Background-Information regarding the incidence of structural valve deterioration (SVD) is used in selecting the type of valve for patients. Standard actuarial statistical techniques have been used widely but do not provide the most appropriate information for patient populations experiencing competing hazards. "Actual," or cumulative incidence, methods may provide a better estimate of the durability of tissue valves for these patients. The purpose of this article is to compare actuarial and actual estimates of the durability of the Hancock Modified Orifice bioprosthesis aortic valve in a multi-institutional study. Methods and Results-Values were implanted between 1976 and 1985 in 3 institutions. This sample contains follow-up data on 727 patients (42% female) with a mean age of 63+/-13 years. The difference between actuarial and actual rates of SVD became more important over time. At 5 years, the difference is significant only in the elderly (aged 165 and 170), whereas at 17 years, the difference was significant for all patients. Similarly, the magnitude of the difference increases over time. Freedom from SVD for patients >65 at 5 years is estimated at 98+/-0.01% by actuarial methods and 100+/-0.00% actual methods. The difference between estimates is larger at 10 years, 93+/-0.02% versus 96+/-0.01%. This difference is greater at 17 years, 78+/-0.04% versus 93+/-0.01%. Conclusions-The results of this study demonstrate that the particular statistical analysis method used to calculate SVD can provide strikingly different conclusions. These observations indicate that the actual method is able to identify the lower risk of SVD in older patients. Generalizability issues must be considered, however, when using actual SVD rates to make decisions regarding valve selection in larger populations. Conclusions-The results of this study demonstrate that the particular statistical analysis method used to calculate SVD can provide strikingly different conclusions. These observations indicate that the actual method is able to identify the lower risk of SVD in older patients. Generalizability issues must be considered, however, when using actual SVD rates to make decisions regarding valve selection in larger populations.
