skip navigation

Key Publications in Biostatistics

Study Examples | Key Publications | Intro

A few key publications follow:

Parts of computed tomographic (CT) images are supposed to be darker or lighter based entirely on the density and atomic number of the material being examined. However, these image attenuation values also vary by scanner and by body size. Therefore, scores for coronary calcium may not necessarily represent how much calcium is present. These researchers used calibration phantoms to adjust attenuation values and standardize calcium measurements, and they recommend that this be done in research studies and clinical practice.

—Nelson JC, Kronmal RA, Carr JJ, McNitt-Gray MF, Wong ND, Loria C, Goldin JG, Williams OD, Detrano R (2005). Measuring coronary calcium on CT images adjusted for attenuation differences. Radiology, 235: 403–414.

Diagnostic tests such as mammography are often evaluated by comparing their outcome to the patient's true disease state; but two complications may interfere: verification bias (the probability of detecting disease may depend on both the test result and the subject’s characteristics); and intrareader correlation (tests interpreted by the same rater may not be independent). This paper suggests that in the context of an observational cohort study where rich covariate information is available, a weighted estimating equations approach may be the best way to address these complications.

—Zheng YY, Barlow WE, Cutter G (2005). Assessing accuracy of mammography in the presence of verification bias and intrareader correlation. Biometrics, 61: 259–268.

This article proposes and compares two approaches for regression analysis of multilevel binary data when clusters are not necessarily nested. The researchers illustrated the methods using data from the Breast Cancer Surveillance Consortium to estimate mammography accuracy from a repeatedly screened population.

—Miglioretti DL, Heagerty PJ (2004). Marginal modeling of multilevel binary data with time varying covariates. Biostatistics, 5 (3): 381–398.

The researchers developed a Bayesian method to compare weekly depression states recalled for a three-month period (longitudinal data) with measurements of current depression assessed during randomly timed phone interviews (cross-sectional data). They examined the accuracy of recalled depression by linking a spline model for recalled depression and a logistic model for current depression. Their proposed model can be modified to examine general relationships between longitudinal and cross-sectional measurements.

—Rutter CM, Simon G (2004). A Bayesian method for estimating the accuracy of recalled depression. Journal of the Royal Statistical Society Series C, 53 (2): 341–353.

This article presents a method for estimating the accuracy of psychological screening scales using receiver operating characteristic curves and associated statistics. The distributions of screening scales typically are nearly symmetric when the target condition is present and highly skewed when it is absent; the researchers model screening scale outcomes using truncated normal distributions that accommodate these different distributional shapes. They use this model to estimate the accuracy of the Symptom Checklist as a measure of major depression.

—Rutter CM, Miglioretti DL (2003). Estimating the accuracy of psychological scales using longitudinal data. Biostatistics, 4 (1): 97–107.

Study Examples | Key Publications | Intro

Top