The long-term clinical outcome and costs of treatment of hyperthyroidism with radioiodine have been examined in two cohorts of patients from Sheffield and Scotland. The majority of patients in both series were considered to have Graves' disease. The Sheffield patients (660) were included in a trial of three radioiodine dose regimens of 3,500 (312), 7,000 (323) and 14,000 (25) rad determined using a formula for accurate dosimetry. The Scottish patients (3,920) drawn from five centres in Aberdeen, Dundee, Edinburgh, Glasgow and Inverness were treated using an arbitrary scale, for the activity of radioiodine administered, related to goitre size. Their results are grouped into five MBq 'dose' bands: 37-185, 186-370, 371-555, 556-740 and 741+. The proportion of patients with persistent hyperthyroidism was higher in both cohorts for low-dose radioiodine regimens, but 15-25% of patients who received high doses showed persistent hyperthyroidism. Early and late onset hypothyroidism was lower after low doses but differences between the treatment groups were small in terms of clinical benefit. Total morbidity at 10 years follow-up, in terms of hyperthyroidism, and hypothyroidism, was highest after low-dose therapy. There was little variation in total costs, but patient costs were lowest for the Scottish regimen and highest for low-dose therapy. A dose of at least 370-555 MBq which will ensure early elimination of hyperthyroidism will also limit the medical workload and total costs.