We have analyzed data on a group of 269 patients with papillary thyroid carcinoma followed on average for 12 years to determine (1) if a prognostic classification scheme can be used to predict an appropriate surgical approach; (2) the effect of treatment on prognosis; and (3) if patients with a "excellent" prognosis benefit from more extensive surgical resection and 131I ablation. Prognostic classification schemes developed by the American Joint Commission, Cady et al., Hay et al., the European Thyroid Association, and our own clinical class scheme each appropriately divided patients into risk category groups. With each system, some patients classified in the low risk group eventually died of the tumor. Considering the excellent but not perfect precision of the prognostic schemes, the need for detailed pathologic analysis, and ideally postoperative thyroid scanning, we conclude that the prognostic classification schemes do not allow the decision at the operating table regarding the appropriate extent of surgery. Patients followed at our institution, operated on by one of three experienced surgeons, and usually given 131I ablation were compared to other patients in the follow-up group operated on by other surgeons and not routinely ablated. There were significantly fewer deaths and recurrences among the patients managed by our method. However, when the groups were restricted to those considering only patients who had more extensive surgery, postoperative 131I ablation, or both, the differences between the groups became insignificant. This finding indicates that the difference in prognosis, comparing patients treated at our institution and those initially treated elsewhere, was primarily due to the routine use of more extensive surgery and postoperative radioactive iodide ablation.(ABSTRACT TRUNCATED AT 250 WORDS)