Digital Lifestyle Interventions to Support Healthy Gestational Weight Gain: Scoping Review

Introduction

Background

Pregnancy represents a critical window of opportunity for improving health outcomes for both mother and child. This transformative period catalyzes significant changes in a woman’s conception of self and social roles, fundamentally altering her capabilities, opportunities, and motivations for health behavior modification [1-3]. Throughout this paper, we use the term “woman’ to refer to individuals assigned female at birth, while acknowledging and respecting all gender identities. As a stage of transition, pregnancy presents an optimal time for encouraging healthy lifestyles, including weight management, with benefits that may persist well beyond pregnancy [4,5].

Current recognized medical guidelines, such as those by the Institute of Medicine (IOM), emphasize the importance of appropriate gestational weight gain (GWG) depending on prepregnancy BMI (ie, underweight BMI <18.5 kg/m²: 12.5‐18 kg; normal weight BMI 18.5‐24.9 kg/m²: 11.5‐16 kg; overweight BMI 25‐29.9 kg/m²: 7‐11.5 kg; and obese BMI ≥30 kg/m²: 5‐9 kg) for improved maternal and fetal outcomes [6,7]. Deviation from these recommendations – either excessive or insufficient GWG – has been associated with adverse outcomes. For women, these include increased risk of gestational diabetes, hypertensive disorders of pregnancy, cesarean delivery, and postpartum weight retention. For infants, risks encompass preterm birth, inappropriate birth weight, macrosomia, and increased likelihood of childhood obesity [8-10].

Despite the well-documented health implications, achieving recommended GWG remains a challenge for many women. Globally, 18%-30% of pregnant women gain insufficient weight, while 37%-51% exceed recommendations [8,11]. This challenge stems from multiple factors: many women are unaware that healthy weight gain targets differ per individual and depend on their preconception BMI [12,13], and even if women are acquainted with recommendations, they may not know how to operationalize them [14]. For example, the majority of pregnant women fail to meet dietary recommendations, such as those for vegetables, fats, and grains [13,15,16], and fall short of achieving the recommended 150 minutes of moderate-intensity aerobic activity per week [17-19]. In addition, common pregnancy symptoms such as nausea, fatigue, and pain, along with lack of knowledge and limited access to information, form additional barriers to maintaining healthy lifestyle practices [14,20-22].

Therefore, pregnant women require support in following healthy lifestyles and achieving healthy GWG. While health care professionals are ideally positioned to deliver this support through prenatal visits [3], they often face significant implementation barriers despite their motivation to address these guidelines [23,24]. For example, they may lack time, access to quality resources, training, and organizational support and perceive their efforts as being ineffective [23-25]. Consequently, many pregnant women report receiving either insufficient or inaccurate counseling on these topics, while they may desire it [6,14,26].

This issue is further exacerbated in maternity care deserts, where many women receive minimal or no prenatal care [27-29]. For example, between 2017 and 2023, in several low- and middle-income countries, less than 60% of pregnant women received at least 4 antenatal care visits while the World Health Organization recommends a minimum of 8 [3,30]. The problem extends to high-income countries, for example, the United States, where over 2.3 million women live in areas without obstetric services, with an additional 3 million having limited access to maternity care [27].

Digital Lifestyle Interventions for Managing Healthy GWG

Digital lifestyle interventions have emerged as a promising solution to the challenges of supporting healthy GWG. Several studies have shown the potential of using digital lifestyle programs in a pregnant population. For example, Redman et al [31] reported that their eHealth intervention combining digital and nondigital components positively impacted healthy GWG, while Feng et al [32] found reduced GWG in participants using a smartphone-based intervention compared with matched controls. However, some interventions have not achieved their intended outcomes [33,34], while others have shown differential effects depending on prepregnancy BMI classifications [35,36]. Despite these mixed outcomes, digital lifestyle interventions offer unique advantages in transcending geographical and social barriers, offering cost-effective, scalable support on top of usual care [28,37-39].

Literature reveals considerable variability in how digital lifestyle interventions are designed and implemented, which contributes to the inconsistent results observed across studies. Whether digital behavior change interventions achieve their intended outcomes depends on multiple factors working in concert, necessitating mapping of intervention characteristics, theoretical foundations, and reported findings to better understand the current landscape of approaches. Even among interventions that have shown promise, uncertainty persists regarding which specific design components and implementation features may contribute to their success.

Some studies suggest there are no clear optimal specifications for implementation features of GWG management interventions, such as duration, contact frequency, intensity of use, delivery format (group vs individual), or dietary approach [40]. Conversely, other research indicates that factors like the timing of intervention initiation, delivery modes (digital-only vs digital-mixed), and the type and frequency of digital components may influence outcomes [37,41]. A possible explanation for these mixed findings is how behavior change interventions are described: variation in reporting practices impedes our understanding of intervention mechanisms, evidence synthesis, and the development of more effective interventions [42].

Design and Implementation of Digital Lifestyle Interventions

Understanding the specific components that contribute to intervention success requires examination of both design and implementation features. Digital lifestyle interventions are a form of behavior change interventions, defined as programs that aim to change behavior with a clear objective and target group [43]. When designing interventions to change behavior, researchers draw on established theories to understand which components are most likely to produce change and what contextual factors might strengthen or weaken the intervention’s impact. The design of these interventions incorporates multiple behavior change techniques (BCTs), which represent the smallest active ingredients of an intervention capable of inducing behavior change [43-45]. BCTs are typically implemented based on theory, and understanding their proposed mechanisms of action may illuminate the processes by which BCTs influence behavior [46]. For example, the BCT “goal setting (behavior)” – defined as setting or agreeing on a goal defined in terms of the behavior achieved, such as agreeing on a 3-mile daily walking goal – works through increasing behavioral self-regulation.

To come to standardized reporting, researchers created a standardized classification system called the Behavior Change Technique Taxonomy to systematically identify and classify the techniques used to change behavior and better understand what makes behavior change interventions effective. Systematically identifying and quantifying BCTs used across interventions can help characterize intervention intensity to some degree, identify common patterns or clusters, and validate theory-based concepts [44,45]. This approach has proven value in previous reviews of interventions targeting smoking cessation [47] and physical activity (PA) [48].

The extent to which a BCT can achieve its intended purpose may vary depending on the target population, the specific behavior being addressed, and how the intervention is implemented [43]. As previously noted, implementation features such as timing, duration, frequency, and delivery modalities are critical as they represent the practical aspects of how interventions are delivered in real-world settings. It is therefore important to further explore how these features interact with BCTs to influence outcomes.

For healthy GWG interventions, implementation features are particularly important given the time-sensitive nature of pregnancy and the need to accommodate evolving physiological and psychological states. Research indicates that key implementation features influencing outcome include the timing of initiation (preconception, early pregnancy, or later stages), intervention duration (spanning the entire pregnancy or focusing on specific trimesters), frequency of delivery (daily contact vs weekly check-ins), and delivery modes (digital-only platforms vs hybrid approaches combining digital and face-to-face elements). The choice of a digital platform itself introduces additional implementation considerations (eg, user interface design, accessibility across devices, integration with existing health care systems, and the level of personalization offered) which are beyond the scope of this paper.

The relationship between these design and implementation features and positive outcomes for healthy GWG remains unclear without a systematic evaluation of existing evidence. Understanding which BCTs are most used, how they are typically implemented, and whether certain implementation approaches are associated with better outcomes requires a comprehensive mapping of the current intervention landscape.

Goal of This Study

To address the knowledge gap surrounding the characteristics and implementation of digital lifestyle interventions for healthy GWG, we conducted a scoping review focusing on interventions developed between 2014 and 2024. Our objectives were to systematically map intervention characteristics and identify design and implementation features that appear to be associated with positive outcomes. We examined theoretical foundations, intervention timing, duration, frequency, delivery modes, and the BCTs used across studies.

We addressed 2 primary research questions. First, what is the scope and nature of evidence on digital lifestyle interventions for healthy GWG? Second, what design and implementation features characterize digital interventions that report positive outcomes?

By addressing these questions, we aim to synthesize existing knowledge about intervention components and provide insights that may inform future research and development of digital interventions for supporting healthy GWG in addition to usual care. This paper presents a descriptive and narrative analysis of our findings.

Methods

Overview

We used a scoping review methodology to map the breadth of existing evidence and identify patterns in the design and implementation of digital lifestyle interventions. This methodology is particularly well-suited for exploring the extent of existing literature, mapping and summarizing evidence, identifying key characteristics, and informing future research directions [49]. This approach was most appropriate given the heterogeneity in digital intervention designs and reporting practices in this field, where the goal was to map evidence patterns rather than to conduct a quantitative synthesis for clinical decision-making.

Our review was guided by Chapter 10 of the Manual for Evidence Synthesis from the Joanna Briggs Institute, specifically the section on scoping reviews [49]. For drafting this paper, we followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension guidelines for scoping reviews) [50] (Checklist 1). A review protocol was developed for internal use but was not registered.

Eligibility Criteria

To define eligibility criteria, we used the Population, Concept, and Context framework [49], with population being “pregnant women” and concept “digital lifestyle interventions for managing gestational weight gain.” With “digital” interventions, we refer to those with at least 1 digital component (eg, app, SMS text message, and website), potentially in addition to nondigital components (eg, paper booklets and in-person sessions). The context was left unspecified to allow for the inclusion of studies conducted across diverse health care settings and populations.

Both articles on primary data (eg, pilot studies and randomized controlled trials [RCTs]) and on secondary data (eg, reviews and meta-analyses) were eligible. To ensure relevance to current digital health practices, the search was limited to articles published within the last 10 years. A preliminary scan of literature before 2014 indicated that earlier studies either did not fit our inclusion criteria or lacked sufficient focus on digital components. Given the rapid evolution of digital health technologies, we considered this time frame appropriate for capturing the most relevant and applicable evidence. Articles were excluded from our selection if they were not available in full in English; were not based on human participant data (eg, computer-generated data); were study protocols without results; were not related to pregnancy; did not include GWG as a primary or secondary outcome; were not peer-reviewed publications; did not investigate digital lifestyle interventions; or focused on high-risk pregnancies, such as placenta previa and gestational diabetes (to maintain focus on general GWG management rather than condition-specific interventions). A complete list of inclusion and exclusion criteria is provided in Multimedia Appendix 1.

Search Strategy

The search strategy was developed collaboratively by all 4 authors. To retrieve a comprehensive set of relevant studies, we translated our main research question into 3 broad search terms: “smartphone application,” “mHealth,” and “gestational weight gain.” These terms were used to search 4 electronic databases: PubMed, Embase, Cochrane, and Web of Science. These databases were selected for their extensive coverage of biomedical and life science literature. Our search strategy confirmed the appropriateness of this selection: while each additional database yielded some unique records, few of these met our inclusion criteria, suggesting that our chosen approach achieved near-saturation of the relevant literature.

The search strings were adapted to meet the syntax requirements of each database. Refer to Multimedia Appendix 2 for the exact strings used per database. The searches in PubMed, Embase, and Cochrane were carried out on March 8, 2024, and the search in Web of Science was completed on March 25, 2024. To match eligibility criteria, the searches were limited to articles published in English between 2014 and 2024 and within the defined population and context scope. Search results from each database were exported as .csv files and merged into a single Microsoft Excel file for screening and deduplication.

Article Selection

First, duplicates were removed using a custom Python script. It recoded each article’s title to an ID by removing all punctuation marks, turning all capitals into lowercase, and removing all spaces. These IDs were used to identify and delete duplicate entries. The remaining, nonduplicate articles were imported and stored in a Zotero library (Corporation for Digital Scholarship) for further screening.

Then, each article was screened independently by different pairs of the 4 authors using the eligibility checklist, examining titles, abstracts, and results sections. Ineligible articles were assigned exclusion criterion codes to ensure traceability. Any disagreements between reviewers were resolved through structured discussion, with a third author consulted when consensus could not be reached.

To enhance comprehensiveness, we also searched the internet for follow-up publications of excluded study protocols and hand-searched the reference lists of all included articles. Any additional articles identified through these methods underwent the same rigorous screening and selection process.

Data Extraction

Selected articles were read in full by at least 2 of the 4 authors in various combinations, and relevant information was extracted. If an article was deemed ineligible at this stage, it was assigned an exclusion criterion code and removed from the database. All exclusion decisions were discussed and agreed upon by at least 2 authors to ensure consistency and transparency.

Extracted data items included study type, study goal, population details, sample size, theoretical foundation, timing, duration, frequency, delivery mode, BCTs used, and results. The full list of data items and their definitions, along with the methodology for handling assumptions and simplifications for certain data items to ensure consistency across studies, is provided in Multimedia Appendix 3.

Data Synthesis

Data synthesis was conducted through a multistage descriptive and narrative analysis led by RAO, with support from the other authors. First, for primary and secondary data articles, we analyzed extracted study characteristics (publication year, country, study design, and so on) to map the scope and distribution of the evidence base. These findings were visualized using frequency tables.

Second, for primary data articles, we analyzed intervention characteristics including theoretical foundations, timing of initiation, duration, frequency, delivery mode, and BCTs. Third, for nonpilot primary data articles and secondary analysis of RCT data and meta-analyses, we categorized reported study outcomes as successful in achieving intended outcomes (yes or no) based on the study authors’ conclusions. We compared intervention characteristics across these outcome categories to identify patterns and potential distinguishing features. Throughout this data synthesis process, we used constant comparison methods to identify patterns, contradictions, and gaps in the evidence base.

Results

Overview

We identified 349 potentially eligible articles through database searches, with an additional 15 articles found through research protocols (n=3) and hand-searching reference lists (n=12). Following full-text screening, 44 articles met the eligibility criteria: 23 primary data articles and 21 secondary data articles (meta-analyses, systematic reviews, and scoping reviews that included articles on digital lifestyle interventions). Figure 1 shows the article identification and selection flow. The complete data extraction file can be found in Multimedia Appendix 4.

‎

Descriptives of Primary and Secondary Data Articles

Primary Data Articles

The 23 primary data articles included 10 pilot RCTs (43%) [33,51-59], 11 full RCTs (48%) [9,31,32,34,35,60-65], 1 real-world user data study (4%) [66], and 1 nonrandomized intervention study (4%) [67]. These studies reported on data gathered between 2011 and 2022, with 6 studies (26%) collecting data (partially) during the COVID-19 pandemic [32,35,59,64,65,67]. As illustrated in Figure 2, pilot RCTs were conducted in most years except 2016 and 2018. From 2016 onwards, full RCTs were also conducted yearly, although these were not follow-ups of the pilot studies. A notable peak in RCTs occurred between 2016 and 2019, potentially reflecting increased interest and funding in digital health interventions during that period.

‎

Studies were conducted in 10 different countries: United States (n=13, 57%) [9,31,34,51,56-63,66], Australia (n=2, 9%) [53,55], Canada (n=1, 4%) [67], China (n=1, 4%) [32], Finland (n=1, 4%) [33], Singapore (n=1, 4%) [54], Spain (n=1, 4%) [64], Sweden (n=1, 4%) [35], Taiwan (n=1, 4%) [65], and the United Kingdom (n=1, 4%) [52]. Sample sizes ranged from 12 to 15,468 participants (median 68, average 916). One study using commercial app data from over 15,000 women drove the high mean [66]. Table 1 summarizes the 23 primary data articles, with complete details in Multimedia Appendix 4.

Secondary Data Articles

The 21 secondary data articles included 9 meta-analyses (43%) [37,38,40,68-73], 5 systematic reviews (24%) [74-78], 3 scoping reviews (14%) [41,79,80], and 4 secondary analyses of RCT data (19%) [36,81-83]. Of the 9 meta-analyses, 4 (44%) investigated lifestyle interventions in general, while the other 5 (56%) specifically focused on digital lifestyle interventions. All systematic reviews targeted digital lifestyle interventions exclusively. Of the 3 scoping reviews, 2 (67%) addressed general lifestyle interventions and 1 (33%) examined digital interventions specifically [1]. Figure 3 illustrates the temporal trend in research methodologies used in the secondary data articles. Meta-analyses, the most prevalent approach, peaked in 2017 with 3 publications. Notably, 2 meta-analyses published in 2017 [38,69] specifically addressed the effectiveness of digital interventions for managing GWG, despite a limited number of RCTs on the topic at that time (Table 1). Systematic reviews maintained consistent usage throughout the search period. Scoping reviews were published less frequently overall but showed an increase after 2021, which roughly coincides with the publication of the PRISMA Extension for Scoping Reviews in 2018 [50]. Secondary analyses of RCT data were typically published 1-4 years following the original data collection.

The secondary data articles synthesized studies published between 1992 and 2022, incorporating 36 publications on average (range 7‐97, median 21), with an average sample size of 526 participants per article (range 53‐2833, median 89). Overlap with our primary data articles varied across the secondary data article types: meta-analyses shared an average of 3 articles (range 2‐6), systematic reviews 2.2 articles (range 0‐5), and scoping reviews 3 articles (range 1‐13) with our selection of 23 primary studies.

In terms of inclusion criteria, the majority of the secondary data articles primarily included RCTs or quasi-RCTs examining healthy lifestyles in pregnant women and GWG management, although systematic reviews allowed for wider inclusion criteria encompassing observational and qualitative studies.

The target populations were predominantly pregnant women in general, with exceptions including 3 articles (14%) focusing specifically on women with overweight or obesity [41,69,70], 1 article (5%) included both healthy women and those with (gestational) diabetes [71], and 1 article (5%) targeting users of pregnancy-related mobile phone interventions [78].

A summary of all included secondary data articles can be found in Table 2, with detailed data extraction available in Multimedia Appendix 4.

‎

Intervention Characteristics of Primary Data Articles

The 23 primary data articles analyzed 22 distinct interventions on which we focus in this section. In general, studies often provided insufficient detail in intervention descriptions, representing a gap in reporting practices that complicated data extraction and interpretation of successful design and implementation features. Despite this, we were able to extract the following information.

Most interventions targeted pregnant women in their first or second trimester who were overweight or obese, had singleton pregnancies, and no medical conditions or pregnancy complications that could affect metabolism or body weight. Exceptions included 2 studies (9%) that also included women with normal weight [34,67], 1 (5%) focusing on sedentary women [61], and 1 (5%) targeting women who had entered their weight gain in a specific app [66].

Recruitment was mostly done through prenatal clinics (n=9), followed by (university) hospital obstetric units (n=6), maternity units (n=3), and social media platforms (n=4). Most interventions used the 2009 IOM GWG guidelines (n=20, 91%). Regarding lifestyle, most interventions targeted both PA and diet (n=18, 82%), while 2 interventions (9%) focused solely on PA [55,57], and 2 (9%) on diet only [54,59].

Theoretical Foundation

Interventions were underpinned by a myriad of behavioral and social theories (Table 3), with social cognitive theory being most prevalent (n=11, 50%). Notably, 4 interventions (18%) used more than 1 theory [9,34,57,60], while 5 interventions (23%) provided no identifiable theoretical basis.

Timing

Intervention timing was relatively consistent across studies, with most starting between gestational weeks 10 and 16 on average. Specifically, 4 interventions started before 14 weeks of gestation (range 6‐13, 18%) [31,32,56,57]; 13 interventions started before 28 weeks of gestation (range 8‐28, 59%) [9,33-35,53,55,59-61,64-67]; and 5 interventions started between 14 and 28 weeks of gestation (range 14‐20, 23%) [51,52,54,58,62].

Duration

The average intervention duration was 21.1 (SD 7.3, range 6‐29, median 24) weeks. Interventions starting in trimester 1 (<14 wk gestation) lasted about 7 weeks longer (average 28, SD 1.0 wk) than those starting in either trimester 1 or 2 combined (<28 wk gestation; average 21, SD 6.9 wk) and about 11 week longer than those starting in trimester 2 (≥14 wk and <28 wk gestation; average 16 wk, SD 1.5 wk).

Frequency

Frequency patterns varied considerably across studies, ranging from on-demand content access to structured daily check-ins combined with in-person sessions and some interventions combining multiple formats (detailed descriptions in Multimedia Appendix 4).

Delivery Mode

Most interventions were digital-only (n=13, 59%) [32-35,51,54,58-60,64-67], with 7 incorporating mobile health (mHealth) tools, such as smart watches, smart scales, or pedometers [34,35,59,64-67]. The other 9 interventions used digital-mixed formats (41%) [9,31,52,53,55-57,61,62], of which 3 also used mHealth tools [55-57].

BCTs

Across all interventions, a total of 227 BCTs were identified, averaging 9 BCTs per intervention (range 3‐18). The majority of BCTs originated from the clusters “goals and planning,” “feedback and monitoring,” and “social support.” The 3 most common individual BCTs were “self-monitoring of behavior” (code 2.3), “goal setting (behavior)” (code 1.1), and “self-monitoring of outcome(s) of behavior” (code 2.4). A full taxonomy of BCTs and their frequency is presented in Table 4.

Outcomes

Digital lifestyle interventions showed varied outcomes for healthy GWG across both primary and secondary studies, with certain design and implementation features appearing more frequently in interventions that achieved their intended outcomes. Among 12 primary data articles coded for achieving their intended outcomes, 7 (58%) reported beneficial effects of digital lifestyle interventions on healthy GWG [9,31,32,60,62,64,66]. These interventions resulted in lower total mean or median GWG (average Cohen d=0.42; range 0.33-0.55) [9,32,60,62,64], reduced weekly rate of GWG [9,64], reduced risk for excessive GWG [66], and improved adherence to GWG guidelines [9,31,60,62]. The remaining 5 studies (42%) did not achieve their intended outcome [34,35,61,65,67]. Although the 10 pilot RCTs were underpowered for definitive conclusions and thus not coded for achieving their intended outcomes (refer to Methods section), most reported beneficial trends for healthy GWG (n=7, 70%) [51-56,59] and 2 (20%) described better adherence to IOM guidelines [52,54].

Secondary data articles reinforce these findings. Of the 9 meta-analyses, 7 (78%) found that interventions achieved their intended outcomes [37,40,69-73], with pregnant women in intervention groups gaining an average of 1.4 (SD 0.71) kg less than control groups. All 4 secondary analyses of RCT data reported positive outcomes for digital lifestyle interventions, including effects for subgroups of participants with mid-high income who consistently used digital interventions (average −2.1, SD 0.26 kg) [81,82], associations between increased digital engagement with lower GWG [83], and differential effects by BMI status [36]. All 5 systematic reviews found digital lifestyle interventions promising for healthy GWG [74-78]. The 3 scoping reviews similarly found that approximately half of the digital interventions included improved healthy GWG outcomes [41,79,80].

Design and Implementation Features Linked to Achieving Outcomes

Figure 4 represents a summary of design and implementation features of the 12 interventions that we coded for achieving their intended outcomes, followed by a description of the main findings per feature.

‎

Theoretical Foundation

The relationship between theoretical grounding and intervention outcomes was inconsistent across studies. In the primary data articles, social cognitive theory was the most frequently used theoretical framework, used in interventions with varying outcomes. Among the 7 interventions that achieved their stated objectives, 2 were based on 2 theories (combining the social cognitive theory and the social ecological model [60] [14%] or the transtheoretical model of change [9] [14%]). Notably, 3 interventions (43%) that met their objectives reported no explicit theoretical foundation [32,62,66]. Of the 5 interventions that did not achieve their stated goals, 3 (60%) were based on social cognitive theory [35,61,65], and 1 (20%) combined the integrative model of behavior prediction with the behavior model for persuasive design [34]. Secondary data articles did not examine theoretical foundations as determinants of intervention outcomes.

Timing

Primary data articles showed that interventions achieving their stated objectives were initiated earlier in pregnancy (average 12, SD 4.1 wk, range 6‐28 wk) compared with those that did not (average 14, SD 1.9 wk, range 10‐20 wk, Figure 4). This pattern was consistent with findings from secondary data sources reporting that digital interventions meeting their outcomes typically commenced in early-to-mid pregnancy [40,41,70,79,80].

Duration

Intervention duration varied between studies with different outcomes, with interventions that achieved their stated objectives having longer durations (mean 25.5, SD 3.3 wk, range 21‐29 wk, median 27 wk) compared with those that did not (mean 21.4, SD 6.9 wk, range 9‐28 wk, median 25 wk). Secondary data sources described similar patterns, with Barroso et al [41] noting that digital interventions meeting their outcomes typically spanned the remaining pregnancy duration. Hutchesson et al [80] identified intervention duration and intensity as areas requiring further research.

Frequency

As stated earlier, the frequency of intervention delivery showed large differences between interventions. There were no clear differences in delivery cadence between interventions that met their intended outcomes and those that did not. Secondary data articles also reported that frequency did not have a consistent impact on interventions achieving their objectives [37,38,40,71].

Delivery Mode

Primary data articles revealed variation in outcomes between delivery modes, with 4 digital-mixed interventions (57%) achieving their intended outcomes [9,31,60,62] and 3 digital-only ones (43%) [32,64,66]. Among the interventions that did not meet their objectives, 1 was digital-mixed (20%) [61] and 4 were digital-only (80%) [34,35,65,67]. Secondary data sources corroborate these patterns, with Barroso et al [41] noting that digital interventions attaining their goals typically included frequent touch points with well-trained lifestyle coaches through in-person sessions. Hutchesson et al [80] similarly identified delivery modes in interventions and involvement of health care providers as areas for further investigation.

BCTs

Primary data articles indicated that interventions meeting their objectives used the same number of BCTs (average 10, SD 4.3, range 3‐16) compared with those that did not (average 10, SD 2.6, range 6‐14). Table 5 maps the BCTs used across interventions that achieved their stated objectives and those that did not meet their goals. BCTs used commonly (≥70%) across both intervention categories were goal setting (behavior) and self-monitoring of outcome(s) of behavior.

In total, 5 BCTs showed notable differences in usage patterns between intervention categories, appearing ≥25% more frequently in interventions that achieved their objectives: goal setting (outcome; 71% vs 40%), discrepancy between current behavior and goal (43% vs 0%), self-monitoring of behavior (86% vs 60%), social support (unspecified; 71% vs 20%), and credible source (71% vs 40%). Conversely, interventions that did not meet their objectives used problem solving (60% vs 29%) and social support (emotional; 40% vs 14%) ≥25% more frequently than those that achieved their goals.

Interestingly, with respect to delivery modes, we observed different patterns in the number of BCTs used in interventions meeting their objective: digital-mixed interventions (n=4, 57%) used 13 BCTs, while digital-only ones used 7 (n=3, 43%). Most commonly used BCTs in successful digital-mixed interventions were goal setting (behavior and outcome) and self-monitoring of behavior. In contrast, successful digital-only interventions primarily relied on self-monitoring of outcome(s) of behavior, information about health consequences, and a credible source.

From the secondary data articles, only the meta-analysis by Rhodes et al [68] examined BCTs in relation to intervention outcomes, reporting an average of 9 BCTs per intervention and confirming that interventions meeting their objectives used twice as many BCTs from the groups “goals and planning” and “feedback and monitoring” categories compared with interventions that did not. They identified 7 BCTs commonly present across interventions that attained their goal: goal setting (behavior), problem solving, review of behavior goals, feedback on behavior, social support, information about health consequences, and information about emotional consequences. Notably, review of behavior goal(s) was observed exclusively in interventions achieving their objectives, with the authors concluding that interactivity is crucial for driving engagement and improving intervention success. A secondary analysis of RCT data [36] described BMI-specific patterns for goal setting, noting beneficial associations for women with normal BMI but adverse associations for those with higher BMI.

Discussion

Evidence Landscape of Digital Lifestyle Interventions for Healthy GWG

Digital lifestyle interventions alongside usual care show potential for supporting healthy lifestyles and GWG management among pregnant women. This scoping review systematically mapped existing literature to identify key design and implementation features, examining theoretical frameworks, timing, duration, frequency, delivery modes, and BCTs. Our mapping of 44 articles (23 primary and 21 secondary data) revealed diverse approaches to digital lifestyle interventions targeting GWG management and highlighted gaps in current research. These findings provide insights into identifying research priorities and informing future intervention development.

Our initial research question focused on the scope and nature of evidence regarding digital lifestyle interventions for GWG. The primary data articles revealed diverse patterns in reported outcomes. Among the 23 primary studies reviewed [9,31-35,51-67], 12 were coded for success [9,31,32,34,35,60-62,64-67] with 7 achieving their intended outcomes [9,31,32,60,62,64,66]. In addition, 7 of the 10 pilot studies [33,51-59], although underpowered, indicated positive trends [51-56,59]. Notably, several studies that did not report improvements in GWG did document benefits such as increased PA and improved dietary quality [35,61,67]. The secondary data articles, particularly meta-analyses, commonly reported that digital interventions supported GWG, although the reported impacts were described as modest. These findings are consistent with broader literature suggesting potential benefits of digital lifestyle interventions for GWG [68].

The substantial body of secondary literature identified in this review indicates ongoing interest in synthesizing evidence on digital lifestyle interventions and their components. This is also reflected in the temporal analysis of the systematic reviews in our selection that revealed an evolving research focus and improvements in the quality of digital lifestyle interventions. Early work by O’Brien et al [74] in 2014 concluded that digital interventions hold potential for complementing traditional health care models but emphasized the low quality and quantity of the interventions available. By 2018, Overdijkink et al [76] reported that app interventions had good overall usability and efficacy, although high dropout rates for several apps persisted. Mertens et al [77] in 2019 confirmed that the quality of interventions improved, but challenges with accessibility and engagement with target populations remained. Subsequently, Hussain et al [78] reinforced the potential of digital interventions for managing healthy GWG but highlighted the continued heterogeneity of studies and their generally small sample sizes. They also called for cost-effectiveness research. In 2021, Iyawa et al [75] reported that apps could positively impact self-management, such as GWG, during pregnancy while noting gaps in longitudinal studies and research in low- and middle-income countries.

The continued heterogeneity of interventions described by Hussain et al [78] was also evident in the primary data articles we studied, with differences in study populations (such as participants from different prepregnancy BMI categories), intervention duration, intensity, and delivery modes. This diversity was explicitly acknowledged in several meta-analyses as a challenge for evidence synthesis [37,40,68,69,71]. In addition, studies used varied approaches to outcome measurement, with some focusing on adherence to IOM guidelines [9,31,60,62], others on the total average [9,31,60,62,63] or median GWG [32,64], weekly weight gain rates [9], patterns of intervention engagement in relation to weight outcomes [66], or any combination of these. This heterogeneity in both intervention design and outcome measurement approaches underscores the complexity of this research field. It points to areas where greater methodological consistency could benefit future research synthesis efforts.

Patterns in Key Design and Implementation Features

Overview

Our second research question explored the key components and characteristics of digital lifestyle interventions for healthy GWG. This exploration involved mapping intervention components and examining patterns across different types of interventions to identify commonly reported features. We examined intervention theoretical foundations, timing, duration, frequency, delivery modes, and BCTs.

A significant gap emerged regarding the comprehensiveness of intervention descriptions across the reviewed articles. Many lacked sufficient detail for reliable interpretation of components and potential research replication. This finding was reinforced by our secondary data articles, which similarly identified gaps in reporting on interventions and their outcomes as a key limitation in the field. These documentation limitations align with previous research highlighting challenges in comparing interventions and synthesizing evidence about their characteristics [42,45,46,68,69,71]. Despite the publication of the Behavior Change Technique Taxonomy in 2013 [44], and the fact that included studies were published afterwards, the majority have not incorporated these standardized reporting guidelines. Despite these limitations, we identified tentative patterns between design features and intervention outcomes, discussed below in relation to existing literature.

Theoretical Foundations

Regarding theoretical foundations, we found that this feature did not appear to differentiate between interventions that achieved their intended outcomes and those that did not: both types of interventions commonly drew upon social cognitive theory. Potential explanations include inappropriate selection or poor application of the theory for a given intervention context [84]. Notably, 5 interventions in our review lacked a clearly defined theoretical foundation, with 3 of these reporting achievement of their intended outcomes. This pattern is noteworthy given available literature suggesting that theory-driven interventions are more likely to be successful [84,85]. Possible explanations include that theoretical frameworks were applied but not explicitly reported, or that intervention developers focused on implementing specific BCTs rather than grounding their approach in a central theory.

Timing and Duration

Earlier intervention initiation and longer duration were more commonly observed in interventions that achieved their stated objectives. This aligns with the understanding that weight gain accelerates in the second half of pregnancy, necessitating early intervention [66,86]. None of the interventions in our review commenced in the third trimester, which aligns with this rationale.

Frequency

No clear patterns emerged between contact frequency and intervention outcomes. Interventions showed substantial heterogeneity in contact points with providers and participant control over interaction with intervention components. Reasons for these differences likely stem from theory-based choices, for example, offering more frequent but less personalized elements versus less frequent but more tailored ones.

Delivery Mode

Digital-mixed interventions were more often associated with success than digital-only formats. As such, in-person components, especially those involving health care provider endorsement, may enhance credibility and behavior change, which is in line with existing research. However, as Rhodes et al [68] observed, methodological heterogeneity between interventions may have limited the ability to detect consistent patterns of association with intervention outcomes. Therefore, studies in which a comparison is made between delivering the same intervention through digital-mixed and digital-only means, such as that by Redman et al [31] that offers an interesting venue for future research.

BCTs

BCTs related to the categories goals and planning, and feedback and monitoring were frequently associated with successful interventions. While previous research by Webb et al [87] and Rhodes et al [68] suggested that a higher number of BCTs may enhance effectiveness, our review did not find a consistent relationship between the number of BCTs and intervention success. This discrepancy may reflect differences in scope: whereas both Webb et al and Rhodes et al examined digital-only interventions, our review included both digital-only and digital-mixed delivery. Despite this, the average number of BCTs per intervention in our review (average 9) matched that reported by Rhodes et al [68], suggesting some consistency in usage patterns.

More importantly, our analysis highlighted that specific BCTs, rather than the total number, were more commonly associated with successful outcomes. These included: goal setting (outcome), discrepancy between current behavior and goal, self-monitoring of behavior, social support (unspecified), and a credible source. These findings are supported by broader literature on diet and PA interventions, in which interventions combining self-monitoring with at least one other technique from control theory [88] – in this case, goal setting (outcome) – were more commonly associated with achieving intended outcomes [89].

When examining delivery modes, we observed that digital-mixed interventions used more BCTs than digital-only ones. This may suggest that digital-only interventions can be effective with fewer, well-targeted BCTs – an area that warrants further research, particularly given the challenges of adapting certain BCTs to digital formats [90].

In Summary

Summarizing, our review identified several implementation and design features that may contribute to digital lifestyle interventions for GWG in achieving their intended goals. While theoretical foundation and frequency of intervention delivery were not commonly associated with success, interventions that started earlier in pregnancy, lasted longer, and combined digital with nondigital components were more likely to achieve beneficial outcomes. In addition, the BCTs’ goal setting (outcome), discrepancy between current behavior and goal, self-monitoring of behavior, social support (unspecified), and credible source were commonly present across interventions accomplishing their intended purpose.

Limitations and Strengths

This review has several limitations. First, we excluded articles involving high-risk pregnancies. Given the greater health stakes in such populations, it is possible that digital lifestyle interventions may have different levels of impact among at-risk women. Future research should explore whether outcomes differ between normal- versus high-risk pregnancies.

A second limitation is the inability to control for variations in usual care across different localities, which can vary substantially between health care settings. This may influence women’s opportunities, motivations, and capabilities to engage in healthy lifestyle behaviors. Consequently, the functional gap that a digital lifestyle intervention needs to address may differ depending on this local health care context.

Third, most included studies were conducted in high-income countries, mainly the United States. This limits the generalizability of our findings to low- and middle-income countries, where digital interventions may be especially valuable due to limited access to maternity care.

In addition, our searches included only publications in the English language and those published until March 25, 2024. This may have excluded relevant studies in other languages or published more recently.

Despite these limitations, our review has several strengths. We included both digital-only and digital-mixed interventions, enabling comparative insights. Our structured and replicable literature search strategy ensured comprehensive coverage of the literature. By focusing on healthy (normal-risk) pregnant populations, we were able to draw clearer comparisons across studies.

Finally, we conducted an extensive analysis of key intervention and design features used across the included interventions, including BCTs. This addressed a gap in the literature, as only a few previous reviews have provided a detailed breakdown of which specific BCTs are commonly used in digital lifestyle interventions with both digital-mixed and digital-only delivery modes for GWG management in pregnant women. Our analysis, therefore, provides a foundation for understanding the current landscape of BCT implementation in this field and may inform future research directions and intervention development considerations. Therefore, we believe that despite the limitations of our review, our conclusions about the key design and implementation features contributing to interventions achieving their intended goals remain valid.

Future Research

Based on our findings, we propose 5 key priorities for future research, which we outline below.

Healthy Lifestyles Versus GWG

First, digital lifestyle interventions reported modest differences in GWG between interventions that met their intended goals and those that did not (average reduction of 1.4, SD 0.71 kg). This is in line with Dodd et al [91], whose meta-analysis of 78 studies on nondigital lifestyle interventions found an average decrease in GWG of 1.1 (95% CI –1.46 to –0.74) kg in pregnant intervention participants compared with controls. In combination with previous research in nonpregnant populations suggesting that more weight loss is needed before health benefits start showing [92,93] and our findings that digital lifestyle interventions can positively affect lifestyles even without successful outcomes for healthy GWG [35,61,67], future studies should explore whether these behavioral changes, rather than GWG alone, are the primary drivers of improved pregnancy outcomes.

Engagement and Standardization

Second, more work on methodological aspects is required, such as differentiating between digital and behavioral engagement in longitudinal interventions. Digital engagement refers to users’ interaction with intervention components, while behavioral engagement describes the adoption of target health behaviors. Understanding their temporal relationship is crucial: interventions typically require initial digital engagement until users establish habits, after which behavioral engagement becomes more meaningful for assessing success. Measuring one type when intending to measure the other could lead to misleading conclusions about what contributes to achieving intended outcomes. It is important to note that long-term engagement—whether digital or behavioral—may not always be necessary for intervention success [90]. Rather than pursuing maximum engagement, future research should focus on identifying “effective engagement” to achieve healthy GWG. In addition, broader use of the BCT taxonomy [44] and clearer reporting of implementation strategies would enhance comparability and replication.

Intervention Development

Third, concerning intervention development, a better understanding of individual BCTs and their implementation is needed. Our findings reveal an overlap between the most prominent BCTs in both effective and ineffective interventions, suggesting that BCT implementation methods may influence outcomes differently than the techniques themselves. For instance, social support can be delivered through forums, group settings, or peer-to-peer pairs, each with potentially different characteristics. Future research could explore BCT implementation approaches and their underlying mechanisms of action.

Another consideration involves recent technological trends in intervention development. As most pregnancy apps are developed by private companies [94], their features reflect current technological capabilities, including comprehensive data collection, wearable integration, and algorithm-driven coaching [95,96]. While these advances may enhance accessibility and personalization, they introduce important ethical considerations [97,98]. Particularly, personalization raises concerns about user autonomy [99] and trustworthiness [100], especially as intervention rules change through usage, potentially making one-time informed consent insufficient. In addition, pregnant women may place varying levels of trust in these systems despite potentially limited understanding of their capabilities [100]. Users may end up over- or undertrusting the system, both of which come at a cost [101]. Future research could explore how ethical considerations might be integrated into intervention development from the start, with attention to meaningful and ethical personalization.

Collaboration Between Academia and Industry

Fourth, we advocate enhanced collaboration between academia and industry in the research and development of digital lifestyle interventions for pregnant women. While our review identified many pilot RCTs, we observed a notable scarcity of follow-up studies scaling these initial findings. This gap may stem from the inherent challenges of transitioning from pilot studies to full-scale RCTs with technologically mature interventions within academic settings, where research typically progresses at a more measured pace and fewer resources for technological development are available. In contrast, industry partners operate on accelerated development cycles with direct access to emerging technologies. This difference in throughput time becomes particularly relevant given that while most published research originates from academic institutions, the majority of pregnancy-related apps in use are developed by private technology companies [94]. Strategic partnerships between these sectors could combine academia’s methodological rigor with industry’s technological agility, accelerating the translation of evidence-based interventions into widely accessible tools while maintaining scientific integrity, to together improve the health behavior of pregnant women and their families.

Societal Aspects

Finally, while there is a widespread assumption among researchers and policymakers that digital interventions for pregnancy could significantly benefit health care desert communities, this hypothesis remains largely unexplored in empirical research [94,102]. This gap is particularly noteworthy because, due to their unique circumstances and limited alternative resources, individuals in health care deserts may actually demonstrate higher success rates with digital interventions in terms of increased capability and opportunity to engage with healthy behaviors. In our scoping review, only Iyawa et al [75] addressed this crucial aspect. Furthermore, the health care desert context emphasizes the importance of conducting rigorous cost-effectiveness analyses, comparing different intervention approaches. Such economic evaluations are essential for making informed decisions about resource allocation, particularly in areas where health care resources are already scarce. This economic perspective becomes especially relevant when considering the potential scalability and sustainability of digital lifestyle interventions in underserved communities [103].

Conclusion

This scoping review examined the landscape of digital lifestyle interventions aimed at supporting healthy (GWG), with a focus on 6 key design and implementation features: theoretical foundation, intervention timing, duration, frequency, delivery modes, and BCTs. Our findings confirm that digital interventions hold promise for promoting healthy GWG. While theoretical underpinnings and frequency of delivery did not consistently predict success, interventions that began earlier in pregnancy and lasted longer were more likely to achieve beneficial outcomes. Digital-mixed delivery modes—those combining digital tools with in-person contact—appeared more effective than digital-only formats. Importantly, 5 BCTs emerged as more commonly used in successful interventions: goal setting (outcome), discrepancy between current behavior and goal, self-monitoring of behavior, social support (unspecified), and credible source.

These findings provide a foundation for designing more effective, evidence-based digital interventions to support maternal health. Future research should continue to refine these components, explore their implementation in diverse populations, and address gaps in reporting and standardization.