If you follow the news about health research, you risk whiplash. First garlic lowers bad cholesterol, then—after more study—it doesn’t. Hormone replacement reduces the risk of heart disease in postmenopausal women, until a huge study finds that it doesn’t (and that it raises the risk of breast cancer to boot). Eating a big breakfast cuts your total daily calories, or not—as a study released last week finds. Yet even if biomedical research can be a fickle guide, we rely on it.

But what if wrong answers aren’t the exception but the rule? More and more scholars who scrutinize health research are now making that claim. It isn’t just an individual study here and there that’s flawed, they charge. Instead, the very framework of medical investigation may be off-kilter, leading time and again to findings that are at best unproved and at worst dangerously wrong. The result is a system that leads patients and physicians astray—spurring often costly regimens that won’t help and may even harm you.

Here’s the rest of the News Week Article.