Article
Adaptive designs with unblinded sample size re-estimation for diagnostic accuracy studies
Search Medline for
Authors
-
Denise Köster
- Department of Medical Biometry, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
-
Annika Hoyer
- Ludwig-Maximilians-Universität München, München, Germany
-
Antonia Zapf
- Department of Medical Biometry, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
| Published: | September 24, 2021 |
|---|
Outline
Text
Diagnostic studies aim to estimate the proportion of correctly classified diseased and non-diseased individuals, namely sensitivity and specificity. To evaluate a diagnostic test, its effectiveness is compared to either a reference standard, which defines the true disease status or another test. If there is no standard test as a comparator, the aim is to demonstrate a pre-specified minimum sensitivity and specificity of the experimental test. If there is a standard test, both tests and the reference standard should be applied to all study participants if it is ethically acceptable and feasible [1]. Otherwise, the two tests should be randomly assigned and the reference standard should be used in all. This results in three study designs: the single-test, and the paired or unpaired two-tests design.
As in any clinical trial, a sample size calculation should be performed to avoid including too many or too few individuals. However, assumptions for the sample size planning are often uncertain. Reasons for biased estimates in diagnostic accuracy studies are analysed in Rutjes et al. [2]. Long-established methods for adaptive designs with sample size re-estimation from the field of interventional studies [3] have not yet been transferred to diagnostic studies with the particularity of co-primary endpoints.
In this talk, an approach for the unblinded sample size re-estimation, which is applicable for all three designs, will be presented and illustrated by an example study. Furthermore, the results of a comprehensive simulation study will be presented, and the potential and limitations of the method will be discussed.
The authors declare that they have no competing interests.
The authors declare that an ethics committee vote is not required.
This contribution has already been published: submitted for GMDS/IBS-CEN 2020, lecture was waived [4].
References
- 1.
- EMA. Guideline on clinical evaluation of diagnostic agents. Doc. Ref. CPMP/ EWP/1119/98/Rev.1. 2010 [date of last access 05/05/21]. Available from: http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003580.pdf
- 2.
- Rutjes AW, Reitsma JB, Di Nisio M, Smidt N, van Rijn JC, Bossuyt PM. Evidence of bias and variation in diagnostic accuracy studies. CMAJ. 2006;174(4):469-76.
- 3.
- Wassmer G, Brannath W. Group sequential and confirmatory adaptive designs in clinical trials. Heidelberg: Springer; 2016.
- 4.
- Köster D, Hoyer A, Zapf A. Adaptive designs with unblinded sample size re-estimation for diagnostic accuracy studies. In: 65th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), Meeting of the Central European Network (CEN: German Region, Austro-Swiss Region and Polish Region) of the International Biometric Society (IBS). Berlin, 06.-09.09.2020. Düsseldorf: German Medical Science GMS Publishing House; 2021. DocAbstr. 419. DOI: 10.3205/20gmds326
