The good news about natural treatments is that scientists have begun to take them seriously and over the past few years have conducted large clinical trials to find out if some are effective. The bad news is that for the most part, the results of these trials have been negative.

You may have noticed a typical progression: from case studies to open label trials to small randomized controlled trials to definitive, large, well-designed controlled trials. Typically, the larger the trial, the poorer the result for the natural medication.

The Chinese herb ginkgo biloba is a good example. After years of open label trials suggesting a benefit for preventing dementia, a 1997 study caused a real stir when it seemed to provide solid evidence that ginkgo worked. Researchers randomized 309 patients with dementia to either ginkgo or placebo and found that after one year, those taking ginkgo averaged a modest improvement in a standard cognition score—namely, a 1.4 point advantage in the ADAS-Cog (Le Bars PL, et al., JAMA 1997;278: 1327–1332).

This promising result prompted others to plan larger studies, and then things started to fall apart for ginkgo. In the largest study, which was funded by the National Center for Complementary and Alternative Medicine, 3,069 elderly patients with normal cognition or mild cognitive impairment were randomly assigned to either ginkgo (120 mg twice daily) or placebo. Every six months, the subjects were assessed for dementia. After a median follow-up period of 6.1 years, a total of 523 patients developed dementia, with no significant difference between ginkgo and placebo. While ginkgo was generally safe, there were twice as many hemorrhagic strokes in the ginkgo group than in the placebo group. This difference did not quite reach statistical significance because the overall numbers of these events were small (Dekosky S et al., JAMA 2008;300 (19):2253–2262).

The conclusion of the Dekosky study was clear—ginkgo does not prevent dementia. Why was this study more trustworthy than the 1997 study? Three factors stand out and bear remembering as you evaluate this and other clinical trials.

1. Sample size. The 1997 Le Bars study enrolled 309 patients; the Dekosky study enrolled 3,069. While determining an adequate sample size can be complicated, bigger samples are generally more reliable than smaller samples. One reason is that when only a small number of patients are randomized to two treatment groups, there is a possibility that the two groups will differ in some way relevant to the treatment. In very large samples, this possibility is greatly reduced, and we have more confidence in the validity of the results.

2. Dropout rate. The dropout rate in the Le Bars study was 50 percent; the dropout rate in the Dekosky study was only 6.3 percent. Similar to the sample size argument, the higher the dropout rate, the less confidence one has in the results of a study because they are based on a small—and potentially unrepresentative—number of patients.

3. Study length. The Le Bars study lasted one year; the Dekosky study lasted six years. Since dementia is gradually progressive, a longer study provides more useful information than a short study.

Another recent example of the debunking of a natural treatment is antioxidant vitamins. Earlier secondary analyses of clinical trials indicated that selenium, vitamin E and vitamin C may prevent cancer, but results were inconsistent and the study designs were not adequate. Therefore, two large placebo-controlled trials were designed specifically to assess the efficacy of these vitamins. One study of 15,000 male doctors found that neither vitamin E nor C helped to prevent cancer (Gaziano J et al., JAMA 2009;301(1):52–62). Another study of 35,000 men found that selenium and vitamin E did not prevent cancer, and in fact, selenium may increase the risk of both prostate cancer and diabetes. Because of these risks, the study was halted early (Lippman S, JAMA 2009;301(1):39–51).

Some vitamins, especially the B vitamins, are marketed as preventing anxiety and depression. This claim is based both on theoretical grounds (some of the B vitamins are involved in the synthesis of serotonin) and on studies showing that patients in psychiatric treatment sometimes have lower levels of B12 and folate (vitamin B9). But until recently, no study had actually compared multi-B vitamins with placebo. In 2008, researchers looked at 299 men older than 75 with histories of depression, but who were not depressed at the time of study. They were randomly assigned to either a combination of B6, B12 and folate, or to placebo. After two years, 84 percent of the men on vitamins remained depression-free vs. 79 percent of those on placebo—not a statistically significant difference (Ford AH, et al., J Clin Psychiatry 2008;69:1203– 1209). While the study had methodological flaws, it is currently the only solid evidence we have regarding the efficacy of B vitamins in psychiatry, and it is not promising.

Even if some complementary medications don’t work, can’t we prescribe them as placebos? The American Medical Association officially says that prescribing placebos without informing the patient is an unethical practice. Nonetheless, according to a recent representative survey of 679 U.S. physicians, about half reported prescribing a placebo on a regular basis (Tilburt J e al., BMJ 2008;337:a1938). Most of these were “active” placebos, meaning pharmacologically active pills that were not thought helpful for the specific condition being treated. Most commonly, these were vitamins or over-the-counter analgesics. When asked how they introduce the placebo to their patients, 68 percent said they describe it as “a medicine not typically used for your condition, but which might benefit you,” while only five percent simply called it a “placebo.”

The question is whether it makes sense for us to prescribe some of the unproven natural remedies in order to take advantage of their placebo effects. We know that the placebo effect is particularly strong in psychiatry, with one meta-analysis indicating that 75 percent of the benefit of antidepressants is due to the placebo effects (Khan A et al., Arch Gen Psychiatry 2000; 57:311–317). With the skyrocketing popularity of supplements, perhaps we finally have a way of harnessing the powerful yet mysterious placebo cure.