Methods for diagnostic studies
An effective therapy requires a correct diagnosis. For this you need diagnostic tests which are evaluated in so-called diagnostic studies. The development of statistical methods for planning and analyzing diagnostic studies and diagnostic meta-analysis was neglected a long time. Therefore, we aim to close the still existing methodological gaps. Currently we are working on the development of flexible designs for diagnostic studies within the DFG-project entitled
Flexible designs for diagnostic studies- from goodness of diagnosis to personalized medicine
Within the project methods for adaptive designs are developed for different research questions in diagnostic studies. For phase 3 trials there are 2 co-primary endpoints (sensitivity and specificity). In phase 4 trials a patient-relevant endpoint is chosen. It is differentiated between blinded and unblinded interim analysis.
The project is divided into the following sub-projects:
- Adaptive designs for blinded interim analysis in phase 3 diagnostic studies
- Adaptive designs for unblinded interim analysis in phase 3 diagnostic studies
- Adaptive designs for blinded interim analysis in phase 4 diagnostic studies
- Adaptive designs for unblinded interim analysis in phase 4 diagnostic studies
- Adaptive seamless designs
In each sub-project comparative studies and non-comparative studies will be considered. Comparative studies are studies where an experimental test is compared to a standard test whereas in a non-comparative study a standard test is not given.
The methods will be implemented in R and/or SAS.
- Todd Alonzo (University of Southern California, USA)
- Norbert Benda (BfArM, Deutschland)
- Christoph Berding (Roche Diagnostics, Deutschland)
- Patrick Bossuyt (University of Amsterdam, Niederlande)
- Jon Deeks (University of Birmingham, UK)
- Tim Friede (Universitätsmedizin Göttingen, Deutschland)
- Oke Gerke (Odense University Hospital, Dänemark)
- Hans Reitsma (University Medical Center Utrecht & University Utrecht, Niederlande)
- Werner Vach (Universitätsspital Basel, Schweiz)
Adaptive trial designs in diagnostic accuracy research. Zapf A, Stark M, Gerke O, Ehret C, Benda N, Bossuyt P, Deeks J, Reitsma J, Alonzo T, Friede T. Stat Med. 2020 Feb 28;39(5):591-601.
Sample size calculation and re-estimation based on the prevalence in a single-arm confirmatory diagnostic accuracy study. Stark M, Zapf A. Stat Methods Med Res. 2020 Oct;29(10):2958-2971.
There is a potential for seamless designs in diagnostic research. Vach W, Bibiza E, Gerke O, Bossuyt PM, Friede T, Zapf A. J Clin Epi, 2020, accepted.
Appraising Heterogeneity. Zapf A. 2018. Diagnostic Meta-Analysis . Biondi-Zoccai G (Hrsg.). 1. Aufl. Berlin: Springer International Publishing, 125-160.
Partial verification bias and incorporation bias affected accuracy estimates of diagnostic studies for biomarkers that were part of an existing composite gold standard. Karch A, Koch A, Zapf A, Zerr I, Karch A. J CLIN EPIDEMIOL. 2016;78:73-82.
A wild bootstrap approach for the selection of biomarkers in early diagnostic trials. Zapf A, Brunner E, Konietschke F. BMC MED RES METHODOL. 2015;15:43.
Nonparametric meta-analysis for diagnostic accuracy studies. Zapf A, Hoyer A, Kramer K, Kuss O. STAT MED. 2015;34(29):3831-3841.
A modified Wald interval for the area under the ROC curve (AUC) in diagnostic case-control studies. Kottas M, Kuss O, Zapf A. BMC MED RES METHODOL. 2014;14:26.
Difference of two dependent sensitivities and specificities: Comparison of various approaches. Wenzel D, Zapf A. BIOMETRICAL J. 2013;55(5):705-718.