Aim: To implement a nutrition education programme (intervention) for adults with type 2 diabetes mellitus (T2DM) adapted from a primary health care setting to a tertiary hospital setting in South Africa and to evaluate the programme's effectiveness on dietary behaviours, clinical status and selected potential behaviour mediators.

Participants and setting: T2DM adults (40-70 years) and at least one year of living with diabetes and poorly controlled diabetes (HbA1c ≥ 8%). The study setting is the outpatient clinic of a tertiary teaching hospital in Tshwane District (Pretoria), South Africa.

Intervention: The intervention will employ a randomised control design with two parallel groups (intervention & control). A total sample of 140 T2DM patients (70 per group) will be needed to detect a 0.5 % change in HbA1c (SD of 1.0 and a power of 80%) at six month and allowing a 10% attrition rate. The intervention is one-year long with the following components: 7-monthly group education sessions; 2 bi-monthly group follow-up sessions at the hospital till one year; participants' workbook for goal setting activities and education materials (pamphlet and wall/fridge poster) for the intervention group. The control group will receive the same education materials with no other education encounters. Both groups will continue with usual care at the diabetes outpatient clinic of the hospital. The education will be offered face to face, will utilize teaching aids including coloured posters and will incorporate interactive group activities and demonstrations. The main facilitator is a qualified dietitian.

Outcomes: Outcomes will be assessed at 6-and 12 months for both groups with the six month being the primary outcome. Outcomes will include clinical [HbA1c (primary outcome), BMI, blood pressure and full lipid profile); dietary behaviours (energy intake, starchy food servings, vegetable and fruits intake, macronutrient intake and their distribution to energy, fibre, meal pattern) and selected potential mediators of behavior (diabetes knowledge and diabetes management self- efficacy).

It is hypothesized that the intervention will lower the HbA1c levels by at least 0.5% at six months and the levels will be significantly lower in the intervention group compared with the control group, and the significantly lower levels will be sustained at 12 months in the intervention group.](streamdown:incomplete-link)

Aim of the study: To implement a nutrition education program (intervention) for adults with type 2 diabetes mellitus (T2DM) adapted from a primary health care setting to a tertiary hospital setting in South Africa and to evaluate the program's effectiveness on dietary behaviors, clinical status and selected potential behavior mediators.

General objectives

• The implementation of the adapted nutrition education program (NEP) at the diabetes outpatient clinic of a tertiary teaching hospital

• The evaluation of the program's effectiveness at six and 12 months on the following outcomes: (i) clinical status: Hemoglobin A1c (HbA1c) (primary outcome), BMI, blood pressure and lipid profile (ii) dietary behaviors (energy intake, starchy food servings, vegetable and fruits intake, macronutrient intake and their distribution to energy as a measure of meal balance, fiber and legume intake) (iii) potential behavior mediators (diabetes knowledge, diabetes management self-efficacy).

Primary outcome objective

• To determine the differences between the intervention and control groups at six months for HbA1c, (primary outcome).

Secondary outcome objectives

• To determine the differences between the intervention and control groups at 12 months for HbA1c
• To determine the differences between the intervention and control groups at six months and 12 months for the secondary outcomes: dietary behaviors, BMI, lipid profile, blood pressure, diabetes knowledge and self-efficacy for diabetes management
• To determine the differences between the intervention and control groups in the proportion of participants who achieve HbA1c levels of less than 7%.

Research hypotheses (i) The intervention will lower the HbA1c levels by at least 0.5% at six months and the levels will be significantly lower in the intervention group compared with the control group, and the significantly lower levels will be sustained at 12 months in the intervention group.

(ii) The intervention group will have significantly better outcomes in the secondary outcomes (blood lipid levels, BMI, blood pressure, dietary behaviors, diabetes knowledge and self-efficacy) at six months and would sustain significantly better outcomes at 12 months compared with the control group.

(iii) The intervention group compared with the control group would have significantly more participants achieving the HbA1c targets (<7%) at six months and at 12 months.

Study design The intervention will employ a randomized controlled clinical trial design using two parallel groups. One group (intervention) will receive the intervention (diabetes nutrition education) including education materials while the other group (control) will only receive the education materials. Both groups will continue with usual medical care at the diabetes outpatient clinic.

Setting and participants: The study setting is the outpatient clinic of a tertiary teaching hospital in Pretoria, South Africa. Participants are male and female T2DM adults (40-70 years) and at least one year of living with diabetes and poorly controlled diabetes (HbA1c ≥ 8%). Participants with no major complications, who are not pregnant or on full time employment and can understand English will be included. A total sample of 140 T2DM patients (70 per group) will be needed to detect a 0.5 % change in HbA1c (SD of 1.0 and a power of 80%) at six month and allowing a 10% attrition rate. A convenience sample of consequent patients will be used. Participants will be recruited face to face during their clinic attendance.

Randomization Participants will be randomised into either the intervention or control group using block randomisation in blocks of varying sizes using a computer generated random schedule. The participants will be stratified based on sex and age. Allocation concealment using sealed sequentially numbered opaque envelopes with the randomisation code will be done (Sex-age stratum will have its own set of the sequentially numbered opaque envelope).

Intervention: The intervention aims to improve glycaemic control and other clinical outcomes (BMI, lipid profile, blood pressure) through improved dietary behaviours and behaviour mediating factors. The intervention is one-year long.

The intervention group will receive the following components:

• Face to face 7-monthly group education sessions, 2 bi-monthly group follow-up sessions at the hospital up to one year and one individual counseling and goal setting activity
• Participate in individual goal setting activities at home using a workbook and receive education materials (pamphlet and wall/fridge poster) for use at home.

The education will be offered face to face, will utilize teaching aids including colorful posters and will incorporate interactive group activities and demonstrations. Participants will be encouraged to bring a family member or a friend. The main facilitator is a qualified dietitian.

The control group will receive the same education materials with no other education encounters. Both groups will continue with usual care at the diabetes outpatient clinic of the hospital.

Outcomes measurement: Outcomes will be assessed at 6-and 12 months for both groups with the six month being the primary outcome.

• Clinical outcomes: HbA1c (primary outcome), BMI, blood pressure and full lipid profile) will be measured using standard procedures.
• Dietary behaviors (energy intake, starchy food servings, vegetable and fruits intake, legume intake, macronutrient intake and their distribution to energy as a measure of meal balance and fibre). Three non-consecutive 24hour recalls (including one weekend day) will assess dietary intake. Standardized household measures will be used to help with portion estimation.
• Potential mediators of behavior (diabetes knowledge and diabetes self-management efficacy). Questionnaires considered as suitable measures for the outcomes and appropriate for use with the target population (through pilot testing) will be used. The questionnaires have been used in diverse settings and geographical locations.

Process evaluation: Quantitative and qualitative measures will be used to confirm fidelity, reach and participants experience with the programme.

Data analysis:

An analysis of co-variance (ANCOVA) will compare the intervention and control groups on the measured outcomes post-intervention, using the baseline values age and gender as covariates. An intention to treat analysis will be performed on all outcomes using the last observed response (carry forward). The level of significance for all tests will be at α < 0.05 for a two-tailed test.

Inclusion Criteria:

• Type diabetes mellitus, poorly controlled (HbA1c > % 8), at least one year living with diabetes, without serious complications, mobile, regular attendance of diabetes clinic, not on full time employment, not planning to leave study site in the next one year.

Exclusion Criteria:

-

Masters student undertaking an aspects of the project (e.g baseline). Data will be available after all baseline data is complete-anticipated 8-months from commencing recruitment

The health professionals (physicians, dietitians, nursing professionals) who serve the patients will only know the patient is participating in the study through a sticker on the medical records but they will not know the treatment allocation (group). Outcome assessors for biochemical parameters will not know the patient allocation as the assessments for the parameters are done alongside other routine assessments. Furthermore, they do not have the information that there are different patient groups