Analysis of Differences in Proportions from Clustered Data with Multiple Measurements in Diagnostic Studies

Summary

Objectives: In diagnostic studies, proportions such as sensitivities are often to be calculated and to be compared between different diagnostic procedures. As statistical unit of analysis, multiple observational units may be assessed within each patient, i.e., multiple lesions in an organ. As a requirement, these are to be assessed by multiple blinded readers. In this paper we propose a method to cover correlations between units within patients, correlations between procedures and correlations between different raters assessing each observational unit.

Methods: The proposed approach is a two-step method to analyze clustered data with multiple measurements to compare diagnostic procedures in a paired modality design and the correlation between the readers in a paired reader design. The performance of the approach was compared to a generalized estimation equations model (GEEs) by power simulations.

Results: Power simulations suggest, that the two-step approach is not inferior to GEEs with regard to the single readers as well as with regard to the average reader.

Conclusions: An intuitive approach was developed next to established methods to analyze “paired modality, paired reader” and “unpaired modality, paired reader” studies with binary endpoints when estimating proportions and differences in proportions for clustered data with multiple measurements. This two-step approach is an alternative method to cover routine designs of diagnostic studies where the difference of proportions is to be estimated directly along with confidence intervals.

Keywords

Clustered data - diagnostic study - proportions - binary data

• References

• 1 FDA. Developing Medical Imaging Drugs and Biological Products. Part 3: Design, Analysis, and Interpretation of Clinical Studies.. Food and Drug Administration (FDA); Guideline 2004
• 2 CPMP. Points to consider on the evaluation of diagnostic agents, Committee for Proprietary Medicinal Products (CPMP) of the European Agency for the Evaluation of Medicinal Products (EMEA), CPMP/EWP/1119/98,2001.
• 3 Rao JNK, Scott AJ. A simple method for the analysis of clustered binary data. Biometrics 1992; 48: 577-85.
• 4 Obuchowski N. On the Comparison of Correlated Proportions for Clustered Data. Statistics in Medicine 1998; 17: 1495-1507.
• 5 Schwenke C, Kaufmann J. Analysis of blinded reader studies: Correlated proportions for clustered data with repeated measurements. Proceedings of the American Statistical Association, Biometrics section (CD-ROM).. Alexandria, VA: American Statistical Association; 2003
• 6 Ahn C. An evaluation of methods for the estimation of sensitivity and specificity of site-specific diagnostic tests. Biometrical Journal 1997; 39: 793-807.
• 7 Jung S-H, Ahn C. Estimation of Response Probability in Correlated Binary Data: A New Approach. Drug Information Journal 2000; 34: 599-604.
• 8 May WL, Johnson WD. Hypothesis Testing and Confidence Interval Construction in 2 χ 2 Tables of Correlated Proportions. Journal of Biopharmaceutical Statistics 1998; 8: 115-30.
• 9 Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986; 73: 13-22.
• 10 Kaufmann J, Werner C, Brunner E. Nonparametric methods for diagnostic test performance. Statistical Methods in Medical Research 2005; 14: 1-18.
• 11 Schwenke C, Nevalainen J. Analysis of proportions from clustered data with missing observations in amatched-pair design. Methods Inf Med 2004; 43: 521-4.