A Smartphone App (TRIANGLE) to Change Cardiometabolic Risk Behaviors in Women Following Gestational Diabetes Mellitus: Intervention Mapping Approach

Abstract

Background: Gestational diabetes mellitus (GDM) is the most common complication during pregnancy and is associated with an increased risk for the development of cardiometabolic diseases. Behavioral interventions can reduce this risk, but current solutions insufficiently address the requirements for such a program. The systematic development of a scalable mobile health (mHealth) promotion program for mothers during the first years post-GDM may contribute to solving this problem.

Objective: The aim of this project was to systematically plan and develop a theory- and evidence-based mHealth intervention to change cardiometabolic risk behaviors in women during the first 5 years post-GDM that meets women's expected standards of commercial health apps.

Methods: The intervention mapping steps 1 to 4 structured the systematic planning and development of the mHealth program described in this paper. Steps 1 and 2 led to a theory- and evidence-based logic model of change for cardiometabolic health. Based on this model, the prevention program was designed (step 3) and produced (step 4) in cooperation with industrial partners to ensure a high technological standard of the resulting smartphone app for the iPhone (Apple Inc). Step 4 included a user study with women during the first 5 years post-GDM once a beta version of the app ("TRIANGLE") was available. The user study comprised 2 test rounds of 1 week (n=5) and 4 weeks (n=6), respectively. The tests included validated questionnaires on user acceptance, user logs, and think-alouds with semistructured interviews.

Results: The novel TRIANGLE app is among the first self-paced smartphone apps for individual habit change in the 3 lifestyle areas of physical activity, nutrition, and psychosocial well-being. The 3 core features-a challenge system, human coaching, and a library-address 11 behavioral determinants with 39 behavior change methods to support lifestyle changes. Participants in the user study showed a high acceptance, high perceived quality, and high perceived impact of the TRIANGLE app on their health behaviors. Participants tested the app regularly, used it intuitively, and suggested improvements. We then adapted the TRIANGLE app according to the insights from the user study before the full TRIANGLE program production.

Conclusions: The intervention mapping approach was feasible to plan and develop an innovative and scalable smartphone solution for women during the first 5 years post-GDM. The resulting TRIANGLE intervention has the potential to support behavior change for cardiometabolic disease prevention. However, the app needs further refinement and testing in clinical trials. Intervention mapping steps 5 (implementation plan) and 6 (evaluation plan) may support the integration of the TRIANGLE intervention into routine care.

Trial registration: German Clinical Trials Register DRKS00012736; https://www.drks.de/DRKS00012736.

Keywords: cardiometabolic disease; diabetes prevention; gestational diabetes mellitus; health behavior; intervention mapping; mHealth; smartphone app.

Figure 1

TRIANGLE app onboarding and workflow.

Figure 2

Screenshots of the TRIANGLE app's core features. Challenge system: (A) activity screen based on active challenges and (B) weekly progress visualization per challenge. Coaching: (C) chat with personal coach and (D) in-app questionnaires. Library: (E) categories per theme and (F) exemplary article.

Figure 3

Results of the user version of the Mobile Application Rating Scale in the TRIANGLE user study stratified by visit and subscale (n=8; all values as mean [SD]).

Figure 4

App usage in the TRIANGLE user study (n=11; group 1: n=5; group 2: n=6). (A) Number of app activities per participant over time stratified by theme; 1 circle per active day, with circle size reflecting the number of activities. (B) Total number of app activities per participant stratified by theme. (C) Number of app activities per participant and time of day; 1 circle per active hour, with circle size reflecting the number of activities. (D) Percent of app activities per subfeature. P: participant.