On calculating the power of a TDT study--comparison of methods

We show that the methods for calculating the power of the TDT of Risch & Merikangas (1996), McGinnis (1998, 2000), Tu & Whittemore (1999) and Knapp (1999) consist of three distinct approaches. The differences between the methods arise either by treating parental transmissions independently or not, and either conditioning on parental heterozygosity or not. Transmissions are only truly independent when the mode of inheritance of the disease is multiplicative, making this assumption invalid and the methods that assume them (Risch & Merikangas, 1996; McGinnis, 1998, 2000) inappropriate. We demonstrate a basic model that allows analytical comparisons of the methods, and suggest a new method that calculates power at least as accurately as any of the three previously published.