With the incidence of chronic lifestyle-related diseases rapidly increasing, cost-effective management is needed [1]. Enabling individually tailored care at a low cost is the ambition when introducing lifestyle interventions as a cornerstone in the prevention and management of chronic diseases [2-6]. For example, type 2 diabetes is strongly related to weight gain in adult life, and pathophysiological studies have established how and why people with type 2 diabetes can be returned to normal glucose control and have shown that even marginal weight loss can have a large impact on disease progression, bringing between 46% and 54% of patients with type 2 diabetes in remission [5,7]. App and web-based interventions aiming to promote a healthy lifestyle have attracted much attention owing to their potential for scalability and accessibility, low cost, privacy and user control, usability in municipal settings, and opportunities for real-time modifications and interactive advice [3,8]. While systematic reviews have concluded that app and web-based interventions can improve lifestyle behaviors, the sustainability of these interventions has been shown to be variable, and long-term sustainability needs to be demonstrated [3,8-12]. The literature shows that the usage of personal feedback from a known health coach on users’ own registrations, group support, and different behavioral change techniques (BCTs) is superior when compared with more automated services, especially on combining BCTs with face-to-face meetings [13].

A collaborative electronic health (eHealth) tool has shown promising results in weight loss among patients with diabetes when implemented in real-life settings [14]. In this study, our aim is three-fold as follows: (1) to measurably demonstrate the effect of a primary care, anchored, collaborative, eHealth lifestyle coaching program (long-term Lifestyle change InterVention and eHealth Application [LIVA] 2.0) on weight loss and diabetes regulation (hemoglobin A_1c [HbA_1c] level) in a strongly scientific and randomized controlled trial (RCT) setting; (2) to evidence the sustainability of a long-term intervention with follow-up measurements at 12 and 24 months; and (3) to assess the cost-effectiveness of a digital intervention in a municipal setting.

The collaborative eHealth tool (version 1.0) was a web-based solution that users accessed using an internet browser on their personal computers. The present version of the collaborative eHealth tool (version 2.0) called LIVA has been developed based on experiences from approximately 140,000 individuals, who used the collaborative eHealth tool 1.0, over a period of 15 years. Besides general experience from using the eHealth tool 1.0, the research team conducted several qualitative research studies with the following three important stakeholder groups within weight loss interventions: patients, general practitioners, and health professionals practicing in eHealth coaching (health coaches) [8,15-19]. The key findings from the qualitative interviews regarding the use of version 1.0 of the collaborative eHealth tool can be summarized as follows: (1) establishment of an empathic relationship with a health coach; (2) an intuitive design that enables ease of use for both users and health coaches; (3) different modes of communication channels allowing for active communication at all weight loss steps among users; and (4) an intuitive backend design, including a text and video library, and communication templates enabling the optimization of tailored personal quality coaching.

A user-driven approach to the design of both the user and coach interfaces has enabled ease of use and has eased communication flow [20], allowing for tailored communication between the health coaches and the end users at all steps of the weight loss program [21]. An important feature of the intervention continues to be the initial establishment of an empathic relationship between the user and the health coach, who delivers effective remote digital coaching based on the user’s own registration. Algorithms have not replaced health coaches, and instead, the features of the eHealth tool 1.0 enable individualized care at minimal effort from the health coach. Afshin et al concluded in 2016 that a direct interaction between a user and a health coach enhances the effectiveness of lifestyle interventions [3]. By establishing a personal relationship outside the digital environment, which is maintained through the platform with backend follow-ups, we believe that we are able to facilitate tailored care and sustained participant engagement over time with limited continued health coach input in the process of successfully changing and sustaining a lifestyle change [8,21].

With version 1.0 of the collaborative eHealth tool, we found that personal eHealth lifestyle coaching combined with various BCTs, such as tailored information, self-monitoring, lifestyle coaching, personal feedback, reminders, and face-to-face support, led to relevant weight loss during a 20-month period [21]. These results were confirmed in an English RCT in a municipality setting, showing weight loss of 5.4 kg among men with type 2 diabetes compared with weight loss of 2.8 kg in a control group receiving standard care [8]. A refinement of the collaborative eHealth tool 1.0 was implemented in 15 Danish municipalities between the summer of 2016 and the summer of 2017, with approximately 12,000 users on the eHealth platform. Besides smaller adjustments from the collaborative eHealth tool 1.0, version 2.0 is a smartphone solution, which is downloaded as an app called LIVA. A feasibility study among patients with type 2 diabetes using LIVA 2.0 in a cross-municipality setting demonstrated relevant weight loss of 4.7 kg among users who had been on the platform for over 90 days. Modeling the association between weight reduction and decreased health care costs indicated cost effectiveness in a municipal perspective 1 year after implementation [14].

Based on our previous research, we hypothesize that eHealth lifestyle coaching with the use of LIVA 2.0 will be effective in improving diet and increasing physical activity levels, thus reducing weight and improving HbA_1c levels after 1 year of intervention, with a sustained effect over time. This will result in increased health and quality of life (QOL) and decreased societal costs [22-27]. We expect that municipalities will find the intervention to be a cost-effective alternative for secondary prevention targeted at citizens who are at risk of developing chronic diseases, such as severe obesity, type 2 diabetes, cardiovascular disease, and cancer, and for tertiary prevention among patients with chronic type 2 diabetes, reducing the progression of diabetes and even resulting in its complete remission. To be able to investigate effects and cost-effectiveness among obese participants and obese patients with type 2 diabetes from a municipal perspective, the two target groups of the intervention are as follows: (1) obese citizens at risk of developing chronic diseases, such as severe obesity, type 2 diabetes, cardiovascular diseases, and cancer and (2) obese patients with type 2 diabetes. The intervention is compared with the conventional preventive program that the municipality offers to these two target groups. Furthermore, we investigate associations between participant characteristics and the success of LIVA 2.0 in providing novel insights regarding the associations between participant characteristics and success with a digital lifestyle intervention from both a 1-year and over 1-year perspective.

With regard to the primary outcome, the research question is as follows: What is the effect of LIVA 2.0 on users’ weight? With regard to the secondary outcome, the research question is as follows: What is the effect of LIVA 2.0 on users’ HbA_1c levels? With regard to the tertiary outcomes, the research questions are as follows: (1) What is the effect of LIVA 2.0 on the need for medicine use? (2) What is the effect of LIVA 2.0 on users’ QOL? (3) What is the cost-effectiveness from a municipal perspective? (4) Which participant characteristics are decisive for the effect and sustainability of LIVA 2.0?

A multicenter RCT with a 1-year intervention and 1-year retention period with collection of clinical and questionnaire data is supplemented with long-term register-based follow-up.

Participants are recruited through advertising on social media platforms, general practices, and patient organizations. Potential participants need to meet the inclusion criteria of the study (Textbox 1). Participants who would like to participate and who think they meet the inclusion criteria can register at the app URL [28]. After registration, potential participants are contacted by phone and are sent an email with more written and detailed information about the study and with an invitation to a face-to-face baseline meeting with a LIVA health coach. The individual face-to-face baseline meeting is conducted over 1 hour with a trained health care professional, who is a clinical dietitian by profession and who has been working with eHealth lifestyle coaching for at least 2 years (referred to as the health coach). The face-to-face baseline meeting is scheduled to take place within 7 to 14 days after the information material is sent by email to the potential participant in order to ensure that the participant has sufficient time to reflect on the decision to participate in the study. At the face-to-face baseline meeting, the participant is invited to bring a friend or a family member. The meeting takes place at the research unit for general practice or in the municipality where the participant lives. At the meeting, the health coach confirms that the participant meets the inclusion criteria and explains the details of the study to the participant, and at the end of the meeting, the coach obtains signed informed consent if the potential participant still wishes to participate. Thereafter, the participant is measured according to defined clinical indicators and fills out questionnaires.

The recruitment process is continued until the necessary number of eligible participants are included in the intervention and control groups within the following two target populations: (1) obese citizens at risk of developing a chronic disease and (2) obese patients with type 2 diabetes.

After all participants successfully complete a web-based questionnaire and undergo baseline measurements, they are randomized via an automated computer algorithm. This procedure ensures that drop-out characteristics can be recorded. Participants are randomized in a 60:40 manner, where 60% of the recruited participants are randomized to the intervention group and the remaining 40% are included in the control group. Randomization is controlled to ensure that 50% of participants in the intervention group and control group will be obese citizens, who have not been previously diagnosed with type 2 diabetes, and the other 50% of participants in the intervention group and control group will be patients diagnosed with type 2 diabetes (BMI ≥30 but ≤45 kg/m²) (Figure 1).

The intervention is not expected to cause any side effects or discomfort. The only recognized risk is in relation to eating disorders (pre-existing or developing during the study), and the dieticians and nurses involved in the study will specifically check for the indications. If an eating disorder is detected, the participant in question will be excluded from the study. The intervention is offered in addition to standard municipal care service, and participants can withdraw from the program at any time. The Regional Ethical Committee has approved the study according to Danish law (no. 18803). Participant data will be handled and stored in accordance with rules approved by the Danish Data Protection Agency. Permission to handle individual participant data from the national registries will be obtained from the Danish Data Protection Agency. All data will be analyzed in anonymized forms.

Participants in the intervention group receive login information for LIVA 2.0 at the first personal face-to-face meeting (either physical or digital), after which the health coach introduces the LIVA 2.0 app. After the first personal meeting, the same health coach will be coaching the participant throughout the period. When a health coach is on vacation or is sick for a short period, coaching is postponed. If the health coach is sick for a long period, a personal meeting with a new health coach is arranged to secure a personal relationship, after which the new health coach will be coaching the participant for the rest of the period. Based on our experiences from earlier studies, a half-time health coach can manage between 200 and 250 individuals. The participant and health coach together agree on goals for diet, physical exercise, sleep, and all other life areas of relevance to the participant. All goals are participant driven on the basis of the initial motivational interview with the health coach. The health coach is required to identify what health initiatives will benefit the participant the most based on the participant’s own wishes and to find out what is possible for the participant taking into account the participant’s personal barriers and facilitators. Using the app, the participant provides a daily record and also enters comments, concerns, and questions for the health coach, who can see the entire profile of the participant. The health coach provides individual asynchronous online consultation according to the participant’s needs and based on the participant’s registrations. The health coach encourages and praises goal attainment and endeavors to keep the participant motivated. Additionally, the health coach provides advice related to setting goals based on the SMART model (specific, measurable, attainable, relevant, and timely) [29] according to a predefined guideline structure (Table 1) [27,29].

Asynchronous online consultations are held weekly during the first 6 months. Thereafter, online consultations are held monthly over a period of 6 months. After 12 months, the participants enter a retention period for 12 months, where the health coach will follow the participants’ registrations and may provide up to a maximum of four coaching sessions during that year. We expect that approximately 20% of participants will not be ready for retention after 1 year and these participants will be offered yet another year of intervention. This evaluation is performed by the health coach. The health coach endeavors to maintain participation through phone calls, text messages, and, if necessary, face-to-face meetings to prevent drop out. After 24 months of follow-up, participants are followed through national registers for long-term follow-up on a number of predefined endpoints. Participants in the intervention group receive the standard municipal preventive care service, such as diabetes education, to the extent that municipalities normally provide these care services throughout the observation period.

Participants randomized to the control group are offered to receive the standard municipal secondary or tertiary preventive care service. A recent study aimed at examining Danish municipal weight loss care services and identifying and describing their content and structure found 234 different municipal weight loss care services. Although they were different, they most frequently contained information about diet and physical activity. They also sometimes included information about how to develop healthy habits, and a few of the care services even included the promotion of well-being and social participation [32]. The standard municipal care service within the control group is therefore not the same, but none of the care services resembled the LIVA 2.0 program. At follow-up measurements, the control group participants were asked to describe if they had participated in any interventions since baseline, and if so, describe the content of these interventions. This enabled a qualitative assessment and summary of the standard care services used in the control group.

Measurements are conducted by the health coach at baseline and at the 6-, 12-, and 24-month follow-ups. Data from national registers are collected before baseline and up to 3 years after the end of the intervention. Measurements are collected in a facility provided by the local municipality, the local general practitioner, or the clinical research center. Participants are scheduled for appointments by phone and with confirmation through e-mail. The facilities consist of a private consultation room with measurement equipment calibrated for use at relevant timelines and a waiting room area. Multimedia Appendix 2 describes all the included variables, definitions, categories, and sources for both assessments and explaining variables. No data from the participant records are obtained directly, and they are obtained only indirectly through registrations in national registers.

Within the scope of a cost-effectiveness analysis, expenditures related to the intervention, including acquisition, deployment, and operational costs, are compared with outcomes of the intervention in relation to the defined assessments. The cost effectiveness of the intervention is primarily assessed from the perspectives of the municipalities. The total cost of the intervention paid by a Danish municipality consists of investment costs as well as operating costs. Investment costs include expenditure for the training of health professionals as well as basic preparation cost. Operating costs cover the annual license fees for the individuals participating in the intervention as well as the individuals in the 1-year retention period. It is assumed that after 1 year in the intervention, participants will move to the retention period, where they will remain for 1 year; however, every year, 20% of the initial population will ultimately leave the intervention. Additionally, the operating costs include the salaries of the health professionals. The costs are compared to the effect of the intervention to evaluate cost effectiveness both in relation to the clinical effect and QOL [39].

A municipal budget impact analysis is performed, subtracting possible savings for the municipality in relation to health care costs, nursing costs, and lost productivity from the investment costs [40]. As part of the budget impact analysis, the annual rate of return of municipality investments is estimated. Follow-up data covering 1 year before baseline until 3 years after are collected from national registers (Multimedia Appendix 2). To investigate the possible impact of the intervention on health care service usage by participants, pharmaceutical consumption and municipal service usage as annual consumption are compared to a reference year 1 year before the intervention (baseline minus 1 year).

As part of the analysis, the following two scenarios are examined: (1) A₀, baseline scenario, where no intervention has been introduced. The health status of the examined population is expected to decrease, whereas the municipality costs are expected to increase; (2) A₁, alternative scenario, where the eHealth lifestyle coaching intervention has been introduced. The health status of the examined population is expected to increase owing to the intervention, whereas municipal costs are expected to decrease.

Since the target group involves participants with chronic diseases or having a risk of chronic diseases, the effect structure of a lifestyle change will first be visible over several years. These data will not be available within the current study design, and hence, the observed effects within the given time frame will be extrapolated over time. We will develop hypotheses for the effect structure according to the literature to be able to model the long-term (5 years) effect of the intervention [41,42].

All data are analyzed according to the intention-to-treat principle [43]. Analyses are mainly carried out according to the two target groups. However, stratification according to participant characteristics and user experiences is applied to investigate associations between success and the different characteristics. Ordinary least square regression and difference between groups over time are used to explore significant associations. Statistical significance is inferred at a two-tailed P value <.05. Robust standard errors are calculated. Data are analyzed by experts in biostatistics. After the study, all data will be made accessible on request in anonymized form to allow full peer scrutiny and facilitate secondary research.

The primary objective of this study is the measurement of changes in body weight and waist circumference. Based on a recent study by Haste et al, which evaluated a web-based weight loss intervention among men with diabetes, we expect a weight loss of at least 4.5 kg at 12 months in the intervention group as compared with 2.5 kg in the control group [8] (Textbox 2).

Based on prior experience, approximately 15% of participants will drop out during the first 2 weeks owing to technological challenges, etc. Likewise, after 3 months, approximately 20% of participants will lose motivation and be less active on the platform [8]. The coach will endeavor to maintain participation through phone calls, text messages, and, if necessary, face-to-face meetings. In the case of exclusion before the end of the trial (eg, due to pregnancy), the participant will be asked to complete a final questionnaire and have objective parameters measured in order to provide data for the intention-to-treat analysis.

From March 2018 to March 2019, 799 potential participants were evaluated for participation. A total of 340 met the inclusion criteria, consented to participate, filled out the web-based questionnaire, and were randomized into study groups. Among participants with type 2 diabetes, 100 (49 female participants) were randomized to the intervention group and 70 (32 female participants) were randomized to the standard care group. Among overweight participants, 100 (81 female participants) were randomized to the intervention group and 70 (52 female participants) were randomized to the standard care group.

The hypothesis is that the intervention group will demonstrate decreased body weight and a much higher percentage of patients with normalization of their HbA_1c levels as compared with the control group. A relevant percentage of participants in the intervention group are expected to decrease their use of glucose-lowering medications and improve their QoL much more as compared with the control group. Operational costs are expected to be lower than standard care and the intervention is expected to be cost-effective for the intervention group.

The first baseline assessments were carried out in March 2018. The trial has now reached the 12-month follow-up period for all included participants, and results are expected by the middle of 2020. The 2-year follow-up will be carried out between March 2020 and April 2021.