Healthy eating is important for the disease course and quality of life (QoL) of people with type 2 diabetes mellitus (T2DM) [1]. Healthy eating may lower hemoglobin A_1c (HbA_1c) to a similar extent as medical treatments for T2DM [2,3]. In some cases, healthy eating can put the disease in remission [4-12]. However, people with T2DM indicated that implementing healthy eating recommendations in their everyday lives is challenging [5,7,8].

Several strategies are available to support people with T2DM with healthy eating. The strategies approach healthy eating in dissimilar manners that could be broadly categorized as dietary prescription, lifestyle counseling, and behavioral therapies [13]. Dietary prescription is characterized by a “top-down” and biomedical approach: the therapist actively provides the solution in the form of a strict calorie-restricted diet, and the main goal is to achieve weight loss or improved HbA_1c levels (eg, [12]). In lifestyle counseling, the participant has a more active role: the therapist provides lifestyle education, meal plan templates, and personalized advice to implement nutritional recommendations. The main goal is adherence to a healthy lifestyle (eg, the Look AHEAD [Action for Health in Diabetes] trial [14] and the Diabetes Remission Clinical Trial [DIRECT]; [15]). Behavioral counseling is characterized by a more “bottom-up” and holistic approach: the starting point is the participant’s preferences; the goals and approaches are collaboratively planned. The therapist has an enabling role in the process: stimulating the participant’s problem-solving abilities and self-efficacy by motivating and supporting the participant. The main goal is achieving a mindset that favors healthy lifestyle behaviors (eg, [16]).

All of these strategies have been successful in initiating health benefits such as weight loss and improved HbA_1c levels [12,16,17] but have struggled with maintaining the benefits in the long term [8,18-21]. Even after the most well-known T2DM interventions—the Look AHEAD [14] and the DIRECT trial [15]—health benefits gradually disappeared. The Look AHEAD trial (restricted caloric intake using portion-controlled meal plans, calorie-counting techniques, and meal replacements combined with moderate physical activity) demonstrated average weight loss of 8.6% (SD 6.9%) in the intervention participants, but half of the initial weight loss was regained after 5 years of follow-up [22,23]. The DIRECT trial (a very low–calorie diet, followed by stepwise food reintroduction) demonstrated that almost half of intervention participants achieved T2DM remission at 12 months, but this percentage declined to 36% at 24 months [24].

Most previous studies used dietary description or lifestyle counseling strategies. These strategies either included strict (ie, very low–calorie) diets or were performed in highly controlled research settings (intensive intervention regarding diet and physical activity combined with frequent contact with health care providers) [12,25,26]. However, people may experience difficulties when they return from these clinical and controlled research settings to their everyday lives. In everyday life, healthy eating requires a scale of personal and social resources, ranging from practical cooking skills to personal agency [27-29]. It is not only about what people eat but also why and how they eat: the ability to prepare food, social economic resources, the food environments, social support, self-image, mental health, etc. This may explain why the Mediterranean diet is considered as one of the healthiest diets with (long-term) health benefits [30]; the Mediterranean diet is not a strict diet, rather a way of life that includes seasonal cooking, freshly cooked meals, and socializing with others [31].

Hence, there seems to be a need for a strategy with an eye for everyday life of people with T2DM to promote long-term healthy eating. A health theory centered around the everyday life context is salutogenesis. Applying salutogenesis to nutritional research leads to the assumption that healthy eating is not regarded as a central goal in life but as a resource for greater enjoyment of life [32]. It requires to accept a holistic viewpoint on health and real-life complexity as starting points for conducting and evaluating nutritional interventions. In salutogenesis, health results from the continuous everyday life interactions between the individual and the social, economic, cultural, physical, mental, and biochemical stressors [33]. The central concept in salutogenesis is the sense of coherence (SoC), which is the individual capability to mobilize and use health-promoting resources [33]. Understanding the resources that facilitate coping with these stressors in a health-promoting manner can help to unravel underlying principles of creating health [34]. Salutogenesis guided the development of the Salutogenic Intervention for Type 2 Diabetes (SALUD) program [35].

The aim of this study protocol is to outline the randomized controlled trial (RCT) in which the SALUD program is compared with the Dutch usual care for T2DM. The theoretical principle behind the program is mobilizing 2 important psychosocial resources required for organizing healthy eating and improved well-being in everyday life: self-identity and social support [35]. The present hypothesis is that mobilizing these resources will initiate a process of empowerment, leading to improved food literacy (direct effect) and improved health (indirect effects). Food literacy is the capability to navigate different food situations to make healthy eating choices [36], which is important for healthy eating and well-being in the long term. It is a measure that combines all relevant aspects for healthy eating in everyday life, that is, knowledge, skills, and behaviors. Hence, the primary objective is to determine the effect of the SALUD intervention on food literacy. Diet quality as well as several physical and psychosocial outcomes will be evaluated as secondary outcomes.

A 2-arm, semiblinded RCT will be conducted with 56 people with T2DM in the Netherlands. The whole study duration will be 24 weeks, with measurements planned at baseline (T0), after 12 weeks (postintervention; T1), and 24 weeks (follow-up; T2). Participants will be randomly allocated to the intervention group (SALUD program) or to the control group (usual T2DM care [37]). The SALUD program will consist of 1 individual intake session (at the start), group sessions for 12 consecutive weeks, and 1 booster session (at 24 weeks). The study procedure is outlined in Figure 1.

Medical Ethical Committee Oost-Nederland has granted medical ethics approval for this study on August 8, 2021 (Medical Ethical Committee number 2021-12949).

The study has been registered at Netherlands Trial Registry with registration number NL8963 [38] (registered on October 12, 2020).

The target population consists of Dutch-speaking men and women with T2DM. Inclusion criteria are as follows: officially diagnosed with T2DM by a general practitioner (GP); aged ≥18-75 years; fluent in Dutch, competent to make own decisions; and in possession of a computer device with internet connection. Exclusion criteria are as follows: pregnant, lactating, or intention to become pregnant in the upcoming 6 months; severe chronic illnesses or condition or comorbidities (other than T2DM, eg, kidney failure, severe form of chronic obstructive pulmonary disease [gold III or IV], heart failure [class 2-4], cancer, dementia, or severe depression); bariatric surgery in the past; and eating disorders (eg, anorexia or bulimia).

Participants are recruited via posters in GP practices, advertisements in local papers, social media, and websites of national diabetes organizations. In addition, several local GPs and practice nurses approach clients during regular health appointments or via phone or mail. Before commencement, all participants receive the information brochure that explains the study in detail and a phone call from the investigator for further explanation. At least 7 days after the phone call, participants sign an informed consent form and send it to the investigators before the start of the study. Participants are informed (written and verbally) that they will be randomly allocated to either the intervention or control group.

Randomization of participants will be performed by an independent researcher (not involved in the study) with use of a web-based randomization tool. The data analysis for the primary outcome will be run blinded by an independent researcher. Treatment codes will be broken after finalization of the data analyses. Participants and researchers are not blinded during the measurements owing to the nature of the intervention.

The participants in the control group receive the usual Dutch T2DM care according to the Dutch General Practitioners Association (Nederlands Huisartsen Genootschap) protocols [37]. The usual care of T2DM is based on medical treatment, education, and lifestyle advice. Newly diagnosed persons and persons who require insulin administration are referred to a dietician by the GP. In some cases, the GP recommends a referral to a psychologist for help with dietary advice or to a physical therapist to help with physical activity. The GP refers to a specialist in case of complications (such as retinopathy and nephropathy). Generally, a person with T2DM has 4 quarterly checkups with the practice nurse and 1 yearly checkup with the GP. However, the exact frequency of the appointments depends on the specific agreements between the health care provider and the patient.

The primary outcome is food literacy; secondary outcomes are self-efficacy, QoL, SoC, diet quality (Dutch Healthy Diet Index 2015 [DHD-15 index]), body weight, BMI, and waist-hip ratio (WHR). Self-efficacy and SoC are measured to evaluate the empowerment process of the SALUD program; diet quality to assess changes in nutritional intake; QoL to assess effects on psychosocial health; and body weight, BMI, and WHR to assess the effects on physical health (Figure 2). All measurements take place at the participants’ homes.

Consecutive to the 12th session, participants are invited to a focus group to provide feedback and share their personal experiences with the program. The participants’ partners are also invited to the focus group. One week before the focus group, the coach instructs the participants to bring an object that symbolizes their experiences with the SALUD and to think about what they liked or disliked about the program. The focus groups will be video recorded and used for qualitative effect analysis of the program. In the video recording, all participants (and partners) are asked to provide informed consent to record the session and to use the video recording for qualitative analysis. The following sample question will be used:

The video recording will be transcribed ad verbatim and pseudonymized. Two researchers of the research team analyze the transcripts independently. First, the transcripts are read and reread to engage with the data. Then, the transcripts are open coded on a descriptive and interpretative level. More specifically, transcripts will be coded on active components of the intervention, the mechanism of action, health-promoting resources, and the influence on everyday life. Subsequently, the codes are clustered in preliminary themes. Then, the 2 researchers compare and discuss the preliminary themes until consensus is reached. The overarching themes are the result of various discussions between the researchers. The themes will be displayed with pertinent participant quotes and detailed interpretative commentary. The Dutch quotes will be translated into English by a professional translator.

Sample size calculation is based on the known mean value and SD of the SPFLS (food literacy measure) for the Dutch general population [36]. There are no RCTs with lifestyle interventions in the scientific literature that assessed food literacy (measured with SPFLS or another measure for food literacy) as a primary outcome. However, as RCTs evaluating lifestyle interventions that targeted psychosocial outcomes such as self-efficacy have demonstrated effect size of between 1.1 and 1.7 [48,49], it seemed reasonable to expect an effect size of 0.9. With use of the Cohen d calculation for effect size, an anticipated mean of 4.20 SPFLS score in the intervention group, and an anticipated mean of 3.83 (SD 0.41) in the control group is expected at T1 (based on references values [36]). With a standard power of 80% and a significance level of 5%, a total of 19 participants per group (38 participants in total) are needed. Adding a 30% dropout rate, the ideal sample size was set at 28 participants per group (56 in total).

Statistical analyses are carried out using IBM SPSS Statistics for Windows, version 28 (released in 2021; IBM Corp). Two-sided P values <.05 are regarded as statistically significant. Analyses are performed with the intention-to-treat principle, meaning that all patients were analyzed in their original allocation group regardless of the extent to which they followed the intervention. Dropouts are counted, and reasons for dropping out are recorded. If missing is completely at random, all available data at T0 (baseline), T1 (12 weeks), and T2 (24 weeks) are used to conduct the analyses. If missing is at random, multiple imputation techniques are used to impute the missing data.

Scores are calculated for multiple-item instruments (ie, DGSES, WHOQOL-BREF, SoC-13, DHD-15 index, and SPFLS). Randomization will be checked, and variables will be normalized if needed. Descriptive statistics are performed to tabulate mean (or median) values of all study characteristics and baseline values of the independent variables. Chi-square tests (for categorical variables) and 2-tailed t tests (for continuous variables) are used to compare the descriptive statistics between study groups and to identify potential covariates. Data are tabulated as mean (SD) or median (IQR) if skewed. Categorical data are presented as n (%) where appropriate.

According to a strict data analysis protocol, primary and secondary outcomes are tested with mixed models. Secondary outcomes are also tested if the primary outcome is not statistically significant. The analyses scripts will be written by the researchers; the analysis of the primary outcome will be run by an independent researcher who is blind to the intervention code. Assumptions for mixed models are checked. The mixed models will test the difference compared with baseline in the scores and values (ie, dependent variables) between intervention and control (ie, independent variable) at T1 and T2. Fixed effects are the treatment condition, time, and the covariates; random effects are the repeated measures per participant at T0, T1, and T2. Multiple models are tested; a crude model, and additional models that are adjusted for confounders commonly adjusted for in literature (eg, sex, age, education, and BMI). Variables that statistically significantly differed at baseline are also included in a model as covariates. The final results are presented in a table displaying the estimated marginal means, 95% CIs, and P values for the treatment effect for all primary and secondary outcomes.

All participants receive a financial compensation of €150 (US $159.5) after completion of the study (ie, after finishing the last measurements at T2). The financial compensation is calculated based on the time commitment and invasiveness of the measurements.

The handling of personal data complies with the European Union General Data Protection Regulation, the Dutch Act on Implementation of the General Data Protection Regulation (in Dutch: Uitvoeringswet Algemene Verordening Gegevensbescherming), and the Findability, Accessibility, Interoperability, and Reusability principles. Participants are assigned a unique participant identifier number (ie, a random 4 digits number) that does not change during the study to ensure that data cannot be tracked back to the participants. This number is linked with the name, address, and date of birth of the participant. The identifier number is used on raw and processed data files. The file that contains the link between the participant’s information and the participant’s identifier number is stored on a password-secured location during the study. Participants indicate in the informed consent form whether their data may be used for other studies. If a participant indicates that they do not allow further use of their data, their data will be removed from the file that can be requested for future research. The file with the link between the participant identifier number and the personal data of the participant will be destroyed after finalizing the data files for future research.

A first strength of this study includes the RCT design, especially because previous Dutch “diabetes reversal” interventions have been evaluated via an observational pre- and posttest design [4,50]. Furthermore, there is also a shortage of evaluating salutogenic programs via RCTs [51]. Although RCTs are considered the gold standard for intervention evaluation, the RCT design has also some limitations for this particular study. The SALUD program is difficult to fit the experimental framework of an RCT, as it is a multicomponent intervention that cannot be strictly standardized [52-54]. Furthermore, RCTs are limited when the implementation context is a determinant of the result [52-54]. In the SALUD study, the success of the program will be dependent on the extent to which the program integrates with the everyday lives of the participants. Nevertheless, RCTs can be helpful for evaluating multicomponent interventions if close attention is paid to contextual factors via qualitative analysis [54]. Hence, a second strength is the proposed qualitative evaluation. This will increases the value of this RCT as it may help to understand the “black box” of the SALUD program (eg, the active components of the SALUD intervention, participant satisfaction, and points of improvement) [54]. In addition, the qualitative evaluation together with the quantitative evaluation may gather new insights regarding the theory of salutogenesis [51,54]. A third strength is that participants can follow the program on the web in the comfort of their own home: it saves the participants’ time and travel efforts. Considering the recent COVID-19 pandemic, it is also a safe manner of group interactions, and the group sessions can continue during a possible lockdown. It also provides valuable information on how to perform a group-based coaching program in a web-based setting. A third strength is the open and flexible nature of the SALUD program, which provides the lifestyle coach opportunities to tailor the program to situational factors and individual needs. A final strength is the inclusion of both physical and psychosocial health measurements that enable assessment of the program’s effect on different dimensions of health. However, because of limited budget, the study does not include HbA_1c and blood measurements to assess T2DM management.

A limitation is the limited follow-up period. Ideally, measurements would be repeated over 1, 5, or even more years to determine whether the program led to sustainable change. However, the current Dutch university and funding structures make it difficult to execute such long-term research projects (eg, temporary research contracts). Challenges the study may face are the recruitment; owing to the COVID-19 pandemic, health care professionals are overflooded with work and are short staffed. Especially in the case of a new infection peak, health care professionals may be unable to assist in the recruitment of participants. Furthermore, the intervention might be specifically suited for more motivated people because of the time investment and because self-identity and self-reflection require a proactive attitude. This might cause dropouts in the intervention group.

The proposed study will provide useful information on the effects of a salutogenic program on healthy eating and well-being in people with T2DM in a daily life setting. The SALUD program—if effective—may serve as a concrete, web-based tool to empower people with T2DM and guide them toward a healthier dietary pattern for the long run. In addition, the study may provide valuable information on what salutogenic interventions entail as well as new insights regarding the theory of salutogenesis. The combination of quantitative and qualitative measures allows a comprehensive evaluation of the intervention’s effects.