© ENVISION/CORBIS

The past couple of years have witnessed a string of disappointing results from long-term studies looking for the benefits of certain diets against chronic disease. First, fruits and vegetables showed no sign of protecting against cancer in general. Then high-fiber eaters found themselves as cancer-prone as the rest. To cap it off, a low-fat diet did nothing to ward off heart disease and colorectal cancer. Given the anti-cancer magic some of us have come to expect from our broccoli and other dietetic food, a question comes to mind: are researchers even doing the studies correctly? Some experts feel the same way. "How do we study this? A number of people are going back to this question," says Harvard University epidemiologist Walter Willett.

The disagreement among researchers turns on the value of expensive clinical studies enlisting lots of people, like the $2 billion Women's Health Initiative (WHI), which aimed to put nearly 20,000 women on a low-fat diet for 8 years. Compliance to the regimen was at most a disappointing 70 percent of the study's goal--to cut fat consumption by nearly half--according to participants' food logs. Even that modest reduction in fat intake by the dieters may have been fudged: blood cholesterol and lipid levels between the low-fat diet and control groups were nearly identical, suggesting that they were dining on similar foods. Perhaps as a result, the only benefit WHI researchers caught any hint of was a slight reduction in the low-fat eaters' rate of breast cancer; only continued follow-up will determine whether or not it is a random bump in the data.

Ross Prentice, lead investigator of the WHI breast cancer study, defends the importance of clinical trials. Smaller-scale studies may be more straightforward to implement, but they are much harder to interpret, he notes. To others, however, clinical trials are almost preordained to fail. The WHI wasn't the first such highly touted clinical trial that did not live up to expectations. In 1982 results came in from the Multiple Risk Factors Intervention Trial, which was designed to reduce the risk of coronary heart disease by counseling an experimental group to stop smoking, lower their blood pressure and cut back on saturated fat. Annoyingly, the control group also started putting down their Camel Lights and Big Macs, and researchers found no difference in outcome between the two groups. Ten years ago Willett and others were already questioning whether the WHI would find any benefits to a low-fat diet, given the compliance problem along with existing evidence that fat intake per se was not a serious health risk. "I think it's going to be a rare situation where a multibillion-dollar kind of study is going to get a clear answer," Willett reflects.

Previous: How Does a Planet Grow?

Next: Does Sprouting New Brain Cells Cure Depression?