The time has come to evaluate critically our practice of cancer pain management and the assumptions on which it is based. We owe it to our patients to maximize the quality of their lives and to provide evidence for them that is based on a scientific approach rather than anecdotal experience. From the information available, opioids do have effects on cognitive and psychomotor function, and although many of these effects diminish once the patient is on a stable dose, the evidence suggests that baseline pretreatment levels are not achieved. In addition, the relationship between measurable effects and the performance of everyday tasks such as driving is unclear. The challenge we now face is to continue the improvements in cancer pain control achieved over the last 25 years. The management of the central adverse effects of opioids must be focused on accurate assessment and careful titration of opioids against pain. Adjuvant analgesic drugs and non-drug measures should be used whenever possible, and drugs should be chosen that will not contribute to existing difficulties. The appropriate use of psychostimulants has yet to be established as has the relative benefit of one opioid over another in cancer pain.