A six-month habit programme left more people in metabolic remission two years later

Lifestyle intervention trials for cardiometabolic risk have a consistent shape: benefits during the programme, attenuation afterwards, and a follow-up paper reporting that most of the effect is gone by the time anyone looks again. The problem is not that lifestyle change fails to work. It is that it fails to persist once the support ends.

The ELM trial, published in the January issue of JAMA Internal Medicine, was designed around that problem specifically: the treatment lasted six months, and the endpoint was measured 18 months after it stopped [s1].

The design

Six hundred and eighteen participants across five US sites were randomised in a single-blind trial running from July 2019 to January 2022, with 24 months of follow-up [s1]. Mean age was 55.5 years, 74.7% of participants were female, and 83% had obesity [s1].

All participants — both arms — received education and activity monitors [s1]. The intervention group additionally attended 19 small group meetings aimed at establishing four daily habits: vegetable consumption, brisk walking, sensory awareness, and emotion regulation [s1]. The comparator group received 24 monthly evidence-based mailings [s1].

The comparator matters here. This was not intervention versus nothing. Both arms got monitoring technology and educational content; the contrast being tested is the group-based habit formation component on top of that.

The endpoint was sustained remission of metabolic syndrome at 24 months — that is, no longer meeting the three-or-more-risk-factor definition, 18 months after the programme ended [s1].

The result

At 24 months, sustained remission occurred in 27.8% of intervention participants versus 21.2% of comparators, an adjusted odds ratio of 1.46 (95% CI 1.01–2.14; P<.05) [s1].

The lower bound of that confidence interval is 1.01 [s1]. That is about as close to the null as a statistically significant result gets. The trial is positive on its prespecified terms, and the result is fragile in the ordinary sense that a handful of differently classified participants would move it.

It is worth stating the effect in the plainest available terms: 6.6 percentage points more of the intervention group met the remission definition [s1]. Roughly seven in ten participants in the intervention arm did not.

Why it is nonetheless notable

Metabolic syndrome prevalence has reached 43% among US adults [s1]. At that scale, small proportional differences represent large absolute numbers, which is the standard argument for population-level interventions and is legitimate.

More specifically, the trial's contribution is the timing of measurement. Most lifestyle trials report their primary endpoint at or near the end of the intervention period. This one asked whether anything remained a year and a half after the group meetings stopped, and found a difference that was still detectable [s1].

The authors' conclusion is framed around that: sustained remission after treatment may be possible by promoting simple habits through a behaviour-based lifestyle programme focused on immediate benefits [s1]. The emphasis on immediate benefits is a design choice worth noting — habit formation literature generally holds that behaviours reinforced by immediate experience persist better than behaviours motivated by distant health outcomes.

What the trial cannot tell you

It cannot tell you which of the four habits did the work, or whether any of them did. Vegetable consumption, brisk walking, sensory awareness, and emotion regulation were delivered together as a package across 19 meetings [s1]; the trial has no arms that separate them.

It cannot tell you whether the group format, the 19 contacts, or the content produced the difference. A programme with that much structured contact differs from monthly mailings in attention, accountability, and social contact as well as in curriculum.

Single blinding means participants knew their assignment [s1]. Metabolic syndrome remission is defined by measured biomarkers rather than self-report, which limits how much that can distort the primary endpoint, but it does affect the behaviours upstream of those biomarkers.

The trial ran through the COVID-19 pandemic, having begun in July 2019 [s1] — a period that disrupted both group meeting formats and participants' baseline behaviour in ways that are hard to model.

And a 55.5-year-old, mostly female sample with 83% obesity prevalence [s1] is a specific population. Generalisation to younger adults, to men, or to people with metabolic syndrome without obesity is not supported by these data.

The comparison nobody ran

The obvious contemporary question this trial does not address is how a 6.6-percentage-point difference in sustained remission compares with what pharmacotherapy achieves in the same population — and, more usefully, what the two do in combination. The trial predates none of that debate but was not designed to enter it.

What to watch

Whether the difference persists at longer follow-up; whether any component analysis is published; and whether the habit-formation framework is tested against, or alongside, current pharmacological approaches to cardiometabolic risk.

This article describes a published randomised trial and is informational only. It is not medical advice and does not recommend any programme, diet, or treatment.

Sources

• [s1] Powell LH, Berkley-Patton J, Drees BM, et al., and the ELM Trial Research Group, Lifestyle Intervention for Sustained Remission of Metabolic Syndrome: A Randomized Clinical Trial, JAMA Internal Medicine, 2026;186(1):67-77.