Effectiveness of an automated text message intervention for weight management in postpartum women with overweight or obesity (Supporting MumS (SMS)): a UK wide, multicentre, two arm, parallel group, randomised controlled trial.

This UK-wide randomized controlled trial found that while a 12-month automated text message intervention did not significantly reduce weight in postpartum women with overweight or obesity compared to an active control group, it successfully improved dietary habits, physical activity levels, and infant feeding practices.

The Gist

Here is the "Supporting MumS (SMS)" study explained in simple, everyday language, using analogies to make the concepts stick.

The Big Picture: A Text Message "Life Coach" for New Moms

Imagine you are a new mom. You've just had a baby, you're tired, you're busy, and you might be carrying a few extra pounds from pregnancy. You know you want to get back to a healthy weight, but you don't have time to go to the gym, you can't afford a personal trainer, and you're too exhausted to cook complex meals.

The Supporting MumS (SMS) trial asked a simple question: Could a friendly, automated text message service act like a 24/7 life coach to help moms lose weight?

The researchers didn't just send generic "eat your veggies" texts. They built a smart system that:

• Sent encouraging messages about food and exercise.
• Asked moms to text back their weight (like a digital diary).
• Let moms text back keywords like "tired" or "craving" to get instant, helpful advice.
• Ran for a full 12 months.

The Experiment: The "Weight Loss" Team vs. The "Baby Book" Team

To see if this actually worked, they recruited 892 moms from all over the UK (from Belfast to London) who had a BMI of 25 or higher (overweight or obese). They split them into two teams, like flipping a coin:

1. The Intervention Team (The "Weight Loss" Team): Got the 12 months of smart, weight-loss-focused text messages.
2. The Control Team (The "Baby Book" Team): Got 12 months of text messages too, but these were about baby development (e.g., "Your baby is 6 months old! Here's what to expect next"). This was the "active control" to make sure the moms felt they were getting something valuable, just not the weight-loss stuff.

The Results: Did the Texts Work?

Here is the twist, which is the most important part of the story:

1. The Scale Didn't Move (Overall)
When they weighed everyone at the end of the year, the "Weight Loss" team did not lose significantly more weight than the "Baby Book" team.

• The Analogy: Imagine two groups of people trying to climb a hill. One group got a map and a compass (the texts), and the other group got a map of the local park (baby info). At the top of the hill, both groups were at roughly the same altitude. The text messages alone weren't enough to push the average mom down the scale significantly.

2. But the Habits Did Change
Even though the scale didn't budge much, the "Weight Loss" team started doing better things:

• They moved more: They were more active.
• They ate better: They chose healthier foods (more fiber, less junk).
• They fed their babies better: They introduced healthier foods to their infants.
• The Analogy: The "Weight Loss" team didn't climb the mountain faster, but they started walking with better shoes, drinking more water, and eating better snacks along the way. They were building a healthier lifestyle, even if the final number on the scale didn't change much yet.

3. The "Super-Engagers" Won
This is the most exciting finding. The researchers looked at who actually replied to the texts.

• The "Low Engagers": Moms who barely replied to the texts. They didn't lose weight.
• The "High Engagers": Moms who replied to the texts often, used the "craving" keywords, and treated the service like a real conversation.
• The Result: The "High Engagers" did lose weight (about 2kg or 4.4 lbs more than the control group).
• The Analogy: Think of the text service like a gym membership. If you just buy the membership and never go, you don't get fit. But if you go every day and actually lift the weights (reply to the texts), you get the results. The tool works, but you have to use it.

Why Was This Study Special?

Most studies on new moms fail because:

• They are too rigid: "Come to our clinic every Tuesday at 9 AM." (New moms can't do that).
• They are too short: "Here is a 6-week plan." (Weight loss takes years).
• They aren't diverse: They only study white, wealthy women.

This study was different because:

• It was flexible: Moms could get visits at their homes, in cafes, or online. They could do it when the baby was sleeping.
• It was inclusive: They recruited moms from all ethnic backgrounds and all income levels.
• It was long: They stuck with it for a whole year.

The Bottom Line

The Verdict: A text message app alone isn't a magic wand that melts fat away for everyone. However, it is a powerful tool that helps moms build better habits (eating well, moving more) and helps those who are willing to engage with it actually lose weight.

The Takeaway for You:
If you are a new mom trying to get healthy, a text message service might be a great starting point to build good habits. But remember: the app is the coach, you are the athlete. The more you interact with the coach, the better your results will be. And even if the scale doesn't move immediately, building those healthy habits is a huge win for your long-term health and your baby's future.

Technical Summary

1. Problem Statement

Postpartum weight retention is a significant public health concern, contributing to long-term risks of obesity, cardiovascular disease, and diabetes. Despite clinical guidelines recommending weight management support within two years of childbirth, there is a scarcity of effective, scalable, and acceptable interventions for this population. Existing studies often suffer from:

• Poor recruitment and retention: Due to the demanding nature of the postpartum period (sleep deprivation, childcare, time constraints).
• Lack of diversity: Samples often lack ethnic and socioeconomic representation.
• Inflexible delivery: Many interventions rely on in-person attendance, which is a barrier for new mothers.
• Limited evidence on technology: While mobile health (mHealth) shows promise, few large-scale Randomized Controlled Trials (RCTs) have tested fully automated text message interventions as the sole delivery mode for postpartum weight loss.

2. Methodology

Study Design:
A UK-wide, multicentre, two-arm, parallel-group, superiority Randomized Controlled Trial (RCT).

Participants:

• Sample Size: 892 women recruited between May 2022 and May 2023.
• Inclusion Criteria: Women aged ≥18 years, 6 weeks to 24 months postpartum, with a Body Mass Index (BMI) ≥25 kg/m², owning a mobile phone.
• Exclusion Criteria: Insufficient English proficiency, history of eating disorders, current weight loss surgery/dietary care, or participation in other weight management studies.
• Demographics: The sample was diverse, with 33.6% from minoritised ethnic groups and over half residing in the two most deprived Index of Multiple Deprivation (IMD) quintiles.

Interventions:

• Intervention Group (n=445): Received a 12-month fully automated text message service ("Supporting MumS").
  • Content: Based on the Health Action Process Approach (HAPA) and maintenance theories. Messages targeted diet, physical activity, and weight management using Behavior Change Techniques (BCTs).
  • Features: Included a core library of 353 messages. Crucially, it featured bidirectional (two-way) messaging: weekly prompts for self-weighing and "trigger word" responses (e.g., "slip-up," "crave") to provide automated, tailored feedback. Frequency decreased from ~14 messages/week (months 1-2) to ~4 messages/week (months 7-12).
• Active Control Group (n=447): Received 12 months of unidirectional text messages regarding child health and development, tailored to the infant's age. This controlled for the "attention" effect of receiving regular texts without the active weight management ingredients.

Data Collection & Outcomes:

• Primary Outcome: Weight (kg) at 12 months.
• Secondary Outcomes: BMI, waist circumference, weight change (>5kg gain/loss), dietary intake (Fat and Fibre Barometer), physical activity (IPAQ-SF), and infant feeding practices.
• Analysis: Intention-to-treat (ITT) using Analysis of Covariance (ANCOVA), adjusting for baseline values, site, ethnicity, and recruitment pathway. Sensitivity analyses included multiple imputation and delta-adjusted methods for missing data.

3. Key Contributions

• Scale and Diversity: This is one of the largest and most diverse RCTs in postpartum weight management, successfully recruiting and retaining a representative sample across four UK nations, including significant numbers of women from minoritised ethnic groups and deprived areas.
• Fully Automated Delivery: It is the first adequately powered RCT to test a fully automated text message intervention as the sole modality, removing the resource intensity of human coaching or in-person sessions.
• Engagement Analysis: The study introduced a pre-specified engagement analysis using hierarchical clustering of bidirectional message responses to differentiate between "high" and "low" engagers, providing insight into the "dose-response" relationship of the intervention.
• Acceptability: The study demonstrated that flexible, home-based data collection and text-based delivery are highly acceptable to postpartum women, achieving high retention rates (84.4% at 12 months excluding pregnancies).

4. Results

Primary Outcome (Weight):

• Overall Effect: There was no statistically significant difference in mean weight change at 12 months between the intervention and control groups.
  • Adjusted mean difference: -0.1 kg (95% CI: -1.0 to 0.8; P=0.84).
• Secondary Anthropometric Measures: No significant differences were found for BMI, waist circumference, or the proportions of women gaining/losing >5kg.

Secondary Outcomes (Behavioral):

• Diet: The intervention group showed a statistically significant improvement in Fat and Fibre Barometer (FFB) scores compared to control at 12 months (Adjusted mean difference: 0.09; P<0.001).
• Physical Activity: Significant increase in physical activity scores (IPAQ-SF) in the intervention group (Adjusted mean difference: +405.3 MET minutes/week; P=0.003). There was a notable shift from low to moderate/high activity levels (OR 1.41; P=0.02).
• Infant Feeding: Significant increases in the weekly portions of eggs, beans/lentils, and nuts given to infants in the intervention group.

Engagement Analysis (Crucial Finding):

• Participants were clustered into "high" and "low" engagers based on their response rates to bidirectional messages.
• High Engagers: Women who responded to ~20% of messages lost significantly more weight than the control group.
  • Adjusted mean weight difference: -1.9 kg (95% CI: -3.2 to -0.6; P=0.005).
• Low Engagers: Showed no significant weight loss compared to control, though they did show improvements in diet scores.

Subgroup Analysis:

• No differential effects were observed across ethnicity, IMD, or site.
• A non-significant trend suggested better weight loss in women who were >52 weeks postpartum at baseline, though this did not hold in the fully adjusted model.

5. Significance and Implications

• Clinical Reality: The trial confirms that while automated text messages can successfully improve health behaviors (diet and activity), they are insufficient on their own to drive significant weight loss for the entire postpartum population. The "one-size-fits-all" automated approach failed to overcome the complex barriers of the postpartum period for the average participant.
• The Role of Engagement: The significant weight loss observed in the "high engagement" subgroup suggests that the intervention is effective if the user actively interacts with the system. This highlights that passive receipt of information is less effective than active participation in behavior change.
• Future Directions:
  • Tailoring: Future interventions may need dynamic tailoring (e.g., using AI/ML) to identify and re-engage users who are disengaging, rather than relying on static automated flows.
  • Hybrid Models: Text messaging may serve best as a scalable "wraparound" support for more intensive interventions or for those who cannot access in-person care, rather than a standalone cure.
  • Policy: The high acceptability and retention rates demonstrate that digital health interventions are a viable, low-burden option for public health delivery, provided they are designed with the specific constraints of new mothers in mind (flexibility, home visits, etc.).

Conclusion:
The Supporting MumS (SMS) trial provides robust evidence that while fully automated text messaging improves diet and physical activity behaviors in postpartum women, it does not result in significant weight loss for the general population. However, for the subset of women who actively engage with the bidirectional features of the system, meaningful weight loss is achievable. This underscores the critical importance of engagement strategies in digital health interventions.