Nonparametric Sample Size Estimation for Sensitivity and Specificity with Multiple Observations per Subject
- PDF / 9,594,783 Bytes
- 8 Pages / 612 x 792 pts (letter) Page_size
- 92 Downloads / 150 Views
Nonparametric Sample Size Estimation for Sensitivity and Specificity With Multiple Observations Per Subject
Fan Hu Department of Statistical Science, Southern Methodist University. Dallas, Texas William R. Schucany Department of Statistical Science, Southern Methodist University. Dallas, Texas Chul Ahn Department of Clinical Sciences, UT Southwestern Medical Center. Dallas, Texas
Key Words lntracluster correlation; Diagnostic test; Empirical power Correspondence Address Chul Ahn. PhD, Department of Clinical Sciences, UT Southwestern Medical Center, 5 3 2 3 Hany Hines Blvd. E 5 . 5 0 6 Dallas, TX 7 5 3 9 0 (email: Chul.Ahn@ UTSouthwestern.edu).
We propare a samplesize calculation approach for the estimation of sensitivity and speciicity of diagnostic tests with multiple observations per subject. Many diagnostic tests such as diagnostic imaging or periodontal tests are characterized by the presence of multiple observations for each subject. h e number of observations F.equently vm'es among subjects in diagnosticimaging experiments or pniodontal studies. Nonparametric statistical methods for the analysis of clustered binary data have been recently devdoped by various authors. In this article, we derive a sample sue formula for sensitivityand specificityof diagnostictests using the sign test while accounting for multiple observations per subject. Application of the
INTRODUCTION Diagnostic tests are of particular importance in medicine since early and accurate diagnosis can decrease morbidity and mortality rates of disease. Some examples of diagnostic tests include positron emission tomography (PET) scans, x-rays, and enzymatic diagnostic tests. The performance of a diagnostic test is often summarized by its sensitivity and specificity. Sensitivity is defined as the probability of a positive diagnostic test in a subject with the disease, and specificity as the probability of a negative diagnostic test in a subject without the disease. In this article we focus on clustered binary observations, which are made from multiple observations on each subject. In this case, observations from each subject are correlated although those from different subjects are independent. For example, in a radiologic study each subject may contribute multiple lesions to the study and an observation is made from each lesion.
samplesizeformula for the design of a diagnostic test is discussed. Since the sample size formula is based on large sample theory, simulation studies are conducted to evaluatethefinite sample perjiormance of the proposed method. We compare the perjiormance of the proposed samplesizeformula with that of the parametric sample size formula that assigns equal weight to each observation. Simulation studies show that the proposed samplesizeformula generally yields empirical powers closer to the nominal level than the parametric method. Simulation studies also show that the number of subjects required increases as the variability in the number of observations per subject increases and the intracluster correlation increases.
F
Data Loading...
