This follow-up study 1 to 5 years after biofeedback therapy, involving 58 patients in six diagnostic groups (migraine headache, tension headache, mixed headache, chronic pain, anxiety, and essential hypertension), revealed that 86% of the patients who continued to practice relaxation techniques improved, while only 50% of those who had stopped practice improved (p = .04). Among the patients who improved, 91% had continued to practice and only 9% had stopped practice, while among the patients who did not improve, 63% had continued to practice and 36% had stopped practice. Patients who were practicing only "occasionally," "as needed," or "when stressed" improved as much as or more than those who practiced regularly and frequently (i.e., at least weekly): 89% versus 77% improved, respectively (p = n.s). There was no difference in the occurrence or frequency of relaxation practice between patients who have been out of therapy 3 to 5 years and those who completed therapy more recently, or between those who were in brief versus longer-term therapy. Although continued relaxation practice is significantly related to the maintenance of long-term improvement, a few patients manage to improve without it, or continue to practice yet relapse. Furthermore, it appears that only occasional relaxation practice after therapy is sufficient to maintain long-term therapeutic gains.