Knee and hip replacements are the most frequently performed elective joint procedures in hospitals in the United States. Pain control in the population of patients with joint arthroplasty presents a challenge for clinicians due to side effects of pharmacologic agents, drug-drug, and drug-disease interactions, especially in older persons. Nonpharmacologic measures could hold promise for pain control in patients undergoing orthopaedic joint procedure. A pilot clinical trial was designed and implemented to evaluate the variability of the effects of a preoperative behavioral intervention (relaxation techniques) on the amount of pain medication used, the number of attempts with the PCA pump, post-operative pain intensity, satisfaction with pain control, and ambulation. Thirty patients, 50 years of age or older, who were undergoing primary knee or hip replacement for osteoarthritis, were randomized to an intervention or control group. Intervention patients received a pamphlet on pain control and practiced relaxation techniques (deep breathing, muscle relaxation and visual imagery) with an audio tape in a preoperative session with a registered nurse. Control subjects received a pamphlet on relaxation techniques. Preliminary analysis of the results showed no significant differences between groups on the pain control outcome measures, and a relatively large variability in individual effects. No significant differences between groups were observed in sociodemographic, health or other variables that could account for the lack of effect of the relaxation intervention.