An adaptive digital lifestyle intervention modestly reduced both the rate of gestational weight gain and total weight gain in patients with overweight or obesity, a cluster-randomized trial found.

Reported in JAMA Network Open, the LEAP study examined whether adding an evidence-based, adaptive mobile health lifestyle intervention to standard care reduces excessive gestational weight gain compared with standard care alone.

More than 60% of US pregnant individuals enter pregnancy with overweight or obesity, and of these, more than half exceed recommended national weight gain guidelines.

These guidelines, said lead author Monique M. Heddersen, PhD, research scientist at the Kaiser Permanente Northern California, Division of Research, in Pleasanton, California, recommend less weight gain for women with higher prepregnancy BMI.

The study focused on patients entering pregnancy with overweight or obesity because they are at high risk of having perinatal complications and excess gestational weight gain exacerbates that risk, Heddersen told Medscape Medical News. “We wanted to see whether we could use mobile health technology to create a more practical, scalable intervention that could be integrated into clinical care.”

Study Details

Patients were randomly assigned at the clinician level. LEAP enrolled pregnant patients with a singleton pregnancy and a prepregnancy BMI of 25 to < 40 who were receiving care within an integrated health system. Data were analyzed from May 2024 to February 2026.

Intervention clinicians received motivational interviewing guidance for discussing weight gain, whereas their patients received personalized automated feedback on weight gain and physical activity via a smartphone application, a wireless weight scale, and an activity tracker.

AI agents were used to monitor the data from the wireless scales and determine who needed personalized text messages from a human lifestyle coach. For those experiencing accelerated weight gain, 13 weekly educational topics were explored via the app, stepwise chats, or phone support from a coach.

A total of 58 participating clinicians enrolled 1265 patients (mean age, 33.4 years; mean prepregnancy BMI, 29.8) — 677 patients in the intervention group and 588 in the standard care arm. Only about 53% of eligible women, however, chose to download the app.

The weekly rate of weight gain was significantly lower in the intervention group vs the standard care group: mean rate, 0.25 kg vs 0.28 kg (mean between-group difference, -0.03 kg; 95% CI, -0.05 to -0.01).

In addition, fewer intervention patients exceeded US Institute of Medicine (IOM) guidelines for weekly gain rate than standard care patients: 51.9% of 351 vs 60.2% of 354 (relative risk [RR], 0.86; 95% CI, 78%-95%).

Total gain was significantly lower in the intervention group than in the standard care group: 9.7 kg vs 10.6 kg (mean between-group difference, -0.87 kg; 95% CI, -1.40 to -0.34 kg).

Fewer intervention patients exceeded IOM guidelines for total weight gain than their standard care counterparts: 299 (44.1%) vs 301 (51.2%; RR, 0.87; 95% CI, 0.77-0.98).

The effectiveness of the intervention did not vary by sociodemographic characteristics but appeared more effective for those entering pregnancy with overweight (BMI, 25-29.9) than for those with obesity (BMI ≥ 30).

“When we focused on participants who actually engaged with the technology, we found that the more they engaged, especially by weighing themselves consistently, the less weight they gained,” Heddersen said.

Suzanne Phelan, PhD, professor of kinesiology and public health at California Polytechnic State University in San Luis Obispo, California, called the findings very promising. “The investigators found modest but meaningful effects in a pragmatic clinical setting. The effects are smaller than in efficacy trials in pregnancy but notable given real-world implementation,” added Phelan, who was not involved in the trial.

Steps to increase the observed low uptake of the app might include game-playing elements and other incentives, Phelan said. “Impact could potentially be increased by refining stepped care thresholds to trigger coaching earlier or more frequently.”

The results “were modest but meaningful and notably consistent with decades of health behavioral intervention research,” wrote Elizabeth R. Stevens, PhD, MPH, and Devin M. Mann, MD, MS, of NYU Grossman School of Medicine and NYU Langone Health in New York City, in a related commentary.

Behavioral counseling is one of the most effective tools available for managing weight during pregnancy, they noted. “Yet despite strong evidence and national recommendations, delivering this kind of counseling at scale remains difficult in routine care. The time, training, and workforce required to provide truly personalized support place a growing burden on already stretched clinical systems.”

Hence, while mobile health offers a pragmatic vision for scalable support, “LEAP highlights the bottleneck that keeps evidence-based counseling from reaching everyone who could benefit, including the reliance on human coaches for personalized care,” they wrote.

A striking finding from the study was the gap between patient eligibility and engagement, they noted. Only about half of patients downloaded the app, and fewer than 20% sustained high engagement.

One remedy is to deliver the personalized component via generative AI, which analyzes data to produce new and varied content, rather than via a human coach. “AI-generated messaging may reduce some of the social and logistical barriers that make it difficult for patients to engage with traditional clinical interactions,” they wrote.

Looking ahead, Heddersen’s group plans to explore ways to increase engagement and examine longer‑term outcomes such as postpartum weight retention and child growth. “If we develop successful strategies to increase patient engagement, we could further increase the effectiveness of the intervention,” she said.

LEAP was supported by grants from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to Heddersen and co-author Susan D. Brown. Brown reported receiving grants from the National Institutes of Health during the study. Co-author Charles P. Quesenberry reported receiving grants from NIDDK during the study. Phelan reported having no conflicts of interest. The editorial commentators reported receiving no funding and having no conflicts of interest.