Concern about abuse/dependence in chronic pain patients taking opioid analgesics may lead to undertreatment of pain, yet little is known about the prevalence of abuse/dependence in these patients and how it differs among analgesic agents. The objective of this study was to assess the prevalence of tramadol abuse compared to nonsteroidal anti-inflammatory drugs (NSAIDs) and hydrocodone-containing analgesics in patients with chronic noncancer pain (CNP). The study had three arms. The first arm consisted of subjects prescribed tramadol alone; the second of subjects randomized to either NSAIDs or tramadol; and the third of subjects randomized to hydrocodone or tramadol. Each investigator received two boxes of prescriptions randomized so that one in every four prescriptions was for tramadol. Upon deciding on the therapeutically appropriate arm, the physician selected the appropriate box, opened the next envelope and completed the enclosed prescription. After the initial randomization, physicians could prescribe whatever medication was therapeutically appropriate. A total of 11,352 subjects were enrolled. Up to nine interviews using a structured questionnaire were conducted over a 12-month period. An algorithm called the "Abuse Index" was developed to identify subjects who were abusing the drug. The primary components of the index were increasing dose without physician approval, use for purposes other than intended, inability to stop its use, and withdrawal. The percent of subjects who scored positive for abuse at least once during the 12-month follow-up were 2.5% for NSAIDs, 2.7% for tramadol, and 4.9% for hydrocodone. When more than one hit on the algorithm was used as a measure of persistence, abuse rates were 0.5% for NSAIDs, 0.7% for tramadol, and 1.2% for hydrocodone. Thus, the results of this study suggest that the prevalence of abuse/dependence over a 12-month period in a CNP population that was primarily female was equivalent for tramadol and NSAIDs, with both significantly less than the rate for hydrocodone.