What Weight-Loss Maintainers Actually Do

Everybody in the registry had already succeeded before they joined#

The National Weight Control Registry is a roster of people who lost at least 13.6 kg and kept it off for at least a year, recruited largely by advertisement and followed with questionnaires. It is the richest description available of what long-term maintenance looks like from the inside, and it is worth reading carefully. It is also, by construction, incapable of answering the question it is most often quoted to answer. You cannot estimate a success rate from a sample selected on success. There is no denominator anywhere in it.

Hold both of those at once and the registry becomes useful in a specific way: as a description of a destination rather than a map of a route or a bet on the odds. What the destination looks like is fairly consistent — a lot of structured exercise, an eating pattern that does not change between Tuesday and Saturday, frequent self-weighing, and a very small tolerance for drift. What the registry cannot tell you is whether those behaviors caused the maintenance, whether they would work for someone who has not already done it, or how many people tried the same things and are not in the database. The practical playbook for the maintenance phase itself is in how to keep the weight off; this page is about what the registry actually measured and what it is entitled to claim.

The entry ticket, and what it selects for#

The founding description covered 784 members — 629 women and 155 men — who had lost an average of 30 kg and held it, with entry requiring a maintained loss of at least 13.6 kg for five years in that first cohort¹. Reported intake averaged about 5,778 ± 2,200 kJ/day with 24 ± 9% of energy from fat, and reported physical activity around 11,830 kJ per week. Slightly over half had lost the weight through a formal program; the rest had done it alone. Nearly all used both diet and exercise. And 77% reported that a triggering event had preceded the successful attempt.

That last cluster of numbers is the most instructive thing in the whole registry and the least quoted. On how they lost it, members diverge completely — programs and self-directed attempts split roughly down the middle, and the diets varied. On how they hold it, they converge. The registry is therefore weak evidence about which diet to pick and much stronger evidence that maintenance is a distinct set of behaviors rather than a continuation of whichever diet happened to work.

One more caveat about intake: those are self-reported figures, and self-reported intake in this population runs low for the same reasons it runs low in every population. Take the composition of the reported diet more seriously than its absolute size.

The 'hour a day' claim, put on an accelerometer#

Registry members report roughly an hour of daily physical activity, and that number has been repeated so often it has become a rule. Somebody eventually measured it. Twenty-six registry members wore triaxial accelerometers for seven consecutive days, alongside 30 never-obese normal-weight controls and 34 overweight controls².

Objectively measured, per day	Registry maintainers	Normal-weight controls	Overweight controls
Moderate-to-vigorous activity in bouts of ≥10 min	41.5 ± 35.1 min	25.8 ± 23.4 min	19.2 ± 18.6 min
Moderate-to-vigorous activity in short bursts (1–9 min)	28.1 ± 10.8 min	32.6 ± 15.4 min	32.3 ± 12.4 min
Bouts per day	1.4 ± 1.0	0.9 ± 0.6	0.8 ± 0.8

Read the second row against the first, because it inverts the usual reading. Maintainers did not accumulate more incidental movement than lean controls — in short bursts they recorded slightly less. Their entire activity advantage sat in deliberate, sustained bouts, and it was statistically clear against the overweight controls (P = 0.004) while only marginal against the normal-weight ones (P = 0.080). The standard deviations are enormous and there were 26 people in the maintainer group, so this is a sketch rather than a measurement of a population.

It is still a useful correction in two directions. The "hour a day" is closer to 40 minutes of structured exercise than to an hour of anything. And this is a maintenance profile, not a weight-loss prescription: raising incidental daily movement remains a good lever for a stalled diet, as the steps evidence sets out. The registry is describing what people who have already arrived tend to do, which is a different question.

The number the encouraging headline leaves out#

The registry's cheerful statistics are real. Its most sobering one gets far less airtime. Following 2,400 members prospectively for two years — people who had lost 32.1 ± 17.8 kg and held it for 6.5 ± 8.1 years — mean weight change was a gain of 3.8 ± 7.6 kg, and 96.4% were still at least 10% below their maximum lifetime weight at year two³.

Now the part underneath. Of the 65.7% who gained any weight in year one, only 11.0% had returned to their baseline weight by year two. Of the 25.5% who regained 5% or more, only 4.7% got back. The authors' summary is blunt: recovery from even minor weight regain was uncommon.

That finding reframes the registry's most-cited behavior. Frequent self-weighing is not on the maintainers' list because measurement is virtuous; it is on the list because in this population, ground given back is rarely recovered, so the only reliably winnable version of the problem is the small one. A regain caught at 2 kg is a different problem from the same regain caught at 8 — the arithmetic of reading the scale often enough to catch it is in how often to weigh yourself.

What a registry can prove, and what it can't#

The methodological objection is real and has been made in print. A critique in the Journal of Nutrition Education and Behavior argues directly against "the claim that the National Weight Control Registry provides data showing that a significant number of adults in the United States have achieved permanent weight loss"⁵. That critique's authors state their own position openly in the abstract — they advocate a health-at-every-size approach and believe weight-loss dieting is "misleading and futile" — which is a declared perspective, not a hidden one, and this blog flags advocacy on both sides of a contested question.

Split the objection into its two halves and they do not fare equally. The denominator point is simply correct: a self-selected, self-reporting, largely white, female and college-educated volunteer registry cannot establish a population success rate, and nobody should quote it as one. The stronger conclusion — that maintenance is therefore futile — does not follow from the same premise, because a sample that cannot measure a rate also cannot show it is zero, and randomized maintenance trials measuring outcomes in unselected people exist and find real, if modest, effects.

The cleanest demonstration of the registry's limit is its own most famous habit. Among 2,959 members, 78% reported eating breakfast every day and only 4% never did⁴ — a striking figure that launched a decade of breakfast advice. The authors were careful, concluding only that breakfast "is a characteristic common to successful weight loss maintainers and may be a factor in their success." When breakfast was later assigned rather than observed, the weight effect largely evaporated — the trial evidence is in does breakfast help weight loss.

That is the whole lesson in one behavior. The registry is a reliable description of what winners have in common and a poor guide to which of those commonalities is load-bearing. Use it to calibrate expectations about effort and vigilance, cross-check any specific habit against a trial that assigned it, and get the daily behaviors themselves from the habits with intervention evidence behind them — because the deficit that produced the loss is not the thing that holds it.

FAQ#

What is the National Weight Control Registry?#

An ongoing observational database, started in 1994, of adults who have lost at least 13.6 kg (30 lb) and kept it off for at least a year, recruited mainly through advertising and followed by questionnaire. The founding cohort of 784 people had lost an average of 30 kg¹. It is a description of successful maintainers, not a trial, and not a random sample of anyone.

Does the registry prove most people can keep weight off?#

No, and it cannot — everyone enrolled had already maintained a loss before joining, so there is no denominator to compute a rate from. That criticism has been published in the peer-reviewed literature⁵, though its authors also draw a stronger conclusion — that weight-loss dieting is futile — which the same missing denominator cannot support either. The registry's value is descriptive: it shows what maintenance looks like when it works.

Do weight-loss maintainers really exercise an hour a day?#

Measured objectively, closer to 40 minutes of structured exercise. Twenty-six registry members averaged 41.5 ± 35.1 min/day of moderate-to-vigorous activity in bouts of at least 10 minutes, against 25.8 in normal-weight controls and 19.2 in overweight controls — but slightly less incidental short-burst activity than either control group². The difference is deliberate exercise, not a generally busier day, and the sample is small enough to treat as a sketch.

Sources#

This article was researched and drafted with AI assistance and reviewed for accuracy by the BurnWeek team. It is general information, not medical advice. How we research and correct our articles →