OBJECTIVE: To assess nationally the use of questionable cancertherapies. DESIGN: Survey of cancer patients or their families regardinguse of questionable methods of cancer treatment; sample survey ofphysicians' perception of use. PARTICIPANTS: 36,000 households; 5,047individuals; 91 physicians. RESULTS: The prevalence of use of questionablecancer methods was nine percent overall. An increase in use was directlyproportional to increased income and education. Prolonged illness andcertain types of cancer were more commonly associated with use. Harmfulside effects of questionable cancer treatments were regarded as modest (sixpercent). There was a wide range in cost. Third-party reimbursement wasreported by 25 percent of patients. Important discrepancies were foundbetween patients' and physicians' perceptions of questionable therapies. CONCLUSIONS: While some questionable therapies are harmless or inexpensive, others have toxic effects and may be costly, and none have scientificallyproven efficacy. Although the percentage of usage reported is relativelylow, overall large numbers of patients are involved, especially in certaingroups. The physician plays a key role in encouraging or preventing the useof questionable methods, and substantial improvements in public andprofessional education are needed.