"TOWARD HEALTHFUL DIETS," the new report of the Food and Nutrition Board of the National Academy of Sciences, not only has increased public confusion over proper diet. It has also soiled the reputation both of the board and of the academy for rendering careful scientific advice.

Its key topic is fat and cholesterol consumption. Offering no new evidence or analysis, the report draws an opposite conclusion -- that Americans should not try to lower their saturated-fat and colesterol intake -- from that reached by virtually every other major medical and public health organization. Though it is nowhere noted, this contradicts the board's own report of just a few months ago -- "Recommended Dietary Allowances" -- which urged that total fat intake "be reduced so fat is not more than 35 percent of dietary energy." Dietary energy means calories and, on average, Americans get 40 percent of their calories from fat.

Why the contradictions? Atherosclerosis -- thickening of the arteries -- and its complications (together known as cardiovascular disease or coronary heart disease) are the leading causes of death in this country. A large body of evidence strongly suggests that high levels of cholesterol and a related type of fat known as LDL in the blood are often associated with atherosclerosis.

It is generally agreed -- until this latest report, it was unquestioned -- that the amounts of saturated fats (those from animal sources) and cholesterol in the diet are related to the levels of cholesterol and LDL in the blood, and that reducing the one will lower the other. Experts are puzzled that this report casts doubt on that relationship, or they flatly disagree. Countless patients who have seen their own cholesterol and LDL levels fall as they follow a doctor's prescribed diet are baffled, too.

True, while all the evidence points in one direction, the key link is still not proven: whether alterations in the diet will definitely reduce the incidence of atherosclerosis and coronary heart disease. Nevertheles, doctors and government agencies must constantly make recommendations on the basis of just this kind of incomplete but suggestive evidence, and there is a consensus on what to do. Even when the precise cause of a disease is unknown, if a certain change in behavior will statistically lower the risk of it, and if the change does not entail new risks or unacceptable economic costs, it should be recommended.

The Food and Nutrition Board agrees with the general principle: "In our present state of knowledge, sound medical and public health practice should be aimed at reducing the known risk factors to the extent possible." But, inexplicably, in the specific it fails to adhere to its own standard. Worse, it follows a double standard, recommending -- on the basis of equally inconclusive evidence -- that Americans lower their salt intake because excessive salt is a risk factor for hypertension.

So how many eggs should we eat, and how much salt should we sprinkle on them? Notwithstanding the Food and Nutrition Board, prevailing medical opinion still is that for the average American it is prudent to lower intake of both. Saturated fats and cholesterol (usually found in the same foods) should be lowered so that fat is no more than 35 percent of total calories.

This moderate reduction involves no known risks. However, a large increase in polyunsaturated fats, though recommended by some groups, may be risky. Reducing fat consumption is also important because fat has twice as many calories per gram as either protein or carbohydrate, and the commonest form of malnutrition in this country is obesity. Obesity -- being 20 percent above proper weight -- is in itself a significant risk factor for hypertension, diabetes, gall bladder disease and coronary heart disease.

A final note: heredity plays a major role in determining likelihood of both coronary heart disease and hypertension. But doctors cannot yet predict which individuals are at risk because of their genes. So particularly those who have a family history of heart disease, but probably everyone, including children, should have blood cholesterol and fat levels determined at least once in the course of a general physical examination.