The primary aim of the intervention is to achieve an improvement in dietary intake and subsequent improvements in clinical variables and lifestyle factors through the use of a smartphone app for healthy eating habits among patients with type 2 diabetes. The content of the HAPPY smartphone app and the healthy eating behavior program is built into an existing digital platform developed by FRISQ AB. The program with its contents has been developed by researchers with experience in behavior change interventions (eg, physical activity), nutritionists, and active clinicians within primary and specialist care. It is based on a theoretical framework including the health belief model and the stages of change model, as well as on social cognitive theory [16]. Several techniques for behavior change, including general information, goal-setting strategies, self-monitoring, and feedback on performance, are included in the program [17]. The specific features of the HAPPY smartphone app are described below. The app is available both for iOS (version 11.4 and higher) and Android (version 5.0 and higher).

Each week the user will be introduced to a new topic on healthy eating habits (eg, vegetable intake). The specific topics for each of the 12 weeks of the active intervention are outlined in Table 1. Following a short written introduction to the current topic, the user is given an activity to perform (eg, to add one portion of vegetables per day during that week or replace sugar-sweetened beverages with water during a week). Daily progress is recorded in the app, and automatic reminders (eg, notifications if the user has not responded to an activity) or feedback on the activity (eg, “Thank you for participating in the activity”) will be given. All study participants receive the same introduction and activity to perform. Reminders and feedback depend on the actions taken by the user.

Each topic (ie, each week) comprises information, recipes, short fun facts, or advice (ie, edutainment and an activity related to the topic). Every Monday, information and an activity that is linked to the theme of the week are available in the app. Recipes to inspire healthy cooking and facts or advice on healthy eating habits are introduced during the week on Tuesday, Wednesday, Friday, and Saturday. Every Thursday, the user receives a reminder of the activity, and at the end of the week, an evaluation about the activity is sent to the user. Each action (ie, reading the information text, recipes, short facts, or advice, and then performing the activity) is marked as read or completed in the app by the user. The status of an action in the app (ie, read vs not read or completed vs not completed) is also visible to study personnel. This information, together with information on each time a user has logged on to the app, is saved within the system. The structure of one week in the healthy eating program is presented in Table 2. Examples from the HAPPY smartphone app are shown in Figure 2.

At baseline, participants in the intervention group will download and be connected to the digital platform, where the study personnel (ie, trained nutritionists) can follow the individual progress of the user during the 12-week course on healthy eating habits in the app. Participants in the control group will be offered to download and be connected to the app at the 3-month follow-up. To avoid overlap with the 4-day food record, the app will be activated on the first Monday following the meeting with study personnel (ie, at baseline or the 3-month follow-up) and the intervention will thereafter follow for 12 consecutive weeks. An individual user account on the digital platform will be created for each participant by study personnel. Users will identify themselves in the app using a personal e-identification .

A web-based questionnaire including assessment of lifestyle factors; a 4-day food record; clinical measurements of height, weight, waist circumference, blood pressure, and body composition; and blood sampling for measurement of HbA_1c and serum lipids will be given to all participants at baseline and at follow-up assessments at 3 and 6 months. At the 12-month follow-up, participants will respond to a final web-based questionnaire.

Dietary intake is measured using a 4-day food record over 4 consecutive days, including at least one weekend day. Participants are given a paper diary and instructed by the study personnel, who are trained nutritionists, to write down everything they eat and drink during the period of recording. Type of meal (ie, breakfast, snack, lunch, dinner, etc) is also recorded. The estimated amount of food and beverages consumed can be recorded in different units, including number of items (eg, number of potatoes), weight (eg, 125 g chicken), or unit of volume (eg, 2 dL of milk or 1 cup of coffee). The participants are requested to maintain their usual diet during the days of recording. The software program Dietist Net (Kost och Näringsdata AB) is used by study nutritionists to calculate nutrient intake from the food records; the nutritionists also check the dietary recordings for completeness when they are returned. In the event of items being recorded in an unspecified way (eg, “fish” or “yogurt” without further specification), information on the most commonly consumed fish or yogurt is obtained from nationwide data [20] and entered into the nutrient calculations. Standard portion sizes available for each food item in the Dietist Net software are used if the amount of food was not specified.

The participants complete the food records, as well as respond to a food frequency questionnaire (FFQ), at baseline and follow-up after 3 and 6 months. While the food records will give detailed information on types of food items, portion sizes, and frequency and timing of intake during the day, data from the FFQs allow for comparison to other studies as well as to follow-up of study participants in this study after 12 months.

Weight, height, waist circumference, and body composition are measured by study personnel at baseline and at 3- and 6-month follow-ups. Weight is measured to the nearest 0.1 kg in light clothing without shoes, and height is measured to the nearest cm in a standing position. Waist circumference is measured around the waist, approximately 2 cm above the umbilicus, to the nearest cm. Body weight and waist circumference are measured once on each occasion. Body composition, including body fat and fat-free mass, is measured using a digital body composition analyzer, Model BC-418 (Tanita). The scale utilizes an 8-electrode bioelectrical impedance analysis with current going from foot to hand and from hand to foot on both sides of the body.

Blood pressure—systolic and diastolic—is measured by the study personnel using the M7 Intelli IT automatic electronic monitor with Bluetooth technology (Omron). Measurements are done with each participant in a seated position with legs uncrossed after the participant has been sitting for at least 5 minutes.

Biomarkers are measured in fasting blood samples at baseline and at 3- and 6-month follow-ups. Study participants visit their closest primary care unit to have the blood samples taken. All samples are sent for analysis to the same lab connected to the Karolinska University Hospital in Stockholm, Sweden. HbA_1c (mmol/mol) is measured using the IFCC (International Federation for Clinical Chemistry and Laboratory Medicine) reference measurement procedure [21,22]. Triglycerides (mmol/L), total cholesterol (mmol/L), and high-density lipoprotein (HDL) cholesterol (mmol/L) are measured using the enzymatic method. Low-density lipoprotein (LDL) cholesterol (mmol/L) is calculated using the Friedewald equation:

The ratio of LDL cholesterol to HDL cholesterol is calculated from the levels of these lipoproteins.

Study participants respond to a web-based questionnaire at baseline and at follow-up after 3, 6, and 12 months. A link to the questionnaire is emailed to study participants on each occasion. A first reminder is sent if the participant has not responded within 1 week, and an additional reminder is sent 1 week after that if there is still no response. Each questionnaire takes approximately 30-45 minutes to complete and is comprised of several different sections as specified below. If not otherwise specified, the web-based questionnaire is comprised of the questions below.

Background information on marital status, education, and tobacco use (ie, smoking and snuff use) is assessed. Participants are also asked to report the year that they were diagnosed with type 2 diabetes; medication use related to hypertension, hyperlipidemia, and diabetes (ie, insulin or other treatment); and if they have changed their medication during the past 30 days. Background information is only collected from the baseline questionnaire.

Dietary intake is assessed using a 94-item semiquantitative FFQ. The FFQ has been validated previously in a random sample of Swedish men [23]. Spearman correlation coefficients for macronutrients comparing intake assessed using the FFQ, for an average of 14 24-hour recalls spread out over 1 year, were .44 (protein), .70 (total fat), .73 (carbohydrates), and .81 (alcohol). Participants report how often, on average, they consume each included food and beverage, including alcohol. An additional six questions developed by the Swedish National Food Agency, that are used in clinical practice, assessing overall dietary habits and risk use of alcohol are also included [24].

Eating behavior is assessed using the 21-item Three-Factor Eating Questionnaire [25]. The questionnaire includes questions related to cognitive restraint (six questions), uncontrolled eating (nine questions), and emotional eating (six questions).

Physical activity and sedentary behavior are assessed using general questions commonly used in clinical practice, regarding time spent doing physical activities at a moderate-intensity level or higher and sitting time [24], using the Active-Q questionnaire [26,27]. Active-Q includes 48 items and assesses habitual activity in four different domains: daily occupation, transportation, leisure time, and regular sports activities.

Health-related quality of life is assessed using the RAND-36, a questionnaire developed by the RAND Corporation [28]. RAND-36 comprises 36 questions within eight domains: physical functioning, role limitations due to physical health, role limitations due to emotional problems, energy and fatigue, emotional well-being, social functioning, pain, and general health. Responses from the individual domains are summarized into an overall physical component summary score and an overall mental component summary score. Additionally, a Swedish translation of the Life Engagement Test comprised of six questions is used to assess purpose in life [29].

Sleeping habits are assessed using a 13-item version of the Karolinska Sleep Questionnaire [30,31]. Participants report the time of going to bed the previous evening, time of waking up in the morning thereafter, and time from going to bed to falling asleep. They are also asked to report on sleep quality.

Stress level is assessed using a 14-item version of the Perceived Stress Scale developed by Cohen et al [32]. Participants are asked to respond to how often they perceive themselves to react in different, potentially stressful situations on a 5-point Likert scale, ranging from “Never” to “Very often.”

Diabetes self-efficacy and distress are assessed using the 20-item Swedish translation of the Problem Area in Diabetes Questionnaire [33]. Participants are asked to rate their distress with having diabetes on a 5-point Likert scale, ranging from “Not a problem” to “A serious problem.”

Perceivedsocialsupport for healthy eating habits is assessed by three questions regarding general support and support from friends and family. The questions were developed for this study and were based on the Physical Activity Social Support Scale [34]. Participants are asked to rate the statements “I experience support and encouragement to eat healthy food from people around me,” “...my friends,” and “...my family” using four response alternatives, ranging from “Agree completely” to “Do not agree at all.”

Evaluation questions about the usability of the app [35] and the contents of the healthy eating program are included in the 3-month follow-up questionnaire for participants in the intervention group and in the 6-month follow-up questionnaire for participants in the control group. This is done so that participants respond to the evaluation directly after having used the app. In total, users are asked to respond to 12 statements about the usability of the app (eg, “The app was easy to use”) and 12 statements regarding the content in the healthy eating program (eg, “The healthy eating program has made me reflect upon my dietary habits”) using a 5-point scale, ranging from “Do not agree at all” to “Completely agree.” In addition, users are also asked to leave a free-text comment regarding (1) what was good about the app and (2) what could be improved.