Estimation of Survivorship in Chronic Disease: The “Actuarial” Method

Abstract In most follow-up studies in chronic disease assumptions on the parametric form of the survivorship function are not justified. A simple frequency estimate, arising from polygonal approximation, is discussed. It is shown that, for proper limitation of the period over which the survivorship is to be estimated, bias due to nonlinearity may be controlled, without loss in efficiency, by proper choice of the sub-intervals. It is concluded that the method proposed is appropriate and highly efficient for large scale follow-up studies such as those carried out by the Mayo Clinic.