We report the first randomized study comparing early hospital discharge with standard hospital-based follow-up after high-dose chemotherapy (HDCT) and PBSCT. Patients aged 18-65 years, with an indication of PBSCT for non-leukemic malignant diseases were randomly assigned between two arms. Arm A consisted of early hospital discharge (HDCT during hospitalization, discharge at day 0, home stay with a caregiver, outpatient clinic follow-up). In arm B patients were followed up as inpatients. In total 131 patients were analyzed (66 in arm A and 65 in arm B). Patient characteristics and hematological reconstitution were comparable between the two groups. In arm A, 26 patients were actually discharged early. Patients in group A spent fewer days in hospital (11 vs 12 days, P=0.006). This strategy resulted in a 6% mean cost reduction per patient when compared with the conventional hospital-based group. The early discharge approach within the French health system, while safe and feasible, is highly dependent on social criteria (caregiver availability and home to hospital distance). It is almost always associated with conventional hospital readmission during the aplasia phase, and limits cost savings when considering the whole population of patients benefiting from HDCT in routine clinical practice.