Table 2.

Implementations of desirable features [14] in the prototype app.

Feature	Implementation
1. Increase competence to manage GDMa with automatic feedback and interactive exploration.	We investigated this first desirable feature in the form of providing interactive self-tracking data exploration (Figure 1b) and feedback in the form of praise and suggestion (Figure 2a).
2. Increase autonomy by enabling personalization.	We sought to customize the app by providing 2 functionalities, (1) a habit tool (see Figure 1c) and (2) the ability to add a profile picture, name, and expected date of birth into the app.
3. Provide social support, especially from the partner.	We incorporated features to provide social support from both the partner and other women with GDM. For the partner, we designed a view where they could receive suggestions from the app (Figure 2a) and a “Shared habits” functionality that provided the possibility for the partners to do things together (Figure 2b). The content would be provided to the partner through a dedicated partner’s version of the GDM app. For women with GDM, we designed a community challenge where they could support each other to achieve a common goal (Figure 2c).
4. Support normal pregnancy and debunk myths about GDM.	The app provided a draft information section about pregnancy (eg, development of the fetus, as shown in Figure 1d), GDM, nutrition, and physical activity.
5. Support dual processing as pregnancy is life changing.	We supported dual processing by implementing 2 different types of visualizations, an area graph to support reflective thinking (Figure 1b) and the latest values to support habitual thinking (Figure 1a). We also created a habit formation and tracking tool to support more autonomous behavior. The reflective visualization was incorporated by visualizing blood glucose levels, nutrition, and physical activity in a single area graph (Figure 1b). To increase the ecological validity of interpreting data, we visualized actual data recorded from 1 GDM woman before the study. The glucose data (mmol/L) were recorded using Medtronic’s continuous glucose meter Enlite [32], nutrition data (macros in grams) were acquired from a food diary (kept for 3 days) and validated by dietitians, and physical activity data (steps) were recorded with a Garmin Vivosmart 3 activity bracelet. The habitual visualization was provided by showing the most recent values on the first page of the app. A small arrow (Figure 1a) indicating the trend in recent blood glucose levels was added. In addition, to follow customized habits, we designed a view where participants could track their habits (Figure 1c) against goals they had set up.
6. Integrate the app with normal pregnancy and existing health care services.	The app showed the gestational weeks on the entry page (Figure 1a), pregnancy information (Figure 1d), and recommendations by a health care professional (Figure 2a) related to this desirable feature.

^aGDM: gestational diabetes mellitus.