Setting
In Dutch primary care, most patients (80–85%) are treated within disease management programs, contracted by so called care groups. These care groups are responsible for the organisation, coordination and delivery of diabetes care in primary care [
13,
14]. Three care groups with 43 general practices (both group and single handed practices, with 89 general practitioners) in the southern part of the Netherlands participate in the current study.
Recruitment
The electronic medical records from the people with T2DM from the participating general practitioners are scrutinised to find eligible T2DM patients. All patients who meet the inclusion criteria will be informed about the possibility to follow a group-based diabetes self-management education program with pre-selection (see further), emphasising that they will be only offered an educational program they are likely to benefit from. People can indicate on a response card whether or not they want to receive oral information about the study. If yes, a research team member will give the patient a telephone call to discuss the study. If the patient is informed and understands the information, informed consent will be completed. Once the patient has given informed consent the pre-selection will take place.
Pre-selection
The pre-selection procedure to assess a patient’s capabilities of self-management will be performed with the use of the SeMaS, a self-management screening tool which is validated in T2DM patients and sent to the patients home [
9]. The SeMaS is a 27 item questionnaire covering self-management abilities of five psychosocial domains [locus-of-control (3 items), self-efficacy (2 items), social support (1 items), coping (9 items), anxiety (4 items) and depression (3 items)], three skills items (computer, group and self-care) and two other items including perceived burden of disease [
9]. All separate items range from 0 to 3 on a 4-point Likert scale (locus-of-control, self-efficacy, coping, skills items) or from 0 to 2 on a 3-point Likert scale (social support, anxiety, depression). For all items higher scores represent higher capability, except for those in the anxiety and depression domain where higher score represent more barriers. The perceived burden of disease is scored on a visual analogue scale from 0 to 10. Patients with high scores on the domains anxiety (score >4 from total 8) and/or depression (score >3 of total 6) will be excluded during the pre-selection. Those patients first need to deal with their mood disturbance before they are able to improve their self-management skills. Patients without barriers for self-management, so with maximal self-management capacity (highest possible scores for the locus-of-control, self-efficacy, social support, coping and for anxiety ≤4 and for depression ≤3) were also excluded, since they would not benefit from the educational course. The general practitioner (GP) will be informed about the SeMaS scores of all participants.
Randomisation
Patients who are eligible after the pre-selection will be randomised (1:1) based on computer-generated random numbers, with sealed, opaque, sequentially numbered allocation envelopes. All patients entering the trial receive a random number from the coordinating research centre (Julius Center). Each number corresponds to either ‘BGI course’ (BGI-group) or ‘no BGI course’ (control-group). As patients will attend a course in the intervention group, it is not possible to blind participants for the treatment allocation. Patients from the same practices and general practitioners will attend a different course, so no large cluster-effect is expected.
Intervention: beyond good intentions (BGI)
This study elaborates on the existing self-management program BGI [
2,
17,
18]. It lasts 12 weeks and consists of one initiating 30-min individual session and four 2.5-h group sessions. In addition, one booster-session is given a few months after the last group session. The course structure is as follows:
Individual session
The patient’s knowledge, attitudes and ambivalence with regard to diabetes management will be discussed. For this purpose the ‘Diabetes Profile’ was specifically designed to help patients gaining insight in the risk factors associated with their diabetes. With a colour scheme, from green to red, their risk for long-term complications is presented. Patients are encouraged to set goals to work on. The trained practice nurse (see further) discusses glycaemic control, BMI, blood pressure and lipids profile with the patient. Patients are asked to formulate a first goal before the first group session and will receive the educational material that will be discussed during the group sessions.
Group sessions
The group sessions are held in groups of six to ten participants and take place every two to three weeks in a location nearby the patients’ homes. The first three sessions will cover topics that are relevant for all patients: physical activity, healthy diet and medication adherence (what do guidelines tell, why it is important to follow these guidelines and how to do it, what are known barriers in this respect, what might facilitate compliance, followed by individual goals to improve). In the fourth session patients will be able to work on a personally relevant goal.
Every group session starts with the introduction of the topic, patients are asked to share their opinions, emotions and experiences regarding the topic. Subsequently, patients write their own individual action plan, according to the proactive coping five-step plan:
-
Step 1: Formulate a concrete and attainable goal.
-
Step 2: What are necessary conditions and possible barriers?
-
Step 3: What may be alternative strategies for crossing these possible barriers?
-
Step 4: Formulate the final plan: what, how, where, and when?
-
Step 5: Indicate how the progress will be evaluated and when you are satisfied, the plan is discussed with the group to improve the final plan.
During the weeks in between the groups sessions patients will practice with their individual action plane, which will be evaluated during the next meeting.
In case of diabetes-related questions, patients are encouraged to collect additional information by themselves, e.g. on the Internet or by referring them to another professional like a physiotherapist or dietician. This way, patients are encouraged to take responsibility for their disease.
Booster session (evaluation)
In this session the past months will be discussed and goals will be evaluated. New goals will be set. In addition, the trainers will motivate the participants to bring up a topic for the booster.
Nurse training
The BGI course will be given by trained practice nurses (n = 3). They receive two 3-h training sessions from the psychologist who is familiar with the BGI training (Train-the-trainer principle). During the training they will work with the five-step proactive coping plan and will be taught on the principles of the self-regulation and pro-active coping theory. The practice nurse should have the role of facilitator and coach, stimulating participants to support each other and become self-confident in seeking information.
Measurements
Clinical measures
In both groups, BMI, systolic blood pressure, HbA1c and lipid profiles will collected from the electronic medical record at baseline and 2.5 years follow-up. Clinical measures in-between will also be extracted from the electronic medical record of the GP.
Questionnaires
Participants will be asked to complete questionnaires to measure diabetes self-management behaviour, proactive coping and quality of life at baseline and 2.5 years follow-up, with the following measures.
Diabetes self-management behaviour
(a)
The SDSCA, revised according to Toobert et al. [
19], includes eleven items assessing several aspects of the diabetes regimen: general diet, specific diet, exercise, blood glucose testing, foot care, and smoking. Ten items are rated on an 8-point Likert scale, measuring how many days an activity is performed in the last week. One item measures smoking status (yes/no) and the amount of cigarettes smoked in the last week. Each of the domains is measured separately.
(b)
The PASE questionnaire [
20] includes 15 items, which cover leisure time physical activities (six items), household physical activities (six items), and work related physical activities (three items). Items differ in scoring range and are translated to the time spent on each activity into a composite score which reflects the amount of energy expended.
(c)
The ‘Kristal food habits questionnaire’ [
21] consists of 22 items assessing patient’s habits with regard to fatty food intake, dietary products and fruit and vegetable intake. The items differ in scoring ranges.
(d)
The MARS-5 [
22], is a brief 5-item instrument which assesses the degree to which patients forget medication, stop taking their medication or change dosages. Items range from 1 (always–never adherent) to 5 (never–always adherent).
Proactive coping
Proactive coping will be measured with the ‘Proactive Diabetes Management Inventory’ (PDMI) [
17], designed to evaluate the success of our course in getting patients to implement the five-step plan in their self-care. The instrument includes 7 statements covering different aspects of the five-step plan. The first six items assess whether patients think about their goals beforehand in a proactive fashion; that is, do they anticipate the conditions needed to reach their goal and do they anticipate barriers they may encounter along the way? Items 7–12 assess whether patients anticipate and plan the strategies needed to achieve their goal. Finally, Items 13–17 assess whether patients monitor their progress and strategies and adapt these accordingly. The items range from 1 (definitely not) to 5 (definitely) [
17].
Quality of life
The EuroQol EQ-5D, Short-Form 36 (SF36) and a diabetes-specific quality of life questionnaire, namely the Audit of Diabetes-Dependent Quality of Life (ADDQol) will be used to measure health-related quality of life as well as health status.
(a)
The EQ 5D is a generic quality of life questionnaire, consisting of a Visual Analogue Scale (EQ 5D-VAS) and a classification system (EQ 5D-Profile) [
23]. The EQ 5D-Profile covers five domains: mobility, self-care, usual activities, pain/discomfort and anxiety/depression of health, each with three levels of functioning: level 1, no problems; level 2, some problems; level 3, severe problems. The EQ-5D health utility scores range from −0.33 to +1.00 and were computed using the Dutch tariff as described by Lamers et al. [
24,
25]. A score of 0 is equal to death, whereas 1 indicates full health. Negative values represent a health state worse than death. The EQ 5D-VAS is a graduated, vertical line, anchored at 0 (worst imaginable health state) and 100 (best imaginable health state). The patient is asked to mark a point on the EQ 5D-VAS that best reflects his/her actual health state.
(b)
The SF-36 generates a profile of scores on eight dimensions of health status: Physical Functioning (10 items), Role Physical (4 items), Bodily Pain (2 items), General Health (5 items), Vitality (4 items), Social Functioning (2 items), Role-Emotional (3 items) and Mental Health (5 items). These scales can be summarised in Physical Health and Mental Health. The 36 items differ in the scoring ranges [
26].
(c)
The ADDQoL, measures the perceived impact of diabetes on the quality of life; it includes 19 items, ranging from −3 to 3 on different questions, with 0 as the neutral score. Scores below 0 reflect a negative influence of the item on quality of life, and those above 0 reflect positive influences. The impact scores are weighted (impact rating x importance rating), so the actual scores per item can range from −9 to 9 [
27].
Cost-effectiveness
Data on health care resources use during the whole study period will be extracted from electronic medical records from the general practitioners and for health status the EQ 5D questionnaire will be used at baseline and after 2.5 years follow-up.
Sample size
In our previous study, the self-management intervention BGI achieved a significant BMI reduction of 0.77 after 9 months in screen-detected T2DM patients (BMI −0.39 in the intervention group vs. +0.38 in the control group) [
2]. For the current study a minimum sample size of 88 (44 per group) is required to achieve 80% power to detect a difference of 0.77 in a design with 6 repeated measurements of BMI (weight is measured every three months as part of usual diabetes care) having a AR (1) covariance structure when the standard deviation is 1.7, the correlation between subsequent observations on the same subject is 0.70, and alpha 0.05. With an expected drop-out rate of 20%, we need to include a total of 106 patients.
Statistical analyses
The efficacy analysis will be performed by intention-to-treat analysis, including all patients from whom a follow-up assessment is available. If necessary a multiple imputation technique will be used for missing data. The primary outcome will be analysed by using Generalized Estimated Equations (GEE) for repeated measures, with baseline BMI as covariate and time, intervention and the interaction of time and intervention as fixed variables. The secondary outcomes will be analysed with ANCOVA on the change from baseline to follow-up. In addition differences in characteristics (e.g. age, sex, education level) of participants and non-participants will be evaluated.
Cost-effectiveness
Lifetime costs and health effects (based on the EQ 5D) will be estimated using a modified probabilistic diabetes model. This validated model has been used before and is described in more detail elsewhere [
28‐
30]. In brief, the model simulates the natural history of type 2 diabetes and calculates costs and quality-adjusted life years (QALYs) for Dutch T2DM patients [
30]. The model includes a health state for cardiovascular disease (CVD; myocardial infarction, stroke, peripheral arterial disease), microvascular complications (end-stage renal disease, retinopathy and neuropathy), uncomplicated type 2 diabetes and death. The model computes the occurrence of the above-mentioned diabetes-related complications and the excess mortality due to diabetes. Based on the estimated events and prevalence of complications, it computes diabetes-related lifetime medical costs and QALYs.
To calculate lifetime costs and outcomes, each health state is assigned a value in terms of medical costs and utility (health-related quality of life), and this value is multiplied by the prevalence of the health states over time.
In addition to the model input data described above, medication costs of glucose-lowering drugs, anti-hypertensive medication and cholesterol-lowering drugs used during the total follow-up period will be included in the cost calculation. Costs of the BGI-course and total health care resources (including consultations of the dietician) use during the whole study period will also be included in the cost calculation. These costs will be extrapolated to estimate lifetime medication costs, assuming that the cost difference between the BGI-group and the control-group remains constant over time.
The model will calculate two lifetime health outcomes (life years, QALYs) and costs for each patients (discounted and undiscounted). To correct for confounding and improve model estimates of the difference in outcome between the BGI-group and the control-group, the following baseline covariates will be used: age, sex, duration of diabetes, history of cardiovascular disease, smoking, HbA1c, systolic blood pressure, total cholesterol and HDL cholesterol.
The primary outcome of the cost-effectiveness analyses will be the cost-effectiveness of BGI versus control, expressed as the incremental cost-effectiveness ratio (ICER), calculated by dividing the incremental costs by the incremental QALYs or incremental life years. Uncertainty surrounding the cost-effectiveness ratios as calculated from the model will be expressed using a cost-effectiveness plane. A cost-effectiveness acceptability curve will be created to determine whether implementation of BGI is cost-effective given different thresholds of willingness to pay for a QALY (e.g. a threshold of € 20.000 per QALY).