The hospital sector in Britain has, over the last decade, achieved substantial reductions in the average length of stay for patients aged 65 and over. One consequence of this may be increased readmission rates. Furthermore, readmission rates are increasingly being proposed as a surrogate measure of outcome after hospital treatment. All admissions of people aged 65 + to two inner-London hospitals in May 1988 were monitored for 6 months after discharge and readmission rates calculated. Of the 386 patients discharged, 130 (38%) were readmitted within 6 months. The 1 week readmission rate was 6%; at 1 month it was 18%. Readmissions showed no variation regarding age or sex of patients but were related to specialty of treatment and length of stay. Consistently, those readmitted has a shorter length of stay than those not readmitted. If readmission rates are to have any utility as a surrogate outcome indicator they must be calculated on a common basis which relates to unplanned readmissions occurring fairly rapidly after discharge, and are related to the index admission.