Every few days, we will be publishing layman’s abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
The article featured today is from Statistics in Medicine and the full article, published in issue 38.22, is available to read online here.
Steyerberg, EW, Nieboer, D, Debray, TPA, van Houwelingen, HC. Assessment of heterogeneity in an individual participant data meta‐analysis of prediction models: An overview and illustration. Statistics in Medicine. 2019; 38: 4290– 4309. doi: 10.1002/sim.8296
Clinical prediction models aim to provide estimates of absolute risk for a diagnostic or prognostic endpoint. The development of such models requires a sufficient sample size. One way to increase sample size is to combine multiple datasets. Such datasets may differ in various aspects, introducing heterogeneity in the analysis. To what extent can we then develop a global model with broad validity and applicability?
We consider different approaches to assess heterogeneity in predictions of absolute risk. We first need to examine differences between settings in terms of study design and the included patients. Next, statistical methods are presented that allow for the development of a global model while including heterogeneity between settings. These models aim to estimate a model with local adjustments, while still borrowing strength across settings. Different visualization methods are described to assess the impact of the heterogeneity between settings on estimated risk for individual patients and the transportability of the developed model. These methods are illustrated in a case study with patients suffering from traumatic brain injury, where we aim to predict 6-month mortality based on individual patient data using meta-analytic techniques (15 studies, n=11,022 patients).