A Systematic Review and Bayesian Network Meta-Analysis Comparing In-Person, Remote, and Blended Interventions in Physical Activity, Diet, Education, and Behavioral Modification on Gestational Weight Gain among Overweight or Obese Pregnant Individuals

Abstract

Background: Despite the well-documented adverse outcomes associated with obesity during pregnancy, this condition remains a promising modifiable risk factor.

Objectives: The aim of this study was to ascertain the most effective treatment modalities for gestational weight gain (GWG) in pregnant women classified as overweight or obese.

Methods: A systematic search was conducted across 4 electronic databases: Embase, EBSCOhost, PubMed, and Web of Science. To assess the quality of evidence, the Confidence In Network Meta-Analysis (CINeMA) approach, grounded in the Grading of Recommendations Assessment, Development, and Evaluation framework, was employed. A Bayesian network meta-analysis was conducted to synthesize the comparative effectiveness of treatment modalities based on GWG outcomes.

Results: The analysis incorporated 60 randomized controlled trials, encompassing 16,615 participants. Modes of intervention administration were classified as remote (R: eHealth [e] and mHealth [m]), in-person (I), and a combination of both (I+R). The interventions comprised 5 categories: education (E), physical activity (PA), dietary (D), behavior modification (B), and combinations thereof. The quality of the evidence, as evaluated by CINeMA, ranged from very low to high. Compared to the control group, the I-D intervention (mean difference [MD]: -1.27; 95% confidence interval [CI]: -2.23, -0.32), I-PADB (MD: -0.60, 95% CI: -1.19, -0.00), and I-B (MD: -0.34, 95% CI: -0.57, -0.10) interventions showed significant efficacy in reducing GWG.

Conclusions: Preliminary findings suggest that the I-D intervention is the most efficacious in managing GWG among pregnant women who are overweight or obese, followed by I-PADB and I-B+R-B(m) treatments. These conclusions are drawn from evidence of limited quality and directness, including insufficient data on PA components used in the interventions. Owing to the absence of robust, direct evidence delineating significant differences among various GWG management strategies, it is tentatively proposed that the I-D intervention is likely the most effective approach. This study was registered with PROSPERO as CRD42023473627.

Keywords: diet; gestational weight gain; lifestyle; physical activity; pregnancy.

FIGURE 1

The PRISMA flow diagram delineates the meticulous procedure employed for the selection of articles included in this review.

FIGURE 2

A network diagram, elucidating the direct evidence derived from all intervention arms, focuses on gestational weight gain. In this diagram, the magnitude of each node correlates directly with the number of participants who underwent the respective treatment. Furthermore, the thickness of the interconnecting lines is indicative of the volume of randomized controlled trials that conducted direct comparisons between the treatment pairs. The treatments are abbreviated as follows: R for remote, I for in-person, e for electronic health, m for mobile health, PA for physical activity, D for dietary, E for education, and B for behavior.

FIGURE 3

A sophisticated network forest plot delineates the comparative effectiveness of diverse treatment modalities in moderating gestational weight gain outcomes. This plot encapsulates an array of treatment strategies, categorized as follows: A, I-B; B, I-B+R-B(e); C, I-B+RB(m); D, I-B+RDB(e); E, I-D; F, I-DB; G, I-DB+R-B(e); H, I-E; I, I-E+R-PA(m); J, I-PA; K, I-PAB; L, I-PAD; M, I-PADB; N, I-PADB+R-B(e); O, I-PADB+R-PAB(e); P, I-PAE; Q, Placebo; R, R-B(e); S, R-B(m); T, R-B(m+e); U, R-E(e); V, R-PA(m+e); W, R-PAB(e); X, R-PADB(e); Y, R-PADB(m); Z, R-PADE(e). The treatments are abbreviated as follows: R for remote, I for in-person, e for electronic health, m for mobile health, PA for physical activity, D for dietary, E for education, and B for behavior.

FIGURE 4

The surface under cumulative ranking (SUCRA) plot constructed based on altered gestational weight gain results. The treatments are abbreviated as follows: R for remote, I for in-person, e for electronic health, m for mobile health, PA for physical activity, D for dietary, E for education, and B for behavior.

FIGURE 5

A funnel plot constructed to assess the altered acclimate in gestational weight gain. This plot includes a comprehensive list of treatment combinations, represented by alphabetic designations (A–Z). A, I-B; B, I-B+R-B(e); C, I-B+RB(m); D, I-B+RDB(e); E, I-D; F, I-DB; G, I-DB+R-B(e); H, I-E; I, I-E+R-PA(m); J, I-PA; K, I-PAB; L, I-PAD; M, I-PADB; N, I-PADB+R-B(e); O, I-PADB+R-PAB(e); P, I-PAE; Q, Placebo; R, R-B(e); S, R-B(m); T, R-B(m+e); U, R-E(e); V, R-PA(m+e); W, R-PAB(e); X, R-PADB(e); Y, R-PADB(m); Z, R-PADE(e). The treatments are abbreviated as follows: R for remote, I for in-person, e for electronic health, m for mobile health, PA for physical activity, D for dietary, E for education, and B for behavior.

FIGURE 6

A loop inconsistency plot based on the altered gestational weight gain outcomes. This graph shows the full spectrum of treatment combinations using the standard abbreviations for treatments as outlined: A, I-B; B, I-B+R-B(e); C, I-B + RB(m); D, I-B+RDB(e); E, I-D; F, I-DB; G, I-DB+R-B (e); H, I-E; I, I-E+R-PA(m); J, I-PA; K, I-PAB; L, I-PAD; M, I-PADB; N, I-PADB+R-B(e); O, I-PADB+R-PAB(e); P, I-PAE; Q, Placebo; R, R-B(e); S, R-B(m); T, R-B(m+e); U, R-E(e); V, R-PA(m+e); W, R-PAB(e); X, R-PADB(e); Y, R-PADB(m); Z, R-PADE(e). The treatments are abbreviated as follows: R for remote, I for in-person, e for electronic health, m for mobile health, PA for physical activity, D for dietary, E for education, and B for behavior. CI, confidence interval; IF, inconsistency factor.