To the Editor We thank Prof Andrew Vickers, PhD, for his Comment1 on our article2 highlighting a problem with the methods used in our model to assess the benefit-to-harm ratio and the cost-effectiveness of risk-stratified breast screening programs compared with a standard age-based screening program and no screening. Our study was published online first on July 5, 2018, and in the November 2018 issue of JAMA Oncology, and the Comment by Prof Vickers now appears online with our article.2 In response to earlier communication from Vickers, we reviewed the statistical code for the base case model, and in doing so, we identified an algebraic error in the calculation of the number of overdiagnosed cases. The calculation inadvertently generated an estimate of overdiagnosis that varied by risk, for which there was no supporting evidence, and which is not compatible with our stated assumption that a fixed proportion of the screen-detected cases are overdiagnosed whatever the prior risk. Consequently, the number of overdiagnosed cases (as shown in Figure 1 of our article2) is smaller than expected. This explains the observation of Vickers that overdiagnosis is substantially reduced when screening is restricted to a small proportion of the population, and the predictiveness of the risk stratifier is limited. This error does not affect any of the other outputs of the model, including the cost-effectiveness analysis.
Pashayan N, Pharoah PDP. Concerns About Methods Used in Modeling Study of Risk-Stratified Screening for Breast Cancer. JAMA Oncol. 2022;8(3):483–484. doi:10.1001/jamaoncol.2021.7311
Artificial Intelligence Resource Center
Customize your JAMA Network experience by selecting one or more topics from the list below.