FitChef

The log: one day.

October 5, 2026

12 corrections on 12 pages. Most arrive together: one paper read again against every page that cites it, and every page fixed the same day.

Bonham et al. 2025 7 corrections · 7 pages

October 5, 2026 · We got it wrong

Night-shift fasting trial: it did not show that having a structured plan, any plan, is what makes weight loss work

What it said
The pages said the trial showed "the active ingredient is dietary structure itself, not the fasting window", that "any structured eating approach works equally well", that "having any plan at all was the lever" and that "the method matters less than the commitment".
What it says now
All three plans in the trial cut energy by a similar amount, and every group had dietitian visits and study food; the authors read the results as consistent with that energy reduction. The pages now report that, and no longer credit the trial's results to structure or to having any plan.
Why
With every group on a calorie cut, a dietitian and study food, and no group without a plan, the trial cannot separate structure from those, and the paper does not claim it. An earlier correction (2026-10-04-night-shift-fasting-trial-no-do-nothing-group) fixed "both beat doing nothing"; these lines made a related claim the paper does not make.
Built on it

2 shorts

The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Night-shift fasting trial: more details of the trial were wrong

What it said
The claim said the trial was "designed to answer exactly one question", that the online fasting advice for night shift workers "had no trial-level support" before it, and credited the authors' "viable alternative" conclusion to a press release; the study page said "Most fasting studies compare fasting to nothing", that a review confirmed the trial as "by far the most rigorous test in the field", that the results "Probably also" applied in the UK, New Zealand and Canada, and that "80 dropped out" (a "32% dropout rate"); the claim also gave a 32% dropout; the guide described calorie cutting as "just slightly smaller portions" and said whether fasting is easier to stick with "stays open".
What it says now
The trial had two co-primary outcomes, weight and insulin resistance; twice-a-week fasting had not been tested in night shift workers before it; "a viable alternative" is the paper's own conclusion; the line about other fasting studies is removed (the paper's background cites trials comparing modified fasting with calorie restriction); the trial is the largest and longest weight-loss trial in night shift workers; calorie cutting meant about 20% less energy with a dietitian's help; the authors say the results may not be generalisable to all groups; people stayed with all three plans at similar rates; 170 of the 250 completed the six months, so 32% did not (the paper counts 70 dropouts).
Why
Each detail was checked against the paper and was wrong as written. An earlier correction (2026-10-04-night-shift-fasting-trial-details) fixed other details of the same trial.
The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Night-shift fasting trial: more lines on the cholesterol result went beyond the paper

What it said
The guide said one fasting group showed "clearly lower cholesterol" and that "the research team who found it called it unproven"; the claim said fasting had no metabolic advantage, that the two diets produced "virtually identical outcomes on every measured dimension", and called the cholesterol result confounded by baseline weight and possibly timing-specific; the study page called it an "unexpected cholesterol benefit" and said it needed "independent confirmation".
What it says now
The group fasting on day shifts or days off ended with lower total and LDL cholesterol than calorie cutting, an exploratory result; the authors say that group's larger weight loss could have contributed and that the finding needs validating in further studies. Insulin resistance did not differ between the groups, and neither did the other secondary measures.
Why
An earlier correction (2026-10-04-night-shift-fasting-trial-cholesterol) fixed the main cholesterol lines; these further lines on the same result went beyond what the paper says.
The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Night-shift fasting trial: it tested weight loss, not a reversal of weight gained on shifts

What it said
The pages said the trial "showed that structured calorie reduction reverses the weight trajectory", "tested whether structured dieting can reverse the trend", that "The reversibility is not" debatable, and that it showed any structured approach "can address" the shift-work body composition change.
What it says now
The trial enrolled night shift workers with overweight or obesity whose weight had been stable for three months, and compared ways of losing weight; the pages now describe it that way.
Why
Its participants had been weight-stable before they joined, and it had no group without a diet plan, so it cannot show a reversal of weight gained on shifts.
The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Night-shift fasting trial: the 61% counts all three plans together

What it said
Next to the 61% figure the pages said "regardless of whether they fasted", "It did not matter which plan", "The number held across all three approaches" and "across ALL arms".
What it says now
61% of those who finished six months lost at least 5% of their weight, with the three plans counted together; counting everyone who started, it was 40%.
Why
The paper gives the 61% for all finishers pooled, not for each plan. An earlier correction (2026-10-04-night-shift-fasting-trial-61-percent-finishers) fixed who the 61% counts; these lines read it as holding in every plan.
The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Night-shift fasting trial: it tested which days to fast, not an eating window

What it said
The pages said the trial showed that the "timing window" or "eating window" made no difference, and that the calorie-cutting group had "no fasting windows"; several lines called the group that fasted on day shifts or days off the "day-shift" group.
What it says now
The fasting groups were asked to keep to 2,100 to 2,500 kJ on two days a week, with no timing set for those days: one group fasted on day shifts or days off, the other on night shifts, and each was compared with daily calorie cutting. The pages now say so.
Why
The paper prescribes no timing on fast days and places no restriction on how energy is spread across them; what it varied was which days were fasting days. Our pages described a different kind of fasting.
The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Night-shift fasting trial: the weight gaps were 0.2 and 0.6 kg, and the results similar, not the same

What it said
The claim and the study page gave the difference between fasting and calorie cutting alone, as "two hundred grams" and "0.2 kilograms"; the study page said "the timing advantage does not exist" and "a larger study would almost certainly have confirmed the same nothing"; the pages called the results "the same", "identical" or "equivalent".
What it says now
The trial compared each fasting schedule with daily calorie cutting: the weight gaps were 0.2 kg and 0.6 kg, neither statistically significant, and the authors judged them too small to matter clinically. The pages now say the results were similar.
Why
The paper reports both comparisons and calls the outcomes similar. It was a superiority trial, not built to show the diets are equivalent, and its authors say no difference is far more likely than a clinically important one the trial was too small to detect, not that a larger trial would certainly confirm it. Our pages reported one comparison and stated the result more firmly than the paper.
The paper
Bonham MP, Leung GKW, Rogers M, et al. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial. eBioMedicine. 2025;117:105803.Read the paper
Found by
Our own reading of the whole page against its paper.
Jeffery et al. 1993 4 corrections · 5 pages

October 5, 2026 · We got it wrong

Meal-delivery study page: one of the paper's Table 3 p values was wrongly said not to match its F statistic

What it said
The page said the printed p values of both centre interaction rows in Table 3, Time x Center and Time x Treatment x Center, do not follow from their F statistics and degrees of freedom.
What it says now
It says this of the Time x Treatment x Center row only.
Why
Recomputed from Table 3 of Jeffery et al. 1993: Time x Center, F(3, 417) = 0.28, gives p = .84, which matches the printed .8414. Time x Treatment x Center, F(12, 417) = 0.28, gives p = .99 against the printed .7347. Our note named both rows, so it said the paper had an inconsistency in a row where it has none.
The paper
Jeffery RW, Wing RR, Thorson C, Burton LR, Raether C, Harvey J, Mullen M. Strengthening Behavioral Interventions for Weight Loss: A Randomized Trial of Food Provision and Monetary Incentives. J Consult Clin Psychol. 1993;61(6):1038-1045.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Meal-delivery trial: the authors' suggestion that free food drew people to the meetings was overstated

What it said
The pages said the trial's authors wondered whether the free food was "simply a reason to show up".
What it says now
They say the authors wondered whether the free food was a reason to show up, without "simply".
Why
Jeffery et al. 1993 list free food, which "may have been an incentive for attendance and completion of homework assignments", as one of three parts of the food programme that may explain its effect, beside a detailed structured diet and food packed in set portions. They say more research is needed to tell which parts mattered. "Simply" made it sound as if the authors suspected the food worked only by getting people to the meetings.
The paper
Jeffery RW, Wing RR, Thorson C, Burton LR, Raether C, Harvey J, Mullen M. Strengthening Behavioral Interventions for Weight Loss: A Randomized Trial of Food Provision and Monetary Incentives. J Consult Clin Psychol. 1993;61(6):1038-1045.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

Meal-delivery trial: both sides had the same weight-loss programme; the paper does not say they shared the same groups

What it said
The pages said both sides of the food comparison "sat in the same weight-loss groups" or went to "the same group meetings", and compared the people given food with "people in the same group without the food".
What it says now
They say both sides went through the same group weight-loss programme, which is what the paper reports.
Why
Jeffery et al. 1993 randomised people to five treatment groups and say the behavioural treatment given with the food was identical to the treatment without it, delivered in groups of about 20. The paper never says people with and without the food met in the same groups. Our pages turned "the same programme" into "the same groups". The comparison itself stands: the programme was the same on both sides, so the difference at one year is food against no food.
The paper
Jeffery RW, Wing RR, Thorson C, Burton LR, Raether C, Harvey J, Mullen M. Strengthening Behavioral Interventions for Weight Loss: A Randomized Trial of Food Provision and Monetary Incentives. J Consult Clin Psychol. 1993;61(6):1038-1045.Read the paper
Found by
Our own reading of the whole page against its paper.

October 5, 2026 · We got it wrong

A claim page showed a trial's food-diary percentages as a count of days

What it said
The Statistical detail box gave the 1993 Jeffery trial's food-diary figures as "Food-diary days completed: SBT + I 81 / 48 / 24; SBT 67 / 33 / 23 (Table 4)", which reads as a number of days.
What it says now
It gives them as the trial reports them: "Percentage of food days completed (Table 4): SBT + I 81% / 48% / 24%; SBT 67% / 33% / 23%", the share of assigned food-diary days that people in the incentive group and in the group without incentives filled in during weeks 1-20, 21-52 and 53-78. The figures are unchanged.
Why
The trial's Table 4 is titled "Percentage of Food Days Completed": each value is the percentage of assigned diary days completed in that period, not a count of days. When we summarised the trial for this page, the label was shortened and the unit dropped, and our number checks passed it because every digit matched the paper. The same slip was caught on our meal-plans claim before that page went live; this page was built from the same kind of summary and went live with it. Our first fact-check on claim pages now confirms each number's unit, not only its digits.
The paper
Jeffery RW, Wing RR, Thorson C, et al. Strengthening Behavioral Interventions for Weight Loss: A Randomized Trial of Food Provision and Monetary Incentives. Journal of Consulting and Clinical Psychology. 1993;61(6):1038-1045.Read the paper
Found by
Our own checks of the site.
Night-shift fasting pages: a smaller meal-replacement trial was said to show similar results

October 5, 2026 · We got it wrong

What it said
"Hospital shift workers in a completely different country — no fasting involved, just structured meal replacements — lost similar weight. Different plan, same story." and "A smaller study of 50 shift workers in Sri Lanka found similar results with a structured meal replacement". The study page also marked that trial as confirming the 24-week trial and called its weight loss meaningful.
What it says now
The pages now describe it as a different plan, a low-calorie meal replacement in place of dinner, tested in 50 hospital shift workers over eight weeks with no fasting involved; the claim adds that weight came down slightly in that time, and the study page no longer marks it as confirming the trial or calls its loss meaningful.
Why
That trial compared a meal replacement at dinner with a usual diet in 50 hospital shift workers for eight weeks. Its authors report slight reductions in weight, nearly 3% in the meal-replacement group, short of the 5% usually counted as clinically meaningful. It did not test fasting, and our pages called its results similar to the 24-week trial's.
The paper
Sooriyaarachchi P, Jayawardena R, Pavey T, King NA. A low-calorie meal replacement improves body composition and metabolic parameters in shift workers with overweight and obesity: a randomized, controlled, parallel group trial. Nutr Metab (Lond). 2024;21:32.Read the paper
Found by
Our own reading of the whole page against its paper.

Newer day: October 6, 2026Older day: October 4, 2026The whole log

The full picture

How every claim on FitChef is checked, page by page.

We did the thinking. You just eat.

Get my week

It takes a minute.

ListenListen0:00Help