ABSTRACT: INTRODUCTION: Our case series prospectively evaluate the concept of Patient/Family-Controlled Sedation with midazolam, as an alternative to sedation by continuous infusion in terminal cancer patients. CASES PRESENTATION: Our method was applied in 8 pts. Midazolam was administered in a Patient Control Analgesia mode. The infusion pump was activated "as-needed" by the pt or a caretaker. Sedation was rated as: 1) awake 2) arousable to voice 3) arousable to light pain or 4) unarousable. Family satisfaction was rated as: 1) good, 2) fair, 3) poor, or 4) unacceptable. Mean midazolam consumption was 12 - 40 mg/24 hours. We did not observe respiratory depression. Death occurred 1-6 days after sedation started. Family satisfaction was mainly good and median sedation was in the range 2 - 3. CONCLUSION: Patient/Family-Controlled Sedation with midazolam was effective in providing comfort, by allowing titration of sedation to each patient's needs.