Accuracy of Meta-Analysis Using Different Levels of Diagnostic Accuracy Measures
Diagnostic studies report results in sensitivity and specificity, or figures of receiver operation characteristics (ROC) curve. Meta-analysis synthesizes these diagnostic accuracy measures from different studies to obtain an overall summary ROC curve. Inc
- PDF / 380,328 Bytes
- 13 Pages / 439.36 x 666.15 pts Page_size
- 95 Downloads / 183 Views
Abstract Diagnostic studies report results in sensitivity and specificity, or figures of receiver operation characteristics (ROC) curve. Meta-analysis synthesizes these diagnostic accuracy measures from different studies to obtain an overall summary ROC curve. Increasingly, meta analysis also uses individual patient level data. However, the pro and con of such an approach are not entirely clear. In this paper, we performed a simulation study to evaluate the accuracy of summary ROC curves derived from different types of data, i.e., the paired sensitivity and specificity from individual study, the study-specific ROC curves, and the individual patient level data. Extensive simulation experiments were conducted under various settings to compare the empirical performance of estimated summary ROC curves using data from three levels. The simulation results demonstrated that the method based on reported ROC curves from individual study provides accurate and robust summary ROC curve compared with alternatives including those based on patient level data and is preferred in practice.
Y. Song Department of Biomedical Data Science, Stanford University School of Medicine, HRP Redwood Building, Stanford, CA 94305-5405, USA Department of Pharmacology and Biostatistics, Institute of Medical Sciences, Shanghai Jiao Tong University, Shanghai, China Y. Lu Department of Biomedical Data Science, Stanford University School of Medicine, HRP Redwood Building, Stanford, CA 94305-5405, USA VA Palo Alto Health Care System, Palo Alto, CA, USA L. Tian () Department of Biomedical Data Science, Stanford University School of Medicine, HRP Redwood Building, Stanford, CA 94305-5405, USA Department of Statistics, Stanford University, Stanford, CA, USA e-mail: [email protected] © Springer International Publishing Switzerland 2016 J. Lin et al. (eds.), Statistical Applications from Clinical Trials and Personalized Medicine to Finance and Business Analytics, ICSA Book Series in Statistics, DOI 10.1007/978-3-319-42568-9_3
21
22
Y. Song et al.
Keywords Sensitivity • Specificity • Receiver operating characteristics (ROC) curve • Meta-analysis • Bivariate normal random effects model • Simulation experiment
1 Introduction Diagnostic and predictive tests are important components of medical care. Clinicians rely on test results to establish diagnosis and guide patient management (Abbas et al. 2007). For example, diagnostic tests using either longitudinal Magnetic Resonance Imaging (MRI) or X-ray scans of hands can measure the total joint erosion and its change overtime to monitor the disease progression of rheumatoid arthritis (RA). A clinical decision on whether to change treatments often depends on the MRI or X-ray test results on RA progression status. MRI, a 3-dimensional technology that provides a better view of bone structure than a 2-dimensional Xray, is anticipated to be able to provide an earlier and more sensitive prediction of disease progression. Comparing results of a diagnostic test with a reference standard (in our RA example, clinical
Data Loading...
