1953 · Ancel Keys
Atherosclerosis: A Problem in Newer Public Health
What this work is
A lecture arguing that coronary heart disease should be treated as a public health problem tied to diet. Keys plots national dietary fat intake, as a share of calories, against heart disease death rates in middle-aged men for six countries and reports a rising curve. He proposes that dietary fat raises blood cholesterol and that cholesterol drives the arterial deposits behind heart attacks.
Then and now
Keys linked heart deaths, blood cholesterol, and diet across countries, and found that eating cholesterol barely changed blood cholesterol. The diet-heart hypothesis grew from this paper into low-fat guidelines and cholesterol drugs. His own data never showed cholesterol causing heart disease, and nobody has seen that.
Verdict
UNSCIENTIFIC
Neither the paradigm nor the work inside it passes. The verdict is unscientific.
1·This work · inside the paradigm
UNSCIENTIFIC
Does not follow 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 work assumes, and how the work 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·This work
Did this work test its claim with methods that are independent, falsifiable, and non-circular?
UNSCIENTIFIC
This work claims that diet, blood cholesterol, and deaths from degenerative heart disease rise together, and that public health should target that link. All three tests fail.
Independently Verifiable
Fail
Death certificates classified causes using the medical model, and the chemical assays produced numbers that the diet-heart framework said were cholesterol and mattered for heart disease.
Falsifiable
Fail
Keys reported that dietary cholesterol did not correlate with serum cholesterol in hundreds of men, and the finding did not end the claim; the paradigm shifted to dietary fat and kept going.
Non-Circular
Fail
The category 'degenerative heart disease' was defined by the same medical model that the serum cholesterol measurements were supposed to validate.
Why
The death category was built by the medical model, and the cholesterol number is the model's reading of an assay. His own finding that eating cholesterol did not change blood cholesterol did not end the claim, which moved on to dietary fat. The same model defines both the disease category and the cholesterol meant to explain it.
2·Paradigm · Diet-Heart Hypothesis
Does the framework this work assumes pass the three tests?
UNSCIENTIFIC
Independently Verifiable
Fail
The diet-heart framework defines heart disease deaths using model-dependent classifications and treats serum cholesterol measurements as meaningful only because the model says so.
Falsifiable
Fail
When dietary cholesterol showed no effect on serum cholesterol in this paper, the paradigm did not die; it absorbed the result and shifted to dietary fat.
Non-Circular
Fail
The model that defines 'degenerative heart disease' is the same model that defines serum cholesterol as the relevant variable to measure.
Why
The diet-heart hypothesis treats model-classified heart deaths as proof of the model. Keys himself wrote that the arterial condition cannot be precisely evaluated in life and is all too seldom verified at death, and he used the death classifications anyway. Blood cholesterol is treated as the problem to push down, not as a sign of what damaged the body. When dietary cholesterol showed no effect, the hypothesis moved to dietary fat and kept going. Countries that do not fit are called confounded.
From the work
the pathological condition cannot be precisely evaluated in life, and is all too seldom verified at death
it is convenient to group them, for the present purpose, as "degenerative heart disease."
extensive surveys of the serum cholesterol concentration in individuals of differing cholesterol intake consistently fail to show a relation between these variables