1968 · The Lancet
Controlled trial of soya-bean oil in myocardial infarction
Verdict
UNSCIENTIFIC
The study only partly follows the scientific method inside its paradigm, and the paradigm itself does not pass. The worse layer decides. The verdict is unscientific.
1·Study methodology · inside the paradigm
MIXED
Partly follows the scientific method within Diet-Heart Hypothesis.
2·Paradigm · Diet-Heart Hypothesis
UNSCIENTIFIC
Fails the tests. The worse layer decides.
How the verdict is decided
The verdict grades both layers: the paradigm this study assumes, and how the study was carried out inside it. Both must be scientific for the whole thing to be scientific. A clean method inside an unscientific paradigm is unscientific. An unscientific method inside a better paradigm is unscientific. A split on either layer is mixed. The worse layer decides.
1·Study methodology
Did this study test its claim with methods that are independent, falsifiable, and non-circular?
MIXED
This study claims that a soybean-oil diet low in saturated fat lowered blood cholesterol by 22% but did not change the rate of relapse or death in men recovering from a first heart attack. The tests split.
Independently Verifiable
Fail
The counted relapses are committee rulings: a blinded panel turned hospital notes and death certificates into 'major relapses' by applying the heart model's criteria for a definite coronary event.
Falsifiable
Pass
The claim under test, that the diet prevents relapse, could and did fail here: cholesterol fell 22% while 25 men died in each group, and the trial reported the null.
Non-Circular
Pass
Men were randomized before any measurement and relapses were judged by a committee that never knew the diet group, so no tool's output was used to prove the thing that built it.
Why
They took 393 men under sixty who had just survived a first heart attack and randomly split them. 199 ate a diet low in saturated fat with 85 grams of soybean oil a day, 194 ate as before, and both groups weighed their food for up to six years. Blood chemistry produced numbers the model calls cholesterol: the diet group's fell 22%, the controls' fell 6%. Suspected new events went to a review committee that did not know the diet group, and that committee applied the heart model's criteria to decide which events were definite reinfarctions or coronary deaths. The hospital admissions and the 25 deaths per group are real. But every 'coronary relapse' in the count is that committee's classification, not a thing anyone watched happen, and the cholesterol figure is an assay reading the model names and interprets.
2·Paradigm · Diet-Heart Hypothesis
Does the framework this study assumes pass the three tests?
UNSCIENTIFIC
Independently Verifiable
Fail
The hypothesis's entity, blood cholesterol as the thing that clogs arteries, appears only as an assay number plus the model's story about what that number does inside the body.
Falsifiable
Fail
This very trial produced the killing result, a 22% cholesterol fall with no change in relapses, and the field kept the hypothesis by blaming the trial's size and the size of the drop.
Non-Circular
Fail
The diet was built to move the model's number, the assay was built to read the model's molecule, and the fall in that number was then reported as the diet working.
Why
The diet-heart hypothesis says saturated fat raises blood cholesterol and blood cholesterol causes heart attacks. Both halves ride on the model's own readings. The cholesterol figure is a chemical reaction's output that the model names and gives a causal role to; nobody has watched that substance do anything inside a living artery. The heart attack itself, the event the hypothesis claims to prevent, is a diagnosis assembled from ECG traces and enzyme readings under the same disease model. The model treats the blood number as the cause rather than asking what damaged the man's heart in the first place. And when this trial dropped cholesterol 22% and relapses did not follow, the field did not bury the hypothesis. It pointed to Oslo's positive result and called the London trial too small, the drop too modest, the diet started too late. Every possible outcome had an explanation ready, so no outcome could end it.
From the study
Suspected relapses were assessed at regular intervals by a review committee unaware of the patient's diet group.
The test diet lowered the serum-cholesterol from a mean initial figure of 272 to 213 mg. per 100 ml. at six months (22% fall); the level in the controls fell from 273 to 259 mg. per 100 ml. (6% fall).
The total number of men who had a major relapse at any time in the trial was 45 in the test group and 51 in the controls; of these major relapses 25 in each group were fatal. None of the differences found is significant.
The results are compared with those from a similar trial in Oslo. There is no evidence from the London trial that the relapse-rate in myocardial infarction is materially affected by the unsaturated fat content of the diet.