Danielle Enserro

Background/Objectives: Proper confidence interval estimation of the area under the receiver operating characteristic curve (AUC), the net reclassification index (NRI), and the integrated discrimination improvement (IDI) is an area of ongoing research. The most common confidence interval estimation methods employ asymptotic theory. However, developments demonstrate that degeneration of the normal distribution assumption under the null hypothesis exists for measures such as the change in AUC (ΔAUC) and IDI, and confidence intervals estimated under the normal distribution assumption may be invalid. We aim to study the performance of confidence intervals derived assuming asymptotic theory and those derived with non-parametric bootstrapping methods. Methods: We examine the performance of ΔAUC, NRI, and IDI in both the logistic and survival regression context. We explore empirical distributions and compare coverage probabilities of asymptotic confidence intervals with those produced from bootstrapping methods through simulation. Results: The primary finding in both the logistic framework and the survival analysis framework is that the percentile CIs performed well regarding coverage, without compromise to their width; this finding was robust in most scenarios. Conclusions: Our results suggest that the asymptotic intervals are only appropriate when a strong effect size of the added parameter exists, and that the percentile bootstrap interval exhibits at least a reasonable coverage while maintaining the shortest width in nearly all simulated scenarios, making this interval the most reliable choice. The intent is that these recommendations improve the accuracy in the estimation and the overall assessment of discrimination improvement.

Examining treatment effects in subgroups of patients defined by demographic, genetic, or clinical characteristics is increasingly of interest given the pursuit of personalized medicine and the importance of representation and equity in treatment decisions. The magnitude or even the direction of the treatment effect may vary across subgroups, and these differential treatment effects could have clinical implications. Subgroup analyses require caution in their interpretation, however, because of the high probability of a false-positive or false-negative conclusion. We outline study design and analysis considerations for responsibly investigating and reporting differential treatment effects across subgroups in oncology trials, with examples from the National Cancer Institute’s National Clinical Trials Network and Community Oncology Research Program. Recommendations include ensuring appropriate representation of patients from subgroups of interest, recognizing power and multiplicity limitations, and treating exploratory subgroup analyses as hypothesis generating rather than practice changing.

Prevention trials in oncology are some of the most important cancer clinical trials that can be designed, implemented, analyzed, and interpreted. They are pivotal in the goal of stopping the development of cancer before it starts. Prevention trials are unique in that they not only have some of the same requirements and challenges as treatment trials but also have challenges that may make their design more complex. This paper aims to discuss some types of prevention trials and highlight their most common design challenges, including large sample size requirements, slow accrual rates with long accrual duration, extended follow-up periods with trial adherence issues and missing data, endpoints that require extended follow-up or have a high level of confounding, and problems with optimizing study design. This article provides real study examples and suggestions for designing prevention clinical trials while mitigating the known issues they face.

Background: The remaining lifetime risk (RLR) is the probability of developing an outcome over the remainder of one’s lifespan at any given age. The RLR for atherosclerotic cardiovascular disease (ASCVD) in three 20-year periods were assessed using data from a single community-based cohort study of predominantly White participants. Methods: Longitudinal data from the Framingham study in 3 epochs (epoch 1, 1960–1979; epoch 2, 1980–1999; epoch 3, 2000–2018) were evaluated. The RLR of a first ASCVD event (myocardial infarction, coronary heart disease death, or stroke) from 45 years of age (adjusting for competing risk of death) in the 3 epochs were compared overall, and according to the following strata: sex, body mass index, blood pressure and cholesterol categories, diabetes, smoking, and Framingham risk score groups. Results: There were 317 849 person-years of observations during the 3 epochs (56% women; 94% White) and 4855 deaths occurred. Life expectancy rose by 10.1 years (men) to 11.9 years (women) across the 3 epochs. There were 1085 ASCVD events over the course of 91 330 person-years in epoch 1, 1330 ASCVD events over the course of 107 450 person-years in epoch 2, and 775 ASCVD events over the course of 119 069 person-years in epoch 3. The mean age at onset of first ASCVD event was greater in the third epoch by 8.1 years (men) to 10.3 years (women) compared with the first epoch. The RLR of ASCVD from 45 years of age declined from 43.7% in epoch 1 to 28.1% in epoch 3 ( P P P P <0.001). Conclusions: Over the past 6 decades, mean life expectancy increased and the RLR of ASCVD decreased in the community-based, predominantly White Framingham study. The residual burden of ASCVD underscores the importance of continued and effective primary prevention efforts with better screening for risk factors and their effective treatment.