Postpartum women’s experiences in a randomized controlled trial of a web-based lifestyle intervention following Gestational Diabetes: a qualitative study

Abstract Introduction Gestational diabetes mellitus (GDM) is associated with an increased maternal risk for the development of type 2 diabetes (T2DM). We previously demonstrated in a randomized trial that a web-based postpartum lifestyle intervention program, Balance After Baby, increased weight loss among postpartum women with recent pregnancies complicated by GDM. The aim of this analysis is to identify the impact of the intervention on study participants as assessed by exit interviews after completion of the 12 month study. Methods We conducted structured exit interviews created with a concurrent-contextual design with subjects randomized to the intervention group at the conclusion of their participation (∼12 months) in the Balance After Baby study, with the objectives of 1) understanding the impact of the intervention on participants and their family members, 2) identifying which program components were most and least helpful, and 3) identifying the perceived best timing for diabetes prevention interventions in postpartum women with recent GDM. Results Seventy-nine percent (26/33) of eligible intervention participants participated in interviews. Participants noted changes in diet and physical activity as a result of the intervention. Several components of the intervention, particularly the online modules and support from the lifestyle coach, were perceived by intervention participants to have had a positive effect on personal and familial lifestyle change, while other components were less utilized, including the community forum, YMCA memberships, and pedometers. Nearly all participants felt that the timing in the intervention study, beginning about 6 weeks postpartum, was ideal. Discussion Results of this study identify the importance of individualized coaching, impact on family members, and demonstrate that postpartum women feel ready to make changes by 6 weeks postpartum. Findings from this study will help inform the development of future technologically-based lifestyle interventions for postpartum women with recent GDM.


Introduction
Gestational diabetes mellitus (GDM), defined as glucose intolerance with first onset or recognition during pregnancy after the first trimester [1,2], affects approximately 7% of all pregnancies in the United States, with prevalence increasing over time [3,4] and reaching 14% in some populations [1]. A history of GDM is associated with a 7-12 fold increased maternal risk of development of type 2 diabetes (T2DM) [2,5,6]. The postpartum period following a pregnancy complicated by GDM can serve as an important "window of opportunity" for interventions to reduce development of T2DM [7][8][9][10].
Diabetes prevention lifestyle interventions have been shown to be effective in decreasing development of T2DM in women with a history of GDM, but often these women are many years removed from their GDM pregnancy [11][12][13][14]. In addition, enrollment in diabetes prevention programs in this at-risk population is low [15], and many factors including identity as a mother, need for support from friends and family, life demands, finances and resources, time, and childcare responsibilities affect maternal adoption of healthy lifestyles post-GDM [16,17]. Given these factors, it is important to understand and integrate women's unique postpartum experiences when designing lifestyle interventions. We and several other investigators have employed technological-based methods for lifestyle interventions as a way to address many of the challenges to delivering postpartum interventions [18][19][20][21]. However, the perceptions of technology-based lifestyle interventions in women with recent GDM are not well understood but have been well-received in previous small, quantitative studies [17,22]. In response to these challenges, we conducted a randomized controlled trial of a web-based lifestyle intervention program, Balance After Baby (BAB), which demonstrated that women with recent GDM randomized to the intervention arm lost significantly more weight in the one year postpartum than women randomized to the control group [18]. Although tailored, technology-based adjuncts to postpartum care have been identified by women with previous GDM as being a potential way to improve postpartum care, the specific aspects of those interventions that may be most helpful remain unclear [22,23]. For the current study, we report on exit interviews at the conclusion of their participation, to better understand the impact of the intervention on the participants and their family members, including which components were most and least helpful, and perceptions regarding ideal timing for interventions in postpartum women with recent GDM.

Study population
The BAB trial (NCT01158131) randomized 75 women with recent GDM to investigate the effect of a webbased lifestyle intervention on weight loss in the first year postpartum. The study was conducted at Brigham and Women's Hospital in Boston, MA from June 2010 through September 2012. The Human Subjects committee at Brigham and Women's Hospital approved the study and all patients gave written informed consent. Full details have been published previously [16]. We included females age 18-45 with a diagnosis of GDM in their most recent pregnancy, no personal history of type 1 or 2 diabetes or bariatric surgery, and a postpartum BMI at the baseline visit 24-50 kg/m 2 (>22 for Asian women).
Participants attended the initial study visit for baseline measurements at 6 weeks postpartum. They were randomized at the end of the visit and all participants were given the National Diabetes Education Program's handout for women with prior GDM, "It's Never Too Early to Prevent Diabetes" [24]. All participants returned for measurements at 6 and 12 months postpartum. Women diagnosed with T2DM did not continue in the study.

Intervention
Women randomized to the intervention group were given the goal to return to their pre-pregnancy weight, and if still overweight, to lose 7% of their 6-week postpartum weight. They were asked to watch web-based educational audiovisual modules supporting healthy dietary choices, increased physical activity, and creating a healthy home environment once per week for the first 12 weeks. Participants were asked to connect with their lifestyle coach, a registered dietitian, weekly (via phone or email) for the first 12 weeks, every other week for the 2 nd 12 weeks (e.g. , and monthly thereafter. If participants were not reaching out to the coach at the appointed time, the coach would reach out to the participant by email, text, or phone call and offer possible times to connect. Since the coach was available remotely, she could be available for women needing more frequent interactions. Participants were asked to record their weekly steps using a pedometer and were offered the option of a free YMCA gym membership.

Data collection and structured interviews
At each study visit, we obtained biometric measurements and administered questionnaires, including the Edinburgh Postnatal Depression Scale (EPDS) [25]. At the end of their 12-month study visit, participants completed a 5-15 min duration structured qualitative interview with a research study team member regarding their experience in the study, which was manually transcribed during the interview by the interviewer. The interviewers included physician investigators and study research assistants, all of whom were trained in qualitative interviewing. Five major topics were covered in the interviews with intervention participants, including the impact of the intervention on participant and familial lifestyle change, the impact of the lifestyle coach on participant experience, components of the intervention that were most and least helpful, and the optimal timing of a postpartum lifestyle intervention. If a participant could not stay for the interview or did not attend the study visit, study staff called the participant and conducted the interview by telephone.

Analysis
Data were analyzed using descriptive analysis [26,27]. Atlas.ti software (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany) was used to aid in coding the structured interviews. Three researchers independently read and re-read interview transcripts and used an inductive approach to code while considering researchers' preconceptions using an interpretive stance from an epistemological perspective. Differences of opinion in coding were resolved through discussion until a consensus was reached. Coded quotes were clustered into related categories which were used to develop themes. Researchers reviewed themes and recoded data until final agreement was reached.

Results
Of 33 women in the intervention arm eligible for the 12-month visit, 26 (79%) participated in a structured interview. The interviews were conducted at the end of the study visits and women who were unable to stay did not complete the interview. We attempted to contact all of these women by phone after then study visit. One of the 26 intervention participants was interviewed by phone, with the others being interviewed in person. Table 1 shows the baseline sociodemographic characteristics of interviewed participants. The participants who did not complete interviews did not significantly differ from those participants who were interviewed (data not shown). Five themes were identified regarding aspects of the intervention that were most and least helpful to participants and their families, as well as perspectives about the effectiveness of the lifestyle group and the optimal intervention timing.

Impact of the intervention on participation lifestyle change
Intervention participants reported that the BAB program positively impacted both eating and physical activity. Specifically, participants discussed the value of learning about portion control, strategies for eating out, choosing healthier versions of their regular groceries, and creating a home environment conducive to healthy lifestyle changes. One woman explained, I've always eaten on the healthy side but needed portion control. This helped me understand what a portion is. My carbohydrate intake in general went down … Junk food isn't even an option anymore. I've moved away from eating butter or margarine because of your modules. (AK30) Another woman explained: The other big thing is using the techniques of creating the right environment … I never thought about that before. A big thing now is putting the food away before you eat. Then you don't keep going back. Knowing that if I am going to be really busy and if there is bad snack food in the house I know I have to have something healthy ready. Otherwise I will eat the bad snack. (GT103) Participants also reported increased physical activity as a result of the intervention. Many implemented suggestions from the BAB modules, such as dancing in the kitchen while making dinner, taking stairs instead of elevators, or parking farther away from buildings to increase walking distance.

Lifestyle coach's impact on participant experience
For the majority of intervention participants, the lifestyle coach was a positive aspect of the program.
The first time I talked to [the lifestyle coach] she pushed me to walk around the block. I was so tired. But I started doing it pushing the baby and now I have been walking. It is really nice … (CG89) Many women attributed their success in accomplishing their goals to a sense of accountability provided by the lifestyle coach. When asked what she found helpful, one participant explained, Accountability. I really enjoyed the weekly check-ins, the support. Just knowing I had to talk to [the lifestyle coach] kept it a priority, because it would have been so easy to focus on everybody else, work, family, etc … . I think she was instrumental in helping me Several women commented that they appreciated having a support system outside of their family unit or friend group whose entire role was to facilitate their success. One woman stated, [ Another said, I was dreading the call because I had to say out loud all the goals I hadn't been making … I don't think she was judging me but I felt so bad. I think talking about it makes you more accountable and I don't think I wanted to be accountable … (JH81)

Impact on family members
Some women identified benefits to their family resulting from their personal efforts to lead healthier lifestyles resulting from participation in the intervention. One mentioned the impact on her children, stating, My daughter (8) was a little chubby before. Now no junk food in the house and now she is not. Other people notice that she lost weight … (RS44) Other women believed the healthy changes seen in their family members were a direct result of the BAB resources. A participant explained, My husband has lost a ton of weight. He has looked as some of the modules and started using some of [the lifestyle coach's] tips as well and ended up losing 30þ lbs. (TE21) Other women found that the intervention provided opportunities to discuss health with and serve as role models for their children. One said, I think it has been great for my kids. We have always talked about nutrition, we would, but it was great to have them watch a parent actively make these choices. Now we talk about how important it is to take care of your body. This has given me an opportunity to model what is really important about how you take care of yourself … (SS60) Components of the intervention perceived to be not helpful or not widely utilized Although most intervention components were well received, several components were not perceived as helpful by participants and/or were not widely utilized, including the YMCA membership, the pedometer, and the community forum.
All women in the intervention arm of the study were offered a YMCA gym membership, many of which provided on-site childcare, at a location of their choosing. However, only one woman used the YMCA consistently throughout the intervention. Women cited scheduling conflicts, lack of proximity to a YMCA location, or having a preexisting membership at a different gym as reasons for not using the membership. Others stated that the locations they joined did not have childcare, or when childcare was available, some women were hesitant or unwilling to utilize it. One woman stated, All women in the intervention arm of the study were provided a pedometer so they could set tangible exercise goals and quantify their physical activity. Many women used the pedometers in the beginning of the study and found them helpful. One said "One day we went to Niagara Falls and I walked over 25,000 steps … It is a great tool. It is very motivating because you want to put the numbers on it." (ER46) However, by the end of the study, many women reported losing the pedometer, damaging it beyond repair (e.g. water damage), or being unable to use it due to dead batteries (although replacements were provided by the study). Other women suggested that they just got out of the habit. For example, "Early on I didthe first 6 months. It was helpful. I was planning to do it again but I didn't." (JH81) and "Yes, not last 3-4 months. Yes, when I was paying attention it was helpful. Cats would get it. Had to put out of sight, then forgot." (GDJP41).
To facilitate group discussion and internal support among intervention group members, a community forum was built into the study website. Despite attempts to foster discussion by the lifestyle coach, who posted prompts and questions related to healthy eating, physical activity, and being a new mom, the community forum was not very active. Several women reported that they would have liked to have a more active forum, but they did not post because of lack of engagement and participation by the overall group. One participant reported: "I did [post] toward the beginning. There wasn't much discussion, so I didn't go on after that … I would have used the community board if other people had been on there too." (JP41)

Optimal timing of a postpartum lifestyle intervention
When asked about the ideal time to start the lifestyle intervention, the majority of the women in the intervention arm reported that 6 weeks postpartum, the time of enrollment specified in the study design, was the ideal time to begin the lifestyle intervention program.
You did it right on. I was so ready to get my body back and start doing something about it. (SS60) I think the 6 week was good because it helps you get back on track, especially because during pregnancy I couldn't give in to any of my cravings because of GDM. So I think I went on a 6 week binge after I gave birth. Starting this program was like, okay, time to get back to reality. (LZ61) I think [6 weeks] was perfect because it was the time when I was ready … I was very preoccupied about the GDM, to make sure that I don't have problems later on, to make sure I will be around for my baby … .You've just gained all this weight, and to start right after [delivery] and see results month after month. It is the right time for it psychologically. (MC94) Sooner the better. Because otherwise you have already entered back into your old habits. This way it stops you. It puts you on guard. It informs you. It is more important when you just had GDM so you want to get rid of it. If our babies are 8 months we might feel like it (gestational diabetes) is gone, we don't care. Whereas if we just had it, it is fresh, we want to get rid of our diabetes. We're trying to get our lives back -we're trying to get back to normal. (ML79) Not all participants felt that 6 weeks was the optimal time to begin an intervention. When asked about the timing, one participant explained, "Depends on whether you are breastfeeding or not. If you are breastfeeding then it is hard to make changes because you are always hungry. Obviously, immediately making changes is better, but it was very hard with breastfeeding." (SR07)

Discussion
Through this qualitative study, we characterize the postpartum experiences of women with recent GDM randomized to a web-based lifestyle modification program providing insight into the impact of program and which components were or were not helpful, as well as the perceived optimal timing of a postpartum lifestyle intervention.
Overall, participants felt they benefited from the program, particularly by increasing their knowledge about positive health behaviors, such as portion control, and by developing skills necessary to execute lifestyle change, such as changing aspects of their home environment to facilitate healthy decision-making. Studies show that women with GDM often desire advice and support related to making lifestyle changes, particularly diet and exercise [28], but do not know where to begin or feel that traditional exercise and nutrition guidance is not compatible with the unique challenges of the postpartum period [29]. During pregnancy, women with GDM have reported stress and frustration because they did not have the skills to address cravings, portions, hunger, and eating in various social environments while being conscious of their blood sugar [30]. This has been demonstrated to continue into the postpartum period; women with a history of GDM reported that the greatest barriers to weight loss were difficulty adhering to a diet, difficulty managing hunger while on a diet, and family responsibilities superseding weight-loss [31]. Through the webbased educational audiovisual modules on the BAB website, women were introduced to a variety of methods by which they could make small changes which they could build into larger ones to improve their health. These modules were well-received, and participants reported improved nutrition, increased physical activity, and the ability to foster home environments conducive to healthy lifestyle changes. These findings are noteworthy, as the postpartum period is a particularly difficult time to make lifestyle changes [32], and women with a history of GDM face significant barriers to lifestyle change including lack of time or childcare, financial constraints, work-related difficulties, and feelings of guilt from being away from their children [16,33]. The reported increase in learning how to eat a healthy diet is significant, as women have reported that in the postpartum period, despite eating healthy diets immediately prior to and during pregnancy, they revert to a consuming a less healthy diet that may be linked to future obesity [29,34].
Most study participants reported that the lifestyle coach was an important aspect of the lifestyle modification program because she provided accountability and support. A recent review evaluating weight management in postpartum women found that interventions including diet and physical activity guidance in conjunction with individualized support and selfmonitoring are the most successful in promoting postpartum weight loss [32]. In the BAB intervention, the lifestyle coach provided personalized physical activity regimens, recipe ideas, and encouragement, via phone, text or email. Because postpartum women desire support and guidance regarding nutrition and exercise but often do not have access to these resources [28,32,35], the lifestyle coach's ability to provide tailored advice may have been particularly impactful. Additionally, as she was coaching remotely, she was both easily accessible to participants when it was most convenient to them and independently checked in when communication frequency was low. Women reported that this instilled a sense of accountability, which has been shown to be a critical aspect of successful health and wellness coaching [36]. In studies targeting weight-loss, participants consistently report on the importance of accountability to their continuing motivation to participate in such programs [37,38], particularly when they lack support from their partner [39]. A systematic review of multimodal, home-based interventions designed to prevent T2DM after GDM found that interventions that utilized individualized programming contributed to improved weight loss and dietary behaviors, particularly when cultural relevance was considered in the designing of the individualized programs [40]. Some of the women, however, did not find the lifestyle coach to be helpful, due to a perceived lack of the lifestyle coach's adaptability to financial and cultural differences, or a feeling of judgment. In future lifestyle interventions, utilizing a lifestyle coach or multiple lifestyle coaches with varying cultural and socioeconomic backgrounds may be helpful to participants. Further, the emphasis of the program should be on supporting participants and encouraging lifestyle change with positive reinforcement.
Study participants indicated that family members benefited from the BAB study as well, with several attributing their family members' weight loss and improved health behaviors to engagement with the BAB website and educational modules. Partners of women with pregnancy complications, including GDM, often report feeling excluded from maternity and postpartum care, as well as a having a desire to be more engaged in postpartum family lifestyle change [41]. It is possible that the accessibility of a web-based lifestyle modification program may allow for greater partner engagement and social support, through either the partners accessing the BAB materials on their own, or through the changes the participants implemented in the home setting. Many women with recent GDM desire social support [16] and report that partner support is critical to making lifestyle changes in the early postpartum period [29,32]. The accessibility and shareability of a web-based lifestyle modification program may affect positive lifestyle change in the participant as well as in her partner through increased individual and joint participation.
In addition to those women who felt their family members benefited directly from their participating in the BAB intervention, many participants attributed their family members' lifestyle changes to downstream or indirect effects of their personal efforts, e.g. weight loss seen in the child of a participant as a result of the mother's decision to remove junk food from the house. This finding is consistent with a study showing that when one parent participated in a weight-loss program the entire family benefited, including partners and children, with all family members becoming more physically active and having discussions regarding the importance of nutrition [39]. This spillover effect has been previously documented in spouses and partners [42][43][44][45] and overweight or obese children [46,47]. Because children with overweight parents are more likely to be overweight or obese themselves [48][49][50], a broad familial benefit of parent participation in a lifestyle intervention could have important impacts in stopping the cycle of obesity.
As discussed above, not every program component was well utilized. Reasons for not using the YMCA memberships are consistent with previously documented barriers to lifestyle change in this population, including lack of time or childcare, work-related obligations, and feeling guilt from being away from their children [16,33]. Despite not attending the YMCA, women reported increased physical activity relative to baseline by exercising at home, work, or around their neighborhoods, which was the type of physical activity emphasized in the program.
The lack of consistent pedometer use found in this study was comparable to that of other studies [51,52]. Many women said they partially tracked their physical activity using the pedometers provided to them, until the pedometers were lost, damaged, or ran out of batteries. Other women suggested that they just got out of the habit. The pedometers used in this study were simple trackers and not integrated into the intervention and website. Although participants entered their step counts into the website, there was no automated gamification or rewards function. In future studies, integrating newer technology, such as wearable bracelets, rings or watches that track activity and are integrated with the intervention and provide gamification may increase use of the device throughout the duration of the study. It has been suggested that women in particular may derive social benefit from gamification [53]. A study evaluating physical activity found that participants using Fitbits TM significantly increased their physical activity while those using traditional pedometers did not [54], and a recent meta-analysis and systematic review reports that Fitbit TM use promotes physical activity and weight-loss [55]. Pedometer studies in adults with cardiometabolic conditions and in women with recent GDM suggest that pedometers in combination with other intervention aspects, such as nutrition coaching, may be more effective than primarily pedometer-focused interventions [51,56,57]. More research is needed regarding the efficacy of pedometers or other similar exercisetracking devices in this population.
The limited community forum participation may be attributed to the intervention group's small size. The 1% rule, or the 90-9-1 principle, states that within internet communities, such as community forums, 90% of users observe and do not participate, 9% contribute sparingly, and 1% generates the vast majority of new content. Although this phenomenon has been largely informally reported, studies have observed this principle in digital health social networks [58] and in internet-based mental health peer support groups [59]. To create a community forum where there is active discussion may require enough participants to fit into the 90-9-1 principle and ensure the 1% is large enough to generate new content and discussion. The number of participants in the BAB study at any given time was likely too small to facilitate such discussion.
When asked about the ideal timing of a lifestyle intervention targeting postpartum mothers with recent GDM, most participants thought 6 weeks postpartum was ideal. The willingness of participants to begin at 6 weeks suggests that a program sensitive to barriers and facilitators to lifestyle change in the first year postpartum can be well accepted by women. This is in contrast to studies which report difficulty in engaging postpartum women in interventions in the first year postpartum [60][61][62]. However, we must consider that these findings may represent confirmation bias, as this is the only web-based postpartum intervention they experienced. In a qualitative study in women with a history of GDM, a majority of women reported at the ideal time for receiving additional support would be around the postnatal check [22].
Limitations of this study include the small sample size. Additionally, all women in our study were required to speak and read English and therefore may not be representative of the general population of women with GDM. Although researchers attempted to consider personal biases in the process of data collection, analysis, and reporting, a more formal process to address researcher reflexivity could have further decreased potential bias. Further, since the completion of this study, significant technological advances have been made, potentially limiting the generalizability of these results. We do, however, believe the components of the study identified as helpful by participants continue to be relevant and should be utilized in future development of technological interventions for women with recent GDM.
In this study, we analyzed exit interviews from 26 women with recent GDM at the conclusion of their participation in the intervention arm of the BAB randomized study. Many women found the intervention helpful, noting the importance of the lifestyle coach, the positive impact the intervention had on their family members, and the development of healthy skills they can use throughout their lives. However, not all aspects of the intervention were considered helpful, including the YMCA memberships, pedometers, and community forum. To improve the helpfulness and efficacy of a future similar intervention, having a larger intervention population to generate more community forum discussion and ensuring the lifestyle coach is able to address any financial and cultural variability in the participant population could have positive impacts on the outcome of the intervention. This study's results may help to inform the development of effective web-based lifestyle interventions in postpartum women with recent GDM and should be considered when designing future studies.