saw palmetto, kava kava, and valerian root). The health content of all websites listed on the first page of the search results were analyzed for a total of 443 sites that met study inclusion criteria. Among the 443 websites, 338 (76 percent) were retail sites. Eighty-one percent of these retail websites made one or more health claims; of these, 149 (55 percent) claimed to treat, prevent, diagnose, or cure specific diseases. More than one-half (153 of 292; 52 percent) of the sites with a health claim omitted the standard federal disclaimer. Bonakdar (2002) conducted a disease-specific search for herbal drugs for cancer. A majority of sites examined claimed that they offered cures for cancer through herbal supplementation with little regard for federal regulations against doing so, with such claims being more common on sites operated from outside the United States.

Ashar and colleagues (2003) conducted a third study that evaluated information contained within Internet sites that advertise and market dietary supplements containing ephedra. Thirty-two products and advertisements were identified and systematically evaluated for deviance from truth-in-advertising standards. Of the 32 websites analyzed, 13 (41 percent) failed to disclose potential adverse effects or contraindications to supplement use. Seventeen (53 percent) did not reveal the recommended dosage of ephedra alkaloids. More importantly, 11 sites (34 percent) contained incorrect or misleading statements, some of which could directly result in serious harm to consumers. These and other studies (Matthews et al., 2003; Sagaram et al., 2002) illustrate that consumers are commonly misled by vendors’ claims that herbal products can treat, prevent, diagnose, or cure specific diseases, despite regulations prohibiting such statements. Closer monitoring of web sites, enforcement of the Dietary Supplement Health and Education Act of 1994 (DSHEA) and Federal Trade Commission regulations, and the creation of a user-friendly authoritative website or criteria for evaluating existing websites on CAM modalities are much needed.

A third common source of information about CAM-related modalities is health food stores. A handful of surrogate patient studies have been conducted in which a researcher poses as a prospective client and asks for advice about what type of herbal medicine he or she (or a family member) should take for a specified ailment. In one study conducted by Mills et al. (2003), eight data gatherers asked employees of all retail health food stores in a major Canadian city what they recommended for a patient with breast cancer. The data gatherers inquired about product safety, potential drug interactions, cost, and efficacy. Employees at 34 stores were queried, and a total of 33 different products were recommended. Twenty-three employees (68 percent) did not ask whether the patient took prescription medications, 15 employees (44 percent) recommended visiting some type of health care professional, and only 3 employees (8.8 percent) discussed the potential adverse effects of the products.