Overweight and obese pregnant women are often at an increased risk of a series of maternal and offspring adverse outcomes [1-3]. In developed countries, overweight and obese women accounted for 30%-50% of fertile women [4-6]. Therefore, the high prevalence of overweight and obesity among pregnant women signifies a substantial burden to public health welfare worldwide. Current clinical care management of overweight and obese women during pregnancy consists of information on healthy eating practices according to standard dietary guidelines both worldwide [7-10] and in Singapore [11,12], in terms of restricting daily energy intake; balancing the proportion of complex carbohydrate (33%-40%), protein (20%), and fat (40%); and lowering cholesterol intake However, the reported effects were equivocal due to limitations such as low compliance, delayed feedback, late initiation, and inefficient delivery of intervention [13,14].

In recent years, the use of mobile technology (such as smartphone apps) for patient care has been increasing. The feasibility and efficacy of such technology has been proven for weight management among pediatric obese patients [15] and for glucose control in mothers with gestational diabetes mellitus during pregnancy [16]. Furthermore, a systematic review summarized 12 studies using phone-based reporting interventions such as video call, phone calls, short messaging service (SMS) and smartphone apps and showed consistent evidence that although such approaches could help pregnant women control their gestational weight gain, they were not effective in preventing other pregnancy outcomes such as gestational diabetes mellitus [17]. Another systematic review of four randomized controlled trials (RCTs) did not draw any firm conclusions on the effects of mobile app interventions during pregnancy on maternal knowledge, behavior change, and perinatal health outcomes due to heterogeneity of interventions, comparators, and outcome measures across all RCTs [18]. In addition, based on the intervention evaluation alone, none of these mobile technologies provide a real-time communication between end users and medical workers or dietitians, thus likely affecting the compliance and efficacy of the phone-based intervention. Given the high prevalence of smartphone app usage among pregnant women, more rigorous studies are needed to optimize the study design and implementation of these technologies in improving maternal health outcomes. In this pilot randomized controlled trial study, we tested a food-coaching intervention program delivered through a smartphone app in overweight and obese pregnant subjects without gestational diabetes mellitus and examined its feasibility, acceptance, and preliminary utility during an 8-week follow-up in second trimester.

Among 30 pregnant women recruited at baseline with 1:1 allocation, 26 (12 in the intervention group and 14 in the control group) completed the 8-week RCT. The detailed RCT flow chart is shown in Figure 1. Based on the log-in records, the uptake of this smartphone app among our intervention subjects was up to 90% at the beginning (n=15) and 70% at the end of the study (n=12), respectively. Based on the 12 returned user evaluation forms, 75% (n=9) of the app users found that the app was easy to operate, more than 80% (n=10) thought the food-coaching guidance was acceptably fast, and 90% (n=11) were satisfied and reported that the app had somewhat or greatly improved their own diet. Therefore, they would recommend the smartphone app to family and friends. Furthermore, according to the backlog records from the smartphone app provider, we assessed compliance by calculating the average log-in frequency in the first and second half of the 8-week follow-up. Each of the 12 subjects in the intervention group logged into the smartphone app 8 times per week in the first 4 weeks, on an average, but this number reduced to 2 times per week in the subsequent 4 weeks.

Tables 1-3 show the baseline and follow-up characteristics between the intervention and control groups. More participants met the Institute of Medicine recommendation for optimal gestational weight gain per week in the intervention group (4-week follow-up: 7/12, 58%; 8-week follow-up: 8/12, 67%) than in the control group (4-week follow-up: 8/15, 53%; 8-week follow-up: 5/14, 36%).

Although not significant, we found a trend among women in the intervention group who tended to have less weight gain than those in the control group at the 4-week follow-up (mean difference=–0.15 kg; 95% CI –1.51 to 1.21) and 8-week follow-up (mean difference=–0.08 kg; –1.80 to 1.63; Table 4). In addition, women in the intervention group tended to consume less cholesterol than those in the control group at the 4-week follow-up (mean difference=–64.87 mg; 95% CI –146.04 to 16.31) and 8-week follow-up (mean difference=–31.73 mg; 95% CI –102.91 to 39.45).

In this in a small sample RCT, we used a smartphone app to guide overweight and obese pregnant women to eat heathier in order to obtain optimal weight gain. We noted a high uptake of the smartphone app, as described above, and found evidence that pregnant women in the intervention group were more likely to have optimal gestational weight gain and consume less cholesterol compared with women in the control group. Although our pilot results were not significant, it provided proof of concept for the feasibility of applying such technology in future RCTs with a larger sample size, an earlier intervention onset, and a longer follow-up for overweight and obese pregnant women.

Although the dietary intervention has been proven to be effective in terms of weight gain control among overweight and obese pregnant women [28,29], there are huge variations in the delivery of dietary recommendation such as timing of intervention initiation, intensity of intervention, and feedback availability [13,28-31]. In a recently published systematic review summarizing 12 phone-based intervention studies on gestational weight gain control outcomes, most of the studies used phone calls or SMS to provide weight control guidelines, encouraged physical activity, and provided educational information about healthy nutrition. Several studies suggested that telephone communication is one of the most cost-effective tools to keep track of pregnant women’s health [32]. However, phone-based interventions are typically initiated by health care providers and might not effectively motivate pregnant women to self-manage their behaviors. Given the high prevalence of smartphone app usage among pregnant women to improve their healthy behaviors, no firm conclusion has been drawn in terms of improving perinatal outcomes [18]. One of the major drawbacks is the lack of rigorous studies examining self-managing and self-regulatory behaviors in pregnant women. Our food-coaching app is more flexible and interactive and has greater variety of communication modalities, thus overcoming all aforementioned limitations in current technology. Given the high smartphone usage rate (up to 80%) among Singaporeans [33] and high prevalence (up to 30%) of overweight and obesity among pregnant women in tertiary hospital settings [4], use of a mobile app to promote healthy dietary intake is feasible and likely a promising intervention strategy. Our findings showed feasibility and acceptability of such a food-coaching smartphone app. For example, 75% of the app users found that the app was easy to operate, and more than 80% thought the food-coaching guidance was acceptably fast. Furthermore, 90% reported that the app had somewhat or greatly improved their own diet.

Interestingly, we did observe a trend of lower energy, carbohydrate, protein, total fat, and cholesterol intake in the first 4 weeks of the intervention, while most of the macronutrients did not maintain the same trend at 8 weeks of follow-up, except for cholesterol. This may be because pregnant subjects were more inclined to remember food items that are high in cholesterol (ie, saturated fat, red meat, full-fat dairy products) and were able to make an effort to avoid such food items compared with others. However, further studies are needed to verify our findings, as such findings with a small sample might be biased. Although the effect estimates are in the desired direction and supported the proof of concept as the primary focus in this pilot RCT, we still need to further verify the utility of such a smartphone app in a larger targeted population with an earlier phase intervention during pregnancy and a longer follow-up.

The strength of our study included a prospective RCT study design and the use of standard protocols in anthropometric measures and dietary assessments. However, our study has significant limitations regarding the loss to follow-up and a small sample for analysis.

In conclusion, our study provides proof of concept that smartphone technology is feasible and acceptable in clinical dietary guidance among overweight and obese pregnant women. In the future, we will adopt this food-coaching smartphone app and test its utility in a larger setting with a targeted population, earlier intervention, and longer follow-up throughout pregnancy.