The continuing controversy between proponents of open and closed commissurotomy might be clarified by analysis of late follow-up with modern actuarial techniques that provide a true perspective of patient risk. We have used open mitral commissurotomy exclusively for 15 years in 100 patients. There was one operative death from pancreatitis and one late death from cancer; the actuarially projected survival rate (+/- the standard error) at 10 years is 97 per cent (+/- 2). Thirteen patients had preoperative emboli, 6 of whom were in sinus rhythm and 7 in atrial fibrillation. Two patients had postoperative emboli, both in sinus rhythm. The actuarial chance of remaining free of embolism at 10 years is 97 per cent (+/- 2). Sixteen patients required reoperation on the mitral valve for functional deterioration. The remaining survivors were in Class I or II when last seen. The actuarial chance of not requiring a reoperation after 5 years is 91 per cent (+/- 4) and at 10 years, 38 per cent(+/- 16). Results in different centers are difficult to compare for many reasons, but imprecise statistical methods further obscure such comparisons. The use of actuarial techniques may help to define the role of open mitral commissurotomy.