Current practice with digoxin was assessed in a group of 42 elderly patients by comparing plasma digoxin concentrations attained on previously established maintenance doses with those generated by a computer programme designed to calculate dosage schedules to suit individual patients. Discrepancies between measured and computed plasma levels and between established and computed doses dictated withdrawal of the drug or revision of dosage in 26 patients (62%), with obvious clinical benefit. An important determinant of dosage was renal function; reduction in creatinine clearance provided good evidence for the loss of ability of the elderly kidney to eliminate digoxin. Simple bedside methods are available which permit a reliable estimate of creatinine clearance without a 24-hour urine collection, provoding a rational basis for the choice of digoxin dosage in the elderly.