General Interest
Case study · Faults 2, 3 and 6
The news says a blood test can now detect Alzheimer’s disease before any symptoms appear. Since 2024 the official definition is a biological process, present whenever one biomarker is abnormal, with or without symptoms. The paper takes three claims in turn. A positive test does not mean you have the disease: for a healthy 70-year-old man it means a lifetime chance of dementia of about 14%, with a range from 7% to 23%. The line between disease and no disease runs through a smooth measurement, and a faulty reagent lot once moved patients across it. And two people with the new diagnosis can share one protein and nothing else. For people who already have symptoms, the tests are a real advance.
PDF · 4 pages · Posted November 2026
Case study · Faults 4, 5 and 6
The Los Angeles Times says creatine is for probably everybody: muscle and bone after 40, memory and mood in menopause, thinking in older adults. The paper takes the claims one at a time against the trials. One holds up: older adults who lift weights two or three times a week gain about 1.4 kg of lean mass. The two largest trials found nothing for bone, and the memory meta-analyses disagree about who benefits. The paper then goes through the harms the story left out: diarrhea at 10 g servings, more people quitting, and a blood test that misreads kidney function. A training supplement, not a supplement for everybody.
PDF · 10 pages · Posted November 2026
Case study · Fault 5
MIT Technology Review reports that eating disorders are on the rise and that GLP-1 drugs could make things worse. The 71% rise comes from a study that ran from 1990 to 2023 and never measured GLP-1 use. The drugs came into wide use in the last two and a half of those 33 years. The one survey that does involve the drugs asked people who already had an eating disorder, had no comparison group, and cannot tell whether the disorder or the drug came first. One fault, the causal leap, and nothing in the story shows that the drugs cause eating disorders.
PDF · 3 pages · Posted October 2026
Case study
A story on Apple News gives two answers. Its expert says most people put the weight back on. The rest of the piece leans on a Cleveland Clinic headline that 45% keep it off. The trials took people off the drug and weighed them for a year, and most of the lost weight came back. The 45% counts something else: people who quit early, about half of them back on some treatment, sorted by whether they were a pound up or a pound down. The paper asks what each number is a fraction of, and which people were counted.
PDF · 4 pages · Posted October 2026
Case study · Faults 1, 4, 5 and 6
Vice reported that GLP-1 drugs may make your fingernails fall off. The study behind it was a questionnaire. People on the drug ticked the nail detachment box more often than people not on it, and no one examined a nail. People who are watching their bodies change look harder at their hands, and about half the users had nail problems before they started the drug. The causal leap is the flaw that settles it, and three smaller faults ride along.
PDF · 4 pages · Posted October 2026
A week of stories about hair loss, nail damage, gallstones, pancreatitis, sudden vision loss and performers losing their voices. A follow-up to Is Taking a GLP-1 Drug Cheating? The paper puts a number on each side effect, in whole people, and finds that apart from nausea, vomiting and diarrhea they are rare, and that those measured at more than one dose rise with the dose. People differ in how much of the drug they need, and in the trials many did nearly as well on the lower dose. Why the dose belongs with a doctor watching the patient, not a schedule run to the maximum.
PDF · 5 pages · Posted October 2026
Cut fat, cut salt, take fish oil, eat five a day, eat plants instead of meat. The paper goes through the studies behind seventeen of the most common pieces of dietary advice and asks what each one rests on. Very little of it was tested by putting people on the diet and counting what happened. Where diet was tested properly, it helped people who already had something wrong. For a healthy adult who eats enough, no study has shown that changing your diet makes you live longer. A survey of a whole field, with a verdict table and what is left once the rest is set aside.
PDF · 17 pages · Posted October 2026
Case study · Faults 1, 3, 4 and 5
For fifty years patients were told to limit eggs. Since then the headlines have gone both ways, from a 17% rise in heart disease to eggs not raising cholesterol at all. The cholesterol you eat is not the cholesterol in your blood test. Eating it raises LDL only a little, and no trial has tested whether cutting it prevents heart disease. Four faults in one story, and a 300 mg limit no one can trace.
PDF · 4 pages · Posted September 2026
Women’s Health and Forbes say there are exact steps for keeping the weight off after stopping a GLP-1 drug. The paper checks each step against its evidence. No trial has compared weaning with stopping outright, the randomized withdrawal trials show most of the loss coming back within a year, and the steps themselves are the diet and exercise advice that failed most people before the drugs existed. The one trial offered in their favor is read through the Bernoulli fallacy.
PDF · 4 pages · Posted September 2026
Case study · Faults 1 and 5
Patients are suing the makers of Ozempic, Wegovy and Mounjaro over sudden loss of sight, and the study behind the headlines put the risk at up to seven times. The paper asks what that multiple leaves out and who the study was actually looking at. Two faults in one headline, and the question to bring to your doctor instead.
PDF · 3 pages · Posted September 2026
Case study · Faults 1, 3, 4 and 5
A headline said a vitamin may cut dementia risk almost in half. The study’s own tables show the same number of cases in the highest and lowest intake groups; the halving appears only after the model’s adjustment. Four faults in one article, and a recommendation aimed at readers who already eat more of the vitamin than anyone in the study.
PDF · 3 pages · Posted September 2026
Case study · Faults 1, 3 and 5
Headlines announced that xylitol is linked to heart attacks and strokes. The study behind them measured blood levels of a substance the body produces on its own, sorted people into four bins and compared only the extremes, and reported a percentage with no baseline. Three faults, and none of them support a warning about the gum.
PDF · 4 pages · Posted September 2026
Case study · Faults 1, 2, 4 and 5
Nicotine pouches are sold as a focus aid, a brain protector, and a safer habit. One wellness claim turns out to be four claims, and each rests on a different fault: who uses it read as what it does, a Parkinson's correlation with the arrow running backward, a small memory trial whose two-year follow-up found nothing, and "safer than smoking" with the comparison left unsaid. A case study in applying the faults to a real claim, written for readers who have not taken the course.
PDF · 5 pages · Posted September 2026
Body weight is regulated by hormones, not by resolve, and GLP-1 signalling runs weak in people carrying extra weight, so the drug supplies an input the body is not producing at strength. Works through what the hormone measurements show a year after a diet, why the concept of cheating does not apply, and what it means that the weight returns if you stop.
PDF · 4 pages · Posted August 2026
Risk of vascular death climbs smoothly with usual pressure, with no safe ledge down to about 115 over 75, which is exactly what a cutoff at 130 or 140 hides. The paper separates what systolic and diastolic each tell you, then builds a small Bayesian network that takes age, sex and the two readings and returns a ten-year probability.
PDF · 6 pages · Posted July 2026
How the BMI curves used in the course are built: shape from a pooled analysis of 239 studies, level from a life table, context from the population distribution. The WHO bands only shade the figures. They never enter the calculation, because the bands are a lossy summary of a smooth signal rather than biological boundaries.
PDF · 4 pages · Posted July 2026