The technology for patient intravenous self-administration of morphine has been successfully implemented in postoperative and other clinical settings and can be used with terminal patients who experience pain. The question of whether patients who use such instrumentation will be vulnerable to over-medication or development of addiction has not been addressed. This report reviews two competing theories that bear upon this question and tests their predictions about self-administration of morphine for pain relief using data obtained from patients in a bone marrow transplant unit. The first, Opponent Process Theory, predicts escalating drug use and the development of addictive behavior in patients who self-administer morphine. The second, Control Theory, predicts that patients will self-regulate pain effectively by administering morphine without developing problems of medication abuse or addiction. Patients self-administering morphine for 2 weeks were compared to controls who received the drug via routine staff-controlled continuous infusion procedures. Self-administering patients used significantly less morphine than controls and still achieved the same amount of pain control; moreover, they terminated drug use sooner than controls. The predictions based upon Opponent Process Theory were not supported in these marrow transplant patients, but Control Theory accounted well for the outcomes. These results support the assumption that self-administration of opioids in a medical setting does not put patients at risk for over-medication or addiction.