Doctors are funny – and naive. They think that eating 5 or more servings of fruit and vegetables clearly reduces the risk of getting cancer in the general public. When they write books, they cite this evidence and make no mention of contrary evidence.
Do doctors not believe in evolution?
Why would plants be wholly beneficial to the animals that eat them? Is not caffeine toxic to at least some of the insects that feed on the coffee plant? Since caffeine is not toxic to us (estimated LD50 of about 70 cups of coffee per day), could it be that we are too large to succumb, or that our ancestors have been selected for resistance to the poisonous alkaloid, or some such related compound?
If we did not rather quickly metabolize caffeine into less toxic substances (into 1,3,7-trimethyl-uric acid, theophylline (and then into 1-methylxanthine, and then into 1-methyl-uric acid), theobromine, and paraxanthine (and then into 1,7 dimethyl-uric acid, AFMU, 1-methylxanthine, and then into 1-methyl-uric acid), we would probably succumb to the poisonous alkaloid as well with a lot less than 70 cups of coffee a day.
There are numerous plant toxins, from cyanide generators like cassava and mustard generators, to lectins (like lima bean lectins to people with type A blood antigens), to canavanine, to solanine, etc. Plants are not wholly beneficial. Anyone claiming this believes in fairy tales like the Garden of Eden myth.
Plant species that sicken or poison the animals that eat their leaves or their seeds have a survival advantage as a species. Fruit is an exception to this: the plant gains by having the animal eat the fruit, or just sample it and throw it away like the fruit of the lemon tree, and survive at least long enough to deposit (and fertilize) the seeds somewhere else.
Animals that harbor mutations that enable them to resist plant toxins or that harbor mutations that make good use of (former) toxins have a survival advantage over animals that succumb to the toxins.
From these facts of evolution, I would predict that plants contain compounds that are demonstrably harmful to our health and plants are likely to contain compounds that are beneficial to our health, mostly as a result of animal adaptations. One such adaptation is our sheer size: plants are unlikely to be able to produce enough toxin to kill us in one feeding, and our strongest taste sensation, that of bitterness, warns us against many of their poisons. Another adaptation is omnivorousness itself: by sampling many different foods, we are less likely to over-consume any one food that may still be toxic to us.
Because of the facts of evolution, the net effect of consuming large amounts of plant foods on survival or on cancer per se would be predicted to be small in any case. This is exactly what is observed if one does not cherry pick the data.
To gain a clear beneficial effect from any plant nutrient, we must purify compounds with a more beneficial effect away from all of those with more harmful effects. That is, we must supplement.
What do we see? Consider bladder cancer: half a dozen studies conclude that consuming 5 or more servings of fruit and vegetables have no net effect on bladder cancer prevention. Slightly more than half a dozen studies find a benefit. Conclusion: if there is a health benefit, in this case measured as resistance to bladder cancer in the general public, it is slight, exactly as predicted from simple evolutionary considerations.
Religious considerations: if a loving God created plants and animals, save detritivores, to be our food, then of course we would expect that all food except detritivores, is simply beneficial and no harmful compounds would ever be found – unfortunately, the data are against this idea. Are our doctors showing vestiges of their religious heritage?