This is a multicentre open (except during the common initial education cycle) randomised two-arm controlled trial (Figure 1). This trial will be associated with a qualitative ethno-sociological study conducted on a sample of 30–40 persons.

Patients will be recruited from five different outpatient secondary settings in the four hospitals of the Reunion Island, and will be seen in outpatient consultations. Detailed information will be given by the investigating physician on the objectives of the study, its nature and constraints, the anticipated risks, and the expected benefits. Patients will be included in the study, and followed-up for two years, once a consent form has been signed.

1. Inclusion criteria: patients aged over 18 with type 2 diabetes, defined in accordance with WHO criteria treated for at least one year (OAD and/or insulin and/or GLP-1 analog); HbA1c ≥ 7.5% for three months or over; treatment regimen stable for at least three months (except for possible changes to insulin doses in the previous three months); living in Reunion Island, affiliated to French national health insurance, and patients having signed a consent form for participation in the trial.

2. Non-inclusion criteria: current or scheduled treatment liable to influence blood sugar levels, such as corticotherapy; current or recent hospitalisation (< three months) associated with the diabetes; structured education intervention in hospital or on outpatient basis in previous three months; severe evolving current complication: ischemic or proliferative retinopathy, severe chronic renal insufficiency (clearance < 15 ml/min), coronary heart disease, foot lesions; current pregnancy or desire for pregnancy during the trial; evolving cancer (except baso-cellular or spino-cellular cutaneous epithelioma); severe physical, cognitive or mental illness handicap; concomitant participation in a therapeutic intervention trial.

The patients randomised in both arms will benefit from a structured initial cycle. Trained educators, blind to the subsequent group allocation, will conduct this initial group education cycle within the 12 weeks following inclusion. At the end of these 12 weeks the blinding will be lifted, and the patients randomised in the two-year education arm will continue to receive on-going structured group education. All group sessions will take place at close quarters, organised by the RéuCARE health network in specific locations in each town (municipal and/or community settings).

Patients are reminded by telephone during the week prior to each educational session to confirm the date, time and place. Patients who fail to attend the session will be called within two days and offered a new schedule.

1. The theoretical basis and framework of the Learning nests’ approach for the structured educational intervention 14 42 .

The complex nature of chronic diseases requires that therapeutic education takes into account both its biomedical and psychosocial dimensions. Learning nests work on knowledge to be built, and social, individual and contextual dimensions of learning, as well as the required decision of actions on a long-term basis, taking into account the heterogeneity of individuals and of their conditions. The learning nests are based on the contributions of socio-constructivism 43 44 , and empowerment 45 , and take into account the enhancement of self-efficacy 18 , through skill mastery, problem solving, decision making and action planning.

Fine components structure the approach of learning nest group sessions:

Component 1: Self analysis of relevant knowledge. The learner is placed in a condition to be able to analyse the relevant knowledge before making the decisions, which will be his own and not subordinated to those of the educator. Each learning nest offers a constructive meeting between the learners' ideas and the reference knowledge. Since the Learning nests do not include lectures, this meeting is dealt with through diversified actions exercised by the learner, on forms of knowledge that are made manifest and manipulable.

Component 2: Acknowledgement of individual contexts and problem-solving. The patient's "ordinary" context is legitimised by virtue of the visibility conferred on it (photographs, handouts, material, foods, ...). The learner can thus identify the affordances of his or her environment. The learner can continue to "manipulate" the reference knowledge in his own particular context. The encounter with the reference knowledge takes place in the presence of the health professional, and this body of knowledge is conveyed into the patients' context in the form of booklets that make sense to them, because the information has been put together there and then.

Component 3: Individual heterogeneity as an asset for self-assessment and action. The heterogeneity of the patients is not considered a problem, but as something of value, which is taken into account in the very act of designing the situation. The analytic dimension of the work to be carried out by patients is of importance. The specific individual characteristics of the patients can be integrated without exception. The outcome of each learning nest thus varies over a range of different solutions, and not a single one.

Component 4: The long lasting dimension of self management. The learning nests integrate the lasting dimension of chronic illness and the necessity of thinking the action over for a long time before seeing the first results. The decision for action always remains the property of the patient, with a space for encouraged actions in parallel, which leads the trainer to work with the patient's individual dispositions, helping the patient to become aware of the advantages and constraints of his environment. At the end of each group session patients identify a goal and an action plan related to their context. At the beginning of the each subsequent session patients are invited to reflect on their experience of working on their self-selected goals, and talk about what they had learned and how it could be applied in the future. Patients are encouraged to discuss events that had occurred and to reflect on how these events had affected self-management and action.

Component 5: Culturally tailored education materials. In addition to ensuring cultural relevance by focusing on the priorities and the personal contexts of the patients in the program, learning nests use education materials that are developed with and for the target audience: local foods, contextual photographs, natural surroundings, occupations, clothing, appropriate reading level, coloured items (for identification).

Learning nests incorporate and emphasise behavioural strategies, rather than a traditional didactic approach, including curricula and handouts that have been developed to meet the specific needs of the target population 14 19 40 46 . They also involve psychosocial dimensions, and the presence of the context in which the chronic illness is dealt 47 with on a daily basis, and the patient’s need for a decision on actions to be taken.

The learning nest can be defined as a learning empowerment situation centred on the construction of bodies of knowledge, in which the knowledge worked on in a given situation will enable the patient to take decisions relating to his health management. The patient has the possibility to modify his everyday practices (nutrition, physical activity, taking medication) while bearing in mind his/her existence in relation with the others, with his environment, his resources and his culture. Patients become aware of the issues they can manage on a priority basis, taking account of not only of medical dimensions (level of HbA1c, other risk factors, treatment, etc.), but also psycho-sociological, contextual, environmental, cultural and economic dimensions. This work results in a decision which incorporates a positive view of health and feasibility in contexts in which the patient will develop. Whatever the specific cultural, religious and socio-economic background of the patients, each can work on the basis of his own data, reinforce his own views and find a way ahead which takes into account the constraints and strengths of his social, family and personal life.

The material, process and the design of each educational group session are detailed in specific booklets for learners, including those with literacy difficulties. Booklets for educators include a curriculum written protocol, available in French and in the English language (EPMC booklets; Education Prévention des Maladies Chroniques 48 ), allowing replication of the educational intervention.

2. Initial education cycle

The initial cycle consists of 3 to 7 educational weekly or bi-weekly 2-hour group sessions for 4 to 10 patients:

Personalised assessment group session (120 minutes): patients work on their own clinical and biological factors that can influence health maintenance (blood glucose, blood pressure, lipids, weight, waist circumference, smoking), before looking at lifestyle measure effects and specific impacts of eventual changes. They then scrutinise the feasibility of action planning and implementation in real life, and goal-setting taking into account all individual, social, and contextual dimensions. At the end of this session, each learner can schedule 2 to 6 specific additional focused sessions, according his/her needs and goal-setting.

Two to 6 specific additional focused sessions of 90 to 120 minutes, spaced at intervals of 1 to 2 weeks (3 weeks maximum). These sessions are selected on the six following themes: self-monitoring and adaptation of treatment - physical activity - food and control of fat intake – understanding of diabetes and treatment – challenges of insulin therapy – prevention of foot lesions.

3. On-going SME support: personalised empowerment follow-up including group sessions of 90 to 120 minutes each, with three to ten patients at 16, 32, 48, 64, 80 and 96 weeks. Each session will be scheduled at the end of the previous session. Patients will work on the progress of their indicators as follows: medical (HbA1c levels, blood sugar, arterial pressure, lipids, waist circumference, smoking, treatments); health coping and goal setting (action(s) initially decided on, actual implementation, obstacles); cognitive (review of knowledge gained through the first session assessment session and additional sessions). The summary of the session comprises objectives for action(s) to be implemented and the scheduling of optional additional sessions to support practical implementation.

4. Educators

The professionals carrying out the sessions have been specifically trained (long-term training by the University Diploma on SME in chronic diseases, or training with experience and skills validated in the network). Training of educators include sessions on biomedical knowledge on diabetes, nutrition and treatment, educational skills, behavioural strategies, learning nest and empowerment approach. In order to “embrace empowerment”, health care professionals have to make a paradigm shift from the traditional approach to care and acknowledge that patient are in control of their daily diabetes care, making autonomous and informed decisions about their diabetes self-management 49 . Additionally, training includes for each thematic group sessions a thorough review of the process, content, educator role, learners actions, learning indicators, specificity of goal setting and problem solving taking into account individual, social and cultural context.

In the control arm, patients receive only the initial structured education programme, conducted in the 12-week period following inclusion.

In both arms, patients will continue to receive quarterly outpatient medical follow-up. Regular examinations and therapeutic adaptation will take place in accordance with the recommendations. The individual nursing or dietician encounters, required in addition to the medical appointments due to the evolution of the disease or therapeutic changes (such as starting insulin, GLP-1 analogue) will be authorised whatever the arm.

The main outcome is the evolution of HbA1c (%) at two years. The HbA1c will be measured by the HPLC method in a centralised laboratory (biochemistry laboratory of the Félix Guyon hospital in St Denis).

The secondary outcomes will be as follows: quarterly evolution of HbA1c (%), fasting blood glucose, anthropometric indicators (body mass index, waist circumference), blood pressure, tobacco consumption, annual screening of lipids (total cholesterol, triglycerides, HDL and LDL cholesterol), antidiabetic treatments (type and dose) and the occurrence or worsening of diabetic complications. Health practices (level of physical activity and food consumption) will be assessed by means of annual questionnaires. These have been used in Reunion Island for a number of descriptive or intervention studies: RECONSAL 50 , REDIA-prev1 51 and REDIA-prev2 31 :

Physical activity: physical activity scores will be assessed by means of a questionnaire developed on the basis of the Baecke self-questionnaire 52 , which explores professional, sporting and leisure exercise. Questions on domestic activity were added to our studies.

Food consumption: food consumption (reported energy intake, macronutrient intake, dietary habits) will be assessed by means of a rapid food frequency questionnaire relating to weekly consumption 53 . Quantities of various foods are assessed by means of a photo album. The validity of the questionnaire and the photo album for the population of Reunion was checked by comparison with food surveys.

The care and treatment courses will be assessed by means of questionnaires at inclusion and at two years, based on the questionnaire used in the French ENTRED study 54 adapted to the situation and context in Reunion and comprising 30 items.

Quality of life will be assessed annually by means of the Diabetes Quality of Life, Brief Clinical Inventory 55 . This scale, including 15 items and translated into French for Reunion, was validated for this trial among a sample of 150 Reunion’s patients presenting diabetes.

Self-efficacy, social support and anxio-depressive state will be assessed by means of psychometric scales validated for the purpose of the trial in Reunion. The feeling of self-efficacy will be measured by means of section 3 of the MDQ (Multidimensional Diabetes Questionnaire) 56 . This questionnaire includes 13 items: seven for self-efficacy and six for outcome expectations. The patent’s perception of the specific social support from the partner with regard to the treatment of the diabetes will be assessed on the basis of 12 items derived from the MDQ. The patient’s anxiety will be assessed by the HADS (Hospital Anxiety and Depression Scale) 57 . The HADS includes seven items and is widely used and validated in outpatient prospective studies. The depression level will be measured by means of the CES-D 58 59 , which comprises 20 items. The French versions of the two latter tools have demonstrated their psychometric qualities.

The food consumption questionnaire will be conducted face to face with the centre nurse trained for this purpose before the start of the research. The other questionnaires will be self-administered, with the assistance of the study nurse if necessary for subjects with low literacy levels.

The efficacy assessment is based on the following main judgement criterion: developments between inclusion (week 0) and two years (week 96) of the percentage of glycated haemoglobin HbA1c (%).

The expected benefit is a decrease at two years of at least 1% in the average HbA1c between the group “structured on-going SME support” and the control group “Initial structured education cycle only”. An absolute difference of 1% at two years between the two groups appears relevant 23 24 29 32 39 60 61 62 .

To show an absolute difference of 1% of HbA1c, with a standard deviation of 2.5%, a type I error rate of 5% and a power of 80%, it is necessary to include 99 subjects per group in a bilateral situation.

Because of an estimated data deficiency of 20% at two years (drop out, refusals, deaths), it is planned to include 120 subjects per group, making a total of 240.

Block randomisation is carried out on two stratification factors: insulin treatment on inclusion (yes/no - two strata), and inclusion centre (five strata). The numbers of participants in each group are balanced with a ratio of 1:1. The 10 randomisation lists are drawn up by the statistician of the CIC-EC of Reunion before the start of the research.

For all patients, the first phase (group-based initial education cycle) will be conducted blind with regard to the randomisation arm for patients, until 12 weeks at the latest, researchers and education team. At the end of this first phase the follow-up will therefore be unblinded.

The statistical analyses will be carried out at the CIC-EC in Reunion, under SAS® (version 9.2, SAS Institute Inc., Cary, NC, USA).

Description of the initial characteristics of the screened patients and of the included patients will be done by usual statistics (number, percentage, average ± standard deviation or median and range).

Effective implementation of the SME program and the effective participation will be documented.

The assessment of efficacy of the intervention will be performed in a single analysis, according to the intention to treat principles (i.e. according to group allocated by randomisation) with type I error equal 5%, and bilateral formulation of Student t-test (primary analysis).

In the two compared groups, the deficient data on HbA1c (%) at two years (week 96) will be replaced by the measurement obtained at week 84.

If the primary analysis shows a significant difference between the two groups, a comparison of the mean change in HbA1c (%) S96-S0 will be done by the level of participation (secondary analysis using 1-factor ANOVA model): i) basic level ("initial structured therapeutic education only"), ii) level 1 (“structured on-going SME support” without specific additional focused sessions), iii) level 2 (“structured on-going SME support” with specific additional focused sessions).

Analysis of longitudinal data gathered during two years will used in generalised estimating equations (GEE) adapted to repeated measurements in order to test the trends described over the duration of the supervision.

Analysis of the time between randomisation and occurrence of a level of HbA1c <7% (first event observed) will be performed by the Kaplan-Meier method, with comparison of survival curves between groups by Log-rank test.

At each annual collection point the anti-diabetic treatments will be classified into four hierarchical categories: 1) non-insulin antidiabetic monotherapy, 2) non-insulin antidiabetic bi-, tri- (or more) therapy, 3) one injection of insulin, 4) two or more injections of insulin. Between two collection points, moves from one category to a higher (or lower) category will be identified as a reinforcement (or reduction) of the treatment. The interaction between the intervention and treatment category will be tested in the appropriate multivariate models (ANCOVA, GEE, time-dependent variable in the Cox model).

The associated qualitative study will be conducted on a sample of 30 to 40 persons in both arms, and will include observation and recording group sessions as well as semi-directive conducted interviews at home.

The objectives of the ethno-sociological qualitative study are: i) to analyse the learning processes during the education sessions, and then in individual contexts to analyse the developments in health ‘literacy’ is taken to mean not only reading and writing ability and numeracy, but also information searching, cognitive work and decision-making 63 in the various areas covered by the SME process. These areas are: food, physical activity, control of risk factors in complications, blood sugar self-monitoring and adaptation of the treatment. 2) analysing the adaptation processes in context and their links with the sessions, interactions in the education session and literacy levels; 3) endeavouring to generate barriers and enablers in developing self-management.

The participants in the qualitative study will be selected on a non-random basis by the head of the ethno-sociological survey (M. Balcou-Debussche), so as to have a sufficiently wide panel of participants, taking account of the age, sex, length of time since the onset of diabetes and its level of control and complications, professional activity, habitat, socio-economic level and level of education. The possibility of inclusion in the ethno-sociological research (including interviews at home) will be specified in the information letter and consent form. Additional oral information will be provided for a qualitative study of participants on the aims and precise conditions of this survey.

An analysis will be made of the content of the themes covered in the interviews and the recurring themes. The categories and signifiers will be developed using the constant comparison method. The emerging themes can be included and developed in new analysis and data-gathering cycles 64 . The session observations will include an ethnographic summary of language interactions, speaking and activities of participants. All the data, themes and recurrence will be the subject of cross-analyses of the same individuals according to the context, sessions or interviews. The analyses will also cover the evolution since the beginning of the research, as well as differences between individuals. All the data will be coded and analysed with the aid of the N-Vivo 9 software QSR international.

Serum collection will be performed at inclusion, as well as at one year and two years for the purpose of a posteriori secondary analyses on the evolution of more specific indicators of the risk of long-term complications of type 2 diabetes. In particular, inflammation and immunity factors, oxidative status and lipid fractions (size and density of LDL, HDL fractions, atherogenic lipoproteins).

This study has received approval from the ethics committee, “CPP Sud-ouest et Outre Mer III” and authorisation from the French Health Products Safety Agency (Agence française de sécurité sanitaire des produits de santé, Afssaps).