gms | German Medical Science

64. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie

08. - 11.09.2019, Dortmund

Unblinded sample size re-estimation for diagnostic accuracy studies

Meeting Abstract

Suche in Medline nach

  • Antonia Zapf - Department of Medical Biometry, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
  • Annika Hoyer - Deutsches Diabetes-Zentrum, Düsseldorf, Germany

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. 64. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e.V. (GMDS). Dortmund, 08.-11.09.2019. Düsseldorf: German Medical Science GMS Publishing House; 2019. DocAbstr. 270

doi: 10.3205/19gmds004, urn:nbn:de:0183-19gmds0048

Veröffentlicht: 6. September 2019

© 2019 Zapf et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe



In diagnostic accuracy studies, sensitivity and specificity are recommended as co-primary endpoints [1]. For sample size calculation, assumptions about the expected sensitivity and specificity of the index test as well as the minimal acceptable diagnostic accuracy (for the comparison with the gold standard) or the expected diagnostic accuracy of the comparator test have to be made.

However, the assumptions are often overoptimistic, which is, for example, due to case-control designs with severe cases and healthy controls or to a data-driven selection of the cutoff value in the previous study [2], [3]. Such overoptimistic assumptions can be found in the study of Yan et al. [4], where the estimated sensitivity was 75.8%, whereas the authors expected 91%.4 On the other hand, it is also possible that the diagnostic accuracy is underestimated, as in the study of Moura et al. [5], where the expected sensitivities of the two diagnostic tests under evaluation were 75% and 49%, whereas the estimated values were 93.8% and 60.4% at the end. Moreover, in some cases no reliable previous knowledge is available, leading to vague sample size calculations as it can be seen in the study of Iglesias et al. [6], where the authors performed sample size calculations for the worst and the most favourable scenario and chose a scenario between them.

Therefore, it is essential to develop methods for a sample size re-estimation in diagnostic accuracy trials. While such adaptive designs are standard in interventional trials, in diagnostic studies they are less common [7]. Some approaches from interventional trials cannot be applied to diagnostic accuracy studies or have to be modified. For example, blinded sample size re-estimation based on the variances are not applicable, as the variances are based on the point estimators. Furthermore, the specific feature of diagnostic accuracy trials are the two co-primary endpoints sensitivity and specificity, which are estimated based on independent subgroups.

In this talk we present an approach for an unblinded sample size re-estimation in diagnostic accuracy studies which is illustrated by results of a simulation study as well as by example studies.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.


EMA. Guideline on clinical evaluation of diagnostic agents. 23 July 2009 [cited 15 July 2019]. Available from: document _library/ Scientific_guideline/2009/09/WC500003580.pdf Externer Link
Rutjes A. Evidence of bias and variation in diagnostic accuracy studies. Canadian Medical Association Journal. 2006;174(4):469-476.
Leeflang M, Moons K, Reitsma J, Zwinderman A. Bias in Sensitivity and Specificity Caused by Data-Driven Selection of Optimal Cutoff Values: Mechanisms, Magnitude, and Solutions. Clinical Chemistry. 2008;54(4):729-737.
Yan L, Tang S, Yang Y, Shi X, Ge Y, Sun W, et al. A Large Cohort Study on the Clinical Value of Simultaneous Amplification and Testing for the Diagnosis of Pulmonary Tuberculosis. Medicine. 2016;95(4):e2597.
Moura D, de Moura E, Matuguma S, dos Santos M, Moura E, Baracat F, et al. EUS-FNA versus ERCP for tissue diagnosis of suspect malignant biliary strictures: a prospective comparative study. Endoscopy International Open. 2018;06(06):E769-E777.
Iglesias K, Sporkert F, Daeppen J, Gmel G, Baggio S. Comparison of self-reported measures of alcohol-related dependence among young Swiss men: a study protocol for a cross-sectional controlled sample. BMJ Open. 2018;8(7):e023632.
Zapf A, Stark M, Gerke O, Ehret C, Benda N, Bossuyt P, Deeks J, Reitsma J, Alonzo T, Friede T. Adaptive trial designs in diagnostic accuracy research. Submitted.