In the in-person condition, the intervention will be delivered via weekly 90-min group meetings for the first 15 weeks and then every other week in weeks 16 to 25. Intervention materials will be provided via handouts. The weight loss counselor will facilitate discussions about weekly topics. During recruitment and baseline procedures, participants will be told that they may bring their babies to in-person intervention meetings but that older children are not allowed to attend. Participants will receive up to US $5 to reimburse them for parking or bus fare.

The intervention will be delivered via a private (secret) Facebook group. Secret Facebook groups are restricted such that group membership is by invitation only, membership in the group does not appear in member profiles or search results, and only members can see content posted within the group [64]. During consent and the orientation webinar, study staff will outline how participants are encouraged to engage with each other (eg, read intervention posts and respond to the weight loss counselor and other participants daily), different ways to access intervention posts (eg, view in news feed and visit the group directly), and encourage participants to adjust their settings to get notifications of new posts in the group. On the first day of the intervention, the weight loss counselor posts a reminder of the format of the intervention, including encouraging participants to post at least one update a week.

We will schedule daily intervention posts to be posted from the weight loss counselor’s account using the Facebook post scheduling tool. There will be 2 posts daily for the first 15 weeks, and 1 post daily for weeks 16 to 25, corresponding to the decrease in contact in the in-person condition. Intervention posts have been developed based on our previous research [30,65-67] to cover the intervention content in that intervention module of the DPP lifestyle intervention through posts that provide information or resources, solicit sharing of thoughts or experiences or challenges related to the topic of the week, ask participants to set goals (posted on Mondays), ask participants to report their progress toward these goals (posted on Sundays), and ask participants to report their weekly weight change (posted on Fridays). Sample intervention posts are shown in Figure 1. The weight loss counselor will facilitate discussions about weekly topics by engaging participants in problem solving, assisting them in setting SMART goals (ie, goals that are Specific, Measurable, Attainable, Relevant, Time-Based), providing constructive feedback, sharing resources, providing positive reinforcement, and answering questions.

At 6 months, immediately following the intervention, participants will attend a focus group with other members of their weight loss group. Research staff will measure participants’ weight at the focus group visit. We will ask participants in both treatment conditions to elaborate on barriers and facilitators to participation. Participants in the Facebook condition will be asked for their reactions to and ratings of intervention posts that received low engagement. We will use their feedback to refine the format of low-engagement intervention posts to make them more engaging. Participants who do not attend the focus group will complete an individual follow-up visit that includes measurement of weight and an individual interview.

At 6 months, participants will also complete an online survey that includes assessment of current pregnancy, depressive symptoms [34], quality of life [40,41], infant feeding [42,43], social support for diet and physical activity changes [44], social media use [45], group cohesion [68,69], contamination, and intervention acceptability [30]. Participants will also report app-tracked time on Facebook over the past 7 days and their Facebook use habits, including what device(s) they used to access Facebook, what proportion of use was from their smartphone (vs other devices), the extent to which other people used Facebook on their phone [46], and whether they consciously changed their Facebook use as a result of becoming aware of the time spent on Facebook and, if so, in what way [70].

We expect that this follow-up visit will take 1 to 1.5 hours, and the survey will take 30 to 45 min. We will email participants a US $40 gift card after they complete the focus group and survey.

At 12 months (6 months after the end of the intervention), participants will attend a follow-up visit to measure weight and complete an online survey. Research staff will provide participants with instructions on how to remove the MyFitnessPal and time tracking apps from their phones. We will email participants a link to complete an online survey that includes measures of current pregnancy, depressive symptoms [34], quality of life [40,41], infant feeding [42,43], social support for diet and physical activity changes [44], and social media use [30,45]. Participants will also report app-tracked time on Facebook over the past 7 days and their Facebook use habits, including what device(s) they used to access Facebook, what proportion of use was from their smartphone (vs other devices), and the extent to which other people used Facebook on their phone [46]. We expect that the survey will take about 30 min, and the visit will take about 15 to 45 min. We will email participants a US $40 gift card after they complete the study visit and survey.

We will examine the feasibility of recruitment, sustained participation, contamination, retention, and feasibility of assessment procedures, particularly measurement of costs related to delivering and receiving the intervention.

We will calculate recruitment rates from the number of individuals contacted, screened, consented, and randomized, overall and by recruitment source. We will record the reasons for ineligibility and nonparticipation, including unwillingness to be randomized to either an online or in-person condition.

We will assess sustained participation in the intervention, that is, treatment retention. For the in-person condition, we will calculate sustained participation as time to last intervention session attended. Attendance at in-person intervention sessions will be recorded by the weight loss counselor at each meeting. For the Facebook condition, we will calculate sustained participation as time to last post, comment, or reaction (based on the date of post or comment reacted to) in the Facebook group. We will download objective engagement data from Facebook and calculate the date of last engagement.

At 6 months, participants will complete a survey that includes questions about whether they participated in other weight loss programs (online or in-person) and whether they sought weight loss support on Facebook or other online social networks [71]; and, if so, to what extent and reasons why they sought this support.

We will calculate retention as the proportion of participants who complete the 6- and 12-month follow-up assessments in each condition.

We will examine the extent and mechanisms of missing data on each measure to be included in the subsequent noninferiority trial, especially procedures used to capture costs. We will systematically track costs associated with delivery of both intervention conditions, capturing information on the costs that would be required to implement each intervention in practice (ie, outside the research context) using methods developed by others [72-74]. We have created an accounting system that captures administrative (eg, setting up the Facebook group and copying participants’ handouts), intervention delivery (eg, leading in-person intervention meetings and counseling via the Facebook group), and participant costs (eg, travel time to intervention meetings, time spent attending intervention meetings, and time spent on Facebook to participate in the intervention) for both conditions [75]. We will document any challenges with procedures for collecting data, including study records, participant-reported measures, measurement of weight, engagement data from Facebook, and time data recorded from tracking tools.

We will randomly select 20% of intervention modules, stratified by phase of the intervention: in-person group meetings weekly (weeks 1-15) or every other week (weeks 16-25). Study staff will review audiotapes from the in-person group sessions and content library of intervention posts from the Facebook condition and rate whether content objectives were met in each module. The weight loss counselor will receive feedback for intervention modules with less than 90% adherence to the protocol.

The possible risks of participating in this study include an injury while being physically active, possible discomfort from completing questionnaires, and breach of confidentiality. Participants will be screened for ability to engage in physical activity, and we will obtain medical clearance from each participant’s primary care provider or obstetrician/gynecologist. Participants will be encouraged to participate in physical activity they are comfortable with, while gradually scaling to meet intensity recommendations, and avoid any exercises that could lead to injury, pain, or discomfort. Participants who experience injury or discomfort from exercise will be asked to meet with a health care provider before returning to physical activity. We will keep study data stored on REDCap [36], on password-protected research drives, or in locked filing cabinets. We will encourage participants to review privacy policies of Facebook, MyFitnessPal, and any time tracking apps, and during the informed consent process, we will review with participants the privacy considerations of participating in an intervention delivered via a secret Facebook group (eg, who can see what they post and what access Facebook has to content they share in the group). We will record and follow up on any adverse events that occur during the intervention, regardless of likely relation to the intervention. We will report any events that are related or possibly related to study procedures to the University of Connecticut IRB within 24 to 48 hours regardless of level of severity, and all other adverse events those that are not related or unlikely related to study procedures, during annual continuing IRB review.

Although not expected to be related to study participation, over the course of the study, research staff may become aware of elevated depressive symptoms among our participants, as depression is common among postpartum women [78]. Study assessments at baseline, 6 months, and 12 months include the EPDS [34]. Elevated depressive symptoms and/or suicidal ideation at baseline are exclusionary. The weight loss counselor will monitor participants’ posts on the Facebook group and discussions at in-person intervention sessions for disclosures of feelings of depressed mood during the intervention phase and will call participants disclosing such feelings to assess for clinical depression and make appropriate referrals if necessary. At any point during the study, we will refer participants for mental health care and/or arrange immediate psychiatric evaluation, as appropriate, and will alert referring providers of elevated depressive symptomology as identified by the EPDS [34].

We will report recruitment using a Consolidated Standards of Reporting Trials (CONSORT) diagram [79,80]. We will compare yield from different recruitment approaches and will examine whether any eligibility criteria are excluding an inordinate proportion of otherwise eligible women. In particular, we will examine the number of otherwise eligible women excluded from participation because one or the other delivery mode is not feasible or unwillingness to be randomized to either treatment condition. If recruitment rates are lower than expected, we will adjust recruitment timelines as we plan the subsequent noninferiority trial.

We will compare sustained participation in both conditions. We will modify the delivery of both intervention conditions to address logistical and other barriers to sustained participation.

We will describe the extent of contamination and reasons women sought these extra sources of support. If a significant proportion of women report seeking weight loss support from other sources, we will adapt our intervention to better meet their needs to reduce the occurrence of contamination in the subsequent trial.

We will report retention rates in each condition using a CONSORT diagram [79,80]. If retention is lower than expected, we will explore reasons for dropout and make changes to the protocol to address these challenges. We will also examine the proportion of participants who become pregnant during the study period.

We will examine the extent and mechanisms of missing data, particularly measures of costs associated with delivering and receiving the intervention. For measures with an unacceptable amount of missingness, we will consider alternative measures in the subsequent noninferiority trial. We will resolve any issues arising in the collection of cost-related data and develop procedures for assessing time spent on Facebook to participate in the intervention.

We will report average percent weight loss in each treatment condition and the proportion that achieved clinically significant weight loss (≥5%).

The purpose of this pilot trial is to examine the feasibility, thus identifying modifications required before examining noninferiority in a large randomized controlled trial. Leon et al [81] state that “power analyses should not be presented in an application for a pilot study that does not propose inferential results.” As they and others recommend [81,82], we based the sample size on necessities for examining the feasibility, thus identifying modifications required to the design of the trial or study procedures before conducting a full-scale noninferiority trial. Conducting 2 waves allows us to iteratively refine how we deliver intervention content via Facebook while assessing the feasibility of recruitment and engagement under the conditions of the subsequent trial. Although we will aim to maximize retention in both conditions, with retention of 80% or greater to be acceptable, a priori, we decided that a retention rate in either condition lower than 60% would indicate that the noninferiority trial is not feasible as designed. With 36 participants per treatment condition, the 95% CI for the estimated retention rate will be within ±13% if observed retention is 80%. Given 36 participants per condition, the lower limit of the 95% CI for the observed retention rate in either treatment condition should not be lower than 60%.

Owing to the variability in estimated standard errors from pilot studies, estimated effect sizes from small pilot studies should not be used to determine sample size to adequately power randomized controlled trials to assess intervention efficacy [81,83]. Thus, as recommended [81,83], we will calculate sample size requirements for the subsequent full-scale noninferiority trial based on a clinically meaningful noninferiority margin for percent weight loss and variance in weight loss observed in previous adequately powered trials of postpartum weight loss interventions [84].