Background
Excessive gestational weight gain (GWG) is a risk factor for postpartum weight retention that contributes to long-term weight gain in women [
1‐
3]. It is also related to increased risk of obesity in offspring [
2,
4]. In the US in 2015, about 39% of normal weight, 61% of overweight, and 55% of obese women, gained more than the recommended amount of weight during pregnancy [
5]. Obesity experts have called for obesity prevention to begin during pregnancy and early infancy [
6].
A recent Cochrane review found diet and/or exercise interventions during pregnancy reduced the risk of excessive GWG by 20% [
7]. The estimate was robust and supported by high-quality evidence. Electronic communication technologies have the potential to reach large numbers of women during pregnancy with behavioral interventions to prevent excessive weight gain. However, there is currently a paucity of data on their effectiveness for this purpose [
8,
9]. There is a growing body of evidence that electronic communication interventions, often called e- and m-health interventions, are efficacious in several different domains of health behavior change including weight loss and weight gain prevention among non-pregnant individuals [
10,
11].
The aim of this study was to evaluate the effectiveness of a self-directed, integrated mobile phone and online behavior change intervention in preventing excessive GWG in a real-world setting.
Discussion
A socio-economically and racially diverse sample of women that was similar in terms of income and BMI to the population from which it was recruited entered into the study. The integrated online and mobile phone intervention that included 3 behavior change tools had no significant effect beyond the informational placebo control on any of the primary or secondary GWG outcomes. The result was robust to several sensitivity analyses.
In effectiveness trials with self-directed, online interventions focused on preventing excessive GWG, the level of engagement required to achieve a successful outcome is not known. In this trial, the level of engagement with the treatment arm specific website, also called adherence, was determined by research participants. In the intervention arm women, adherence at a level considered to be a minimum possibly effective dose was 46.1% (Table
2), despite the weekly e-mail reminders to login and view new content on the website that was also listed in the message.
Currently five relatively small pilot or feasibility efficacy trials in the literature used mobile and web-based communication technologies to address excessive GWG [
22‐
26]. Among those with positive results, engagement with behavior change interventions was higher than for this study, ranging from 61 to 86%. These findings suggest that the low prevalence of adherence in the intervention arm may have contributed to the null findings of this effectiveness trial with its self-directed, online intervention.
While the placebo control and intervention participants had a similar number of web page views of informational content (the majority of page views), the intervention arm members had only 6 page views of the behavioral change tools. In addition, 135 of the 1126 women in the intervention group (12%) only viewed placebo control content, essentially self-assigning themselves to the control group. The weight gain tracker was the most widely used behavior change tool with 70% of intervention arm women using the tool at least once [
14] and 27.3% using it consistently across pregnancy [
27]. In an analysis of usage of intervention website features, using the weight gain tracker consistently across pregnancy was associated with a reduction in the proportion of women with excessive total and weekly GWG and mean total GWG among not-low income women [
27,
28]. These results support the supposition that the low prevalence of consistent engagement with a key behavior change tool, the weight gain tracker, was not sufficient to produce a difference in weight outcomes between treatment arms and demonstrate the challenge of implementing the IOM guidelines in public health contexts.
Pregnancy has been described as a teachable moment when women are highly motivated to engage in healthy behaviors [
29]. For motivated women in the placebo control arm, the information and blogging tools may have been sufficient for preventing excessive GWG. Another RCT with a health coaching intervention to prevent excessive GWG that had an education only control condition also saw no difference in weight gain outcomes between treatment groups [
30]. However, no associations were found in this trial between the amount of use of the website and weight outcomes in the placebo control arm decreasing the likelihood of this explanation for the null findings [
28].
In this study, personal contact with study staff was limited and similar across treatment arms. A recent review of mobile phone interventions for weight management has shown that the amount of personal contact was significantly (
p = 0.05) related to differences in weight outcomes [
31].
This study has several limitations related to the measurement of weight outcomes and the design of the intervention. This study used a weight < 14 weeks gestation from women’s medical records for calculating total GWG. These medical records came from 29 different practices and clinics likely leading to measurement error and inflating the variability in GWG. However, the sensitivity analyses showed that using self-reported pre-pregnancy weights for calculating GWG, as is done in most studies of this topic, did not alter trial outcomes, indicating that the lack of an intervention effect is not highly sensitive to measurement error.
The study relied on self-reported weight and height for calculating BMI for strata assignment and identifying the upper limit for GWG for the weight gain tracker. If women under-reported their pre-pregnancy weight, which was likely, that would lead to a lower BMI and potential misclassification by BMI group. The results would be that under-reporting women would have been told by the study website to gain too much weight during pregnancy. This would decrease the likelihood that the intervention would have a positive effect compared to no weight advice given to the placebo control condition. In future trials, women should be weighed at entry in early pregnancy by research staff in order to correctly identify the upper limit of GWG and avoid giving incorrect advice on appropriate GWG.
Another limitation of the study was that the intervention was self-directed. This was done to meet the very diverse needs of the population of women [
32] and to mimic real-world conditions for this translational research trial. The goal was to provide an intervention that was similar to what women could find on their own in the online environment. This may have been an unwise decision. A recent review of the literature on e-behavioral nutrition interventions showed that structured, tailored interventions were more likely to be successful in achieving dietary change [
33].
A major strength of this study that supports the importance of publishing this null findings paper is that it is the first large RCT to test a self-directed, integrated e- and m-health intervention to prevent excessive GWG in a real-world setting, much like what any woman could access on her own in an online environment. A related strength is that the trial included a socio-economically, racially, and ethnically diverse sample and used a theory-based suite of behavior change intervention features.
Conclusion
This trial failed to show a positive effect of a self-directed, integrated online and mobile phone behavior change intervention on the proportion of the sample with excessive total GWG compared to an information only placebo control condition. The lack of effect was robust to several sensitivity analyses. There are likely multiple explanations for this result. Our view is that in the overall sample, the behavior change features in the intervention were used too little by too few women to produce a difference from the information only placebo control condition. The behavior change features, particularly the weight gain tracker, could possibly have had an effect if they had been more widely and consistently used across pregnancy. Finally, obtaining an accurate measured body weight very early in pregnancy is essential for giving the correct advice on GWG and for the calculation of the GWG outcome.
Several large trials are currently underway to evaluate the efficacy of various intervention approaches to preventing excessive GWG in overweight and obese women [
34]. The results of these trials plus the results of this trial have the potential to enhance the understanding of how to utilize e-and m-health interventions in population-wide obesity prevention programs during pregnancy.
Acknowledgements
The authors would like to thank professors Geri Gay and Jeff Niederdeppe from the Department of Communication at Cornell University for their work on developing the pregnancy intervention and Christian Schimke, Joe Tunis, and Jeremy Larkin of Form Collective for the design, development, and maintenance of the participant website. The authors are grateful to Eva Pressman, Chair of Obstetrics and Gynecology at the University of Rochester School of Medicine and Dentistry and Patrick Stover, Director of the Division of Nutritional Sciences at Cornell University for their administrative leadership that allowed completion of this project.