sheltered housing. We hypothesize that:
· Between baseline and the end of the intervention and after a 6-month follow up period, significant differences in the primary outcomes ‘body weight’,‘Body Mass Index’ (BMI), ‘waist circumference’ (WC), ‘fat mass’ between the intervention and control group will be identified;
· Between baseline and the end of the intervention and after a 6-month follow up period, significant differences in the secondary outcomes ‘quality of life’ (QOL), ‘PA levels’, ‘eating habits’, and ‘psychiatric symptom severity’ between the intervention and the control group will be identified;
· The health promotion intervention is cost-effective.
Study design and setting
The design consisted of a cluster preference randomized controlled trial. An overview of the study design can be found in Figure
. The study was conducted in sheltered housing organisations (SHOs) in the Flanders region (Belgium) with the SHOs as the unit of randomization. In SHOs, support on several domains (e.g. psychological, domestically, occupational, relational) is offered to the patients. In this type of health care service, patients are living alone in a studio or apartment or together with other patients in ‘community houses’. For this reason, cluster randomization was appropriate. Individual randomization at the level of the individual patient would decisively cause contamination bias due to the risk of participants in the intervention and control group living together. Cluster randomization to the intervention or control group therefore occurred at the level of the SHOs.
Health promotion intervention: Study design.
Mental health professionals working in the intervention SHOs were asked to lead and to support the health promotion intervention. This implied a significant involvement and workload above their usual workload. Also, as mentioned before, centers being assigned to a non-preferred intervention arm could be disappointed, which may reduce their interest to participate in the study [12
]. Moreover, a substantial risk of non-participation was also assumed based on the results of previous qualitative research indicating that lack of time due to the high workload in the daily care of patients is a common barrier for mental health professionals to engage in health promotion programs [21
]. Consequently, it appeared to be necessary to provide a detailed explanation about the expectations when serving as an intervention SHO. For this reason, preference randomization appears to be appropriate.
An invitation letter and response form with a self-addressed postage envelope was sent to the managers of all SHOs in the Flanders region. They were asked if they were interested to participate in the study having (i) no preference to serve as intervention or control group and to be randomized or, (ii) a preference to serve as intervention group (see a detailed description of the intervention below) or, (iii) a preference to serve as control group. A concise explanation of the aim of the study and of the expectations and content when participating as intervention group was included in the letter. If necessary, a second mailing was foreseen. If a SHO was not prepared to participate, they were asked to report the reason for non-participation.
Subsequently, based on the responses, SHOs were either assigned to the intervention or control group according to their preference or randomly assigned to the intervention or control group when they expressed no preference. Randomization occurred by an external person not involved in the study. Finally, the patients living in the intervention and control SHOs received both written and oral information about the study. The written information consisted of a detailed explanation about the study and an informed consent.
The study population consisted of people with MD aged between 18 and 75 years living in a SHO in the Flanders region (Belgium). There are 42 SHOs in the Flanders region, including 2662 approved places [23
]. Exclusion criteria included people aged <18 or older than 75 years, having a gastric ring or pacemaker placed, having cognitive impairments (assessed by the mental health professionals) compromising the understanding of the psycho-educational and behavioral sessions of the health promotion intervention.
Development of the materials
The theoretical framework of the intervention was developed using elements of several theories including the social-cognitive theory [24
], the self-determination theory [25
], and the control theory [26
]. The health promotion intervention was developed using the mediating variable approach including the mediating variables knowledge, skills, self-efficacy and motivation [24
]. A schematic overview of the theoretical framework can be found in Figure
. Knowledge is a necessary component of behavior change [24
]. For example, how to select appropriate food portion sizes, how to distinguish between sedentary and moderate or vigorous physical activities. Behavior-specific skills are those specifically related to the targeted behavior [27
]. For example, how to interpret the level of physical shape by measuring the pulse rate. Self-regulatory skills include goal setting and problem solving [27
]. Self-efficacy is confidence in one’s ability to successfully perform a task or behavior and is influenced in two ways: personal success and observing others successfully perform the behavior [24
]. Two types of motivation can be distinguished. People can be motivated because they value an activity either from a sense of personal commitment or because there is strong external motivation and support [25
Theoretical framework of the health promotion intervention targeting physical activity and healthy eating.
The staff manual was developed based on the manual ‘Health promotion on well-balanced eating and healthy physical activity’ developed by the Flemish Institute of Health Promotion and Disease Prevention [28
]. As the target population of this manual is the general population, some adjustments were made to focus especially on the lifestyle behavior of our target population (for example how to choose a more healthy lifestyle despite the presence of barriers associated with the MD). The manual was built around ten themes focusing on PA and healthy eating: 1) PA and healthy eating: Introduction, 2) Awareness of the consumption of fat and fibres, 3) A healthy lifestyle: advantages and barriers, 4) The food triangle, 5) Using the food triangle throughout the day, 6) Label reading, 7) The influence of the environment & Budget issues, 8) & 9) Physical activity, 10) A quiz regarding PA and healthy eating.
Study duration and intervention components
The study period consists of an intervention period of 10 weeks followed by a post-intervention period of six months. In addition to treatment as usual, the intervention groups (n
14) receive the 10-week health promotion program targeting PA and healthy eating. In the intervention group the following intervention components are offered:
· Psycho-educational and behavioral group sessions
This part of the program consists of 10 group sessions in a 10-week period and includes discussions on PA and healthy eating, problem solving, written exercises, quizzes and plans to increase PA levels and to stimulate a more healthy eating behavior. All participants in the intervention group receive the same information in the same format. The program is delivered by the mental health professionals working in the intervention SHOs.
· Supervised exercise
In the same 10-weeks period a weekly 30-minutes supervised walking session is organized. These sessions are also led by one or more mental health professionals.
· Individual counseling
During the 10-week intervention period, all participants in the intervention group receive individual support from the mental health professionals (for example motivation to persist, discussing of experiences).
Implementation of the intervention
The manager of each SHO in the intervention group was asked to discuss with their team of mental health professionals the selection of one or two persons, who would serve as contact person with the research team and who would be responsible for the sessions. Every intervention SHO was visited by the same researcher. The aim of this visit was to instruct the mental health professionals who would lead and supervise the sessions. Preferably, also other staff members were present during this training session. During the study period, it is possible to contact one of the researchers by phone or e-mail. If necessary, visits of one of the researchers to the SHO will also be possible.
Evaluation of the intervention
At the end of the study period, a process evaluation of the health promotion program will be organized for all participating SHOs in the intervention group. This evaluation will consist of a questionnaire with both closed and open-ended questions including topics on experiences, advantages and disadvantages of the program, lessons learned, and suggestions for further research.
Sample size calculation
The sample size calculation is based on an average change of the primary outcome body weight of 3.5 kg between the intervention and the control group at the end of the study. This change is based on the results of a previous literature review performed by the research team on the effectiveness and cost-effectiveness of lifestyle interventions on PA and eating habits in people with MD [12
]. Cluster randomized trials require larger sample sizes than the individually randomized design. This can be explained by the fact that observations on individuals in the same cluster tend to be correlated, and so the effective sample size is less than the total number of individual participants. This reduction in effective sample size and the degree of correlation within clusters is known as the intraclass correlation coefficient (ICC) [20
]. As no ICC for this kind of intervention in people with MD was found in the literature, an assumption was made by multiplying the sample size with a design factor of 1.5. A sample size of 371 participants in each group would provide a sample large enough to detect a difference in mean body weight change of 3.5 kg across the two groups with 80% power at a significance level of 0.05.
Data collection and outcome measurements
Participants will be asked to complete a questionnaire on sociodemographics including sex, age, duration of stay in sheltered housing, marital status, occupational status, contacts with relatives, tobacco and alcohol use, and medication use.
Primary outcome measures
The primary outcomes of the study consists of changes in body weight, BMI, WC and fat mass. Body weight is measured in all participants wearing light clothing without shoes by a member of the research team using a TANITA BC-420 SMA digital weighing scale (TANITA, Tokyo, Japan). A member of the research team measures height in a standardized way using a Seca 225 stadiometer (Seca GmbH & KG, Hamburg, Germany). The BMI is calculated by dividing the body weight in kilograms by the square of the height in meters. WC is measured with a Seca 200 tape (Seca GmbH & KG, Hamburg, Germany) by one of the researchers according to the guidelines described in the National Heart, Lung, and Blood Institute report ‘Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. The evidence report’ [30
]. The calculation of the fat mass occurs using the TANITA BC-420 SMA digital weighing scale (TANITA, Tokyo, Japan). This occurs at the same time as the weight assessment. Both the stadiometer and the digital weighing scale were placed on a flat surface to assure correct measurement of height, body weight, and fat mass.
Secondary outcome measures
Changes in PA are assessed using the Dutch long version of the self-administered International Physical Activity Questionnaire (IPAQ), as this questionnaire appears to be a reliable and valid PA measurement tool [31
]. The analysis of the IPAQ is based on self-reported data. Therefore, PA levels are also measured using pedometers as a more objective tool. The Yamax Digiwalker SW-200 (Yamax, Tokyo, Japan) is used as this is known as accurate and reliable for counting steps [32
]. Dietary habits of the participants are assessed using an adapted version for adults of the “Young Children’s Nutrition Assessment on the Web ”[33
]. Quality of life is assessed using the SF-36 Health Survey questionnaire. Finally, psychiatric symptom severity is assessed through the use of the Brief Symptom Inventory (BSI). This questionnaire is considered as a reliable and a valid tool useful in patient groups with different psychiatric diagnoses [34
Data on all primary and secondary outcome measures are collected at baseline and at 10 weeks. At the end of the study (at 36 weeks) only data on body weight, WC and fat mass will be collected and BMI will be calculated. At that time participants will be asked to only complete again the SF-36 Health Survey.
As there is growing need for health economic research in health care and health policy, the cost-effectiveness of the health promotion intervention will also be examined. This will occur by calculating the difference in costs between intervention and no intervention (usual care), by calculating the expected health gain expressed in quality-adjusted life years (QALY) through health economic modeling, and by calculating the Incremental Cost-Effectiveness Ratio (ICER). The ICER is calculated as the ratio of the net cost to the net health gain: ICER
) where I is intervention and NI is no intervention.
QALYs are calculated by multiplying the utility level for a given disease status (a health-related quality-of-life weight ranging between 0 and 1) with the number of years an individual suffers from that disease. A utility of 0 is assigned to death, while an utility of 1 represents perfect health.
Parametric and non-parametric tests are used at the individual level to compare the intervention and control group at baseline, depending on the distributions of the quantitative variables. The X²-test is used in qualitative variables. Repeated measure analyses will be used to evaluate differences in the primary outcome variables body weight, BMI, WC and fat mass between pre- and post-intervention in the intervention and control group. Because preference randomization occurred at the level of SHO and not at the level of the individual patient, cluster effects will also be examined. The analyses of the primary outcomes will be performed on an intention to treat (ITT) basis. Secondary outcome variables will be evaluated per protocol. A P-value ≤0.05 is considered statistically significant. For statistical analyses, SPSS®19 will be used.
To examine the cost-effectiveness of the intervention, a Markov decision-analytic model assuming a public payer perspective will be constructed to project health outcomes and costs of the health promotion intervention compared with usual care. Overweight and obesity are substantial risk factors for the high prevalence of type 2 diabetes and cardiovascular disease in individuals with MD[35
]. Therefore, the Markov model will be used to estimate, for both the intervention and control group, the development of cardiovascular disease and type 2 diabetes over time and the associated costs. The time horizon will be a 10-year period. In both the intervention and control arm several health states will be included in the Markov model. All future costs and health outcomes will be discounted respectively at 3% and 1.5% annually. The costs will be calculated by analyzing the direct health care costs and the costs of the program. The direct health care costs include hospitalization, medication, GP consultations, and other health professionals costs. The program related resource use (staff time, materials) and resulting costs will be calculated making a distinction between the resource used and costs related to the research purpose and those related to the intervention itself.
Permission to perform the study was obtained from the Ethics Committee of the University Hospital of Ghent (Belgium). Written consent for participation is obtained from all participants. Participation in the study is voluntary and all participants are informed that the data analysis will be anonymous and that they could withdraw from the study at any time. A reward (pedometer) for the participants in the intervention group who completed the program was foreseen.