Dr. Moskowitz's primary research interests involve the development of statistical methodology for evaluating diagnostic tests and biomarkers. Within this field her work is currently concentrated in two general areas. The first focuses on developing meaningful ways of quantifying and comparing the predictive accuracy of multiple tests. The methodology that she has developed and continues to work on can be used to assess instruments such as biomarkers, risk prediction models, and radiographic imaging tools in their ability to detect disease and predict future outcomes. The second area on which she is currently focused in collaboration with colleagues at MSKCC involves developing methodology for evaluating diagnostic tests when each subject contributes multiple observations per test.
Dr. Moskowitz also actively collaborates with colleagues in the Department of Radiology and the Epidemiology Service. She has worked on the design and analysis of studies assessing various imaging modalities and potential biomarkers.
Selected Bibliography:
1. Alonzo TA, Pepe MS, and Moskowitz CS. Sample size calculations for comparative studies of medical tests for detecting presence of disease. Statistics in Medicine 2001;21:835-852.
2. Alonzo TA, Braun TM, and Moskowitz CS. Small sample estimation of relative accuracy for binary tests. Statistics in Medicine 2004;23:21-34.
3. Moskowitz CS, Pepe MS. Quantifying and comparing the predictive accuracy of continuous prognostic factors for binary outcomes. Biostatistics 2004;5:113-127.
4. Moskowitz CS, Pepe MS. Quantifying and comparing the accuracy of binary biomarkers when predicting a failure time outcome. Statistics in Medicine 2004;23:1555-1570.
5. Cai T and Moskowitz CS. Semiparametric estimation of the binormal ROC curve. Biostatistics 2004;5: 573-586.
