What do we know? Limitations of the two methods most commonly used to estimate the length of the prospective wait

Abstract

Health service researchers, policy analysts and other commentators have overlooked the limitations of existing approaches to the estimation of waiting times. If urgent cases are given priority, there are no instances when census-based data can supply accurate estimates of the length of the prospective wait. But there are three occasions when event-based data supply accurate estimates of the prospective wait of those who chose to enrol and we can predict the direction of error when the relevant conditions are violated if we know whether the list was open or closed, and whether it grew in size or shrank. Without this additional information, we cannot determine whether the changes we observe over time or the differences we see between one list and the next are spurious or not. The period life-table provides a timely and bias-free alternative to the existing cross-sectional approaches for a modest increase in the complexity of calculation.