Using meta-analysis to inform the design of subsequent studies of diagnostic test accuracy

Different evidence summaries have implications for contextualizing findings of meta-analysis of diagnostic tests

OBJECTIVE

To evaluate diagnostic tests, analysts use meta-analyses to provide inputs to parameters in decision models. Choosing parameter estimands from meta-analyses requires understanding the meta-analytic and decision-making contexts.

STUDY DESIGN AND SETTING

We expand on an analysis comparing positron emission tomography (PET), PET with computed tomography (PET/CT), and conventional workup (CW) in women with suspected recurrent breast cancer. We discuss Bayesian meta-analytic summaries (posterior mean over a set of existing studies, posterior estimate in an existing study, posterior predictive mean in a new study) used to estimate diagnostic test parameters (prevalence, sensitivity, specificity) needed to calculate quality-adjusted life years in a decision model contextualizing PET, PET/CT, and CW.

RESULTS

The mean and predictive mean give similar estimates, but the latter displays greater uncertainty. Namely, PET/CT outperforms CW on average but may not do better than CW when implemented in future settings.

CONCLUSION

Selecting estimands for decision model parameters from meta-analyses requires understanding the relationship between decision settings and meta-analysis studies' settings, specifically whether the former resemble one or all study settings or represents new settings. We provide an algorithm recommending appropriate estimands as input parameters in decision models for diagnostic tests to obtain output parameters consistent with the decision context.

Parametric multistate survival models: Flexible modelling allowing transition-specific distributions with application to estimating clinically useful measures of effect differences

Multistate models are increasingly being used to model complex disease profiles. By modelling transitions between disease states, accounting for competing events at each transition, we can gain a much richer understanding of patient trajectories and how risk factors impact over the entire disease pathway. In this article, we concentrate on parametric multistate models, both Markov and semi-Markov, and develop a flexible framework where each transition can be specified by a variety of parametric models including exponential, Weibull, Gompertz, Royston-Parmar proportional hazards models or log-logistic, log-normal, generalised gamma accelerated failure time models, possibly sharing parameters across transitions. We also extend the framework to allow time-dependent effects. We then use an efficient and generalisable simulation method to calculate transition probabilities from any fitted multistate model, and show how it facilitates the simple calculation of clinically useful measures, such as expected length of stay in each state, and differences and ratios of proportion within each state as a function of time, for specific covariate patterns. We illustrate our methods using a dataset of patients with primary breast cancer. User-friendly Stata software is provided.