Generalized linear mixed models for multi-reader multi-case studies of diagnostic tests.

Diagnostic tests are often compared in multi-reader multi-case (MRMC) studies in which a number of cases (subjects with or without the disease in question) are examined by several readers using all tests to be compared. One of the commonly used methods for analyzing MRMC data is the Obuchowski-Rockette (OR) method, which assumes that the true area under the receiver operating characteristic curve (AUC) for each combination of reader and test follows a linear mixed model with fixed effects for test and random effects for reader and the reader-test interaction. This article proposes generalized linear mixed models which generalize the OR model by incorporating a range-appropriate link function that constrains the true AUCs to the unit interval. The proposed models can be estimated by maximizing a pseudo-likelihood based on the approximate normality of AUC estimates. A Monte Carlo expectation-maximization algorithm can be used to maximize the pseudo-likelihood, and a non-parametric bootstrap procedure can be used for inference. The proposed method is evaluated in a simulation study and applied to an MRMC study of breast cancer detection.

Can sequential parallel comparison design and two-way enriched design be useful in medical device clinical trials?

Sequential parallel comparison design (SPCD) was proposed for trials with high placebo response. In the first stage of SPCD subjects are randomized between placebo and active treatment. In the second stage placebo nonresponders are re-randomized between placebo and active treatment. Data from the population of "all comers" and the subpopulations of placebo nonresponders then combined to yield a single p-value for treatment comparison. Two-way enriched design (TED) is an extension of SPCD where active treatment responders are also re-randomized between placebo and active treatment in Stage 2. This article investigates the potential uses of SPCD and TED in medical device trials.

A causal model for joint evaluation of placebo and treatment-specific effects in clinical trials.

Evaluation of medical treatments is frequently complicated by the presence of substantial placebo effects, especially on relatively subjective endpoints, and the standard solution to this problem is a randomized, double-blinded, placebo-controlled clinical trial. However, effective blinding does not guarantee that all patients have the same belief or mentality about which treatment they have received (or treatmentality, for brevity), making it difficult to interpret the usual intent-to-treat effect as a causal effect. We discuss the causal relationships among treatment, treatmentality and the clinical outcome of interest, and propose a causal model for joint evaluation of placebo and treatment-specific effects. The model highlights the importance of measuring and incorporating patient treatmentality and suggests that each treatment group should be considered a separate observational study with a patient's treatmentality playing the role of an uncontrolled exposure. This perspective allows us to adapt existing methods for dealing with confounding to joint estimation of placebo and treatment-specific effects using measured treatmentality data, commonly known as blinding assessment data. We first apply this approach to the most common type of blinding assessment data, which is categorical, and illustrate the methods using an example from asthma. We then propose that blinding assessment data can be collected as a continuous variable, specifically when a patient's treatmentality is measured as a subjective probability, and describe analytic methods for that case.

Inference and sample size calculation in the fit assessment of filtering facepiece respirators.

Filtering facepiece respirators have recently been cleared by the U.S. Food and Drug Administration (FDA) for use by the general public in public health medical emergencies such as pandemic influenza. In the fit assessment of these devices it is important to distinguish between the two sources of variability: population heterogeneity and random fluctuations over repeated donnings. The FDA Special Controls Guidance Document (SCGD) which describes these devices and their evaluation, recommends that the fit performance of a filtering facepiece respirator be evaluated in terms of the proportion of users who will receive a specified level of protection 95% of the time. A point estimator of this proportion is easily obtained under an analysis of variance model, and the SCGD suggests bootstrap as one possible approach to interval estimation. This paper describes a closed-form procedure to obtain confidence intervals and provides sample size formulas. Simulation results suggest that the proposed procedure performs well in realistic settings and compares favorably to two simple bootstrap procedures.

Abrasion resistance of medical glove materials.

Due to the increasing demand for nonlatex medical gloves in the health-care community, there is a need to assess the durability of alternative glove materials. This study examines durability characteristics of various glove materials by abrasion resistance testing. Natural rubber latex (latex), polyvinyl chloride (vinyl), acrylonitrile butadiene (nitrile), polychloroprene (neoprene), and a styrene-ethylene/butylene-styrene block copolymer (SEBS) were tested. All test specimens, with the exception of the vinyl, were obtained from surgical gloves. Unaged out-of-the-box specimens as well as those subjected to various degrees of artificial aging were included in the study. After the abrasion sequence, the barrier integrity of the material was assessed through the use of a static leak test. Other traditional tests performed on these materials were viral penetration to validate the abrasion data and tear testing for comparative purposes. The results indicate that specific glove-material performance is dependent upon the particular test under consideration. Most notably, abrasion, even in controlled nonsevere conditions, may compromise to varying degrees the barrier integrity of latex, vinyl, SEBS, nitrile, and neoprene glove materials. However, as evidenced by the results of testing three brands of neoprene gloves, the abrasion resistance of any one glove material may be significantly affected by variations in production processes.

Biaxial flex-fatigue and viral penetration of natural rubber latex gloves before and after artificial aging.

Barrier integrity of unaged and oven-aged (at 70 degrees C) natural rubber latex examination gloves was assessed with a biaxial flex-fatigue method where failure was detected electronically, and by live viral penetration testing performed according to a modified version of ASTM F1671-97a. When no change in barrier properties was detected during flex testing, no virus passage was found after viral challenge. Conversely, when a change in the barrier properties was indicated by the electrical signal, virus passage was found in 74% of the specimens. Flex-fatigue results indicated that unaged test specimens from powdered (PD) and powder-free (PF) nonchlorinated gloves had significantly longer fatigue lives than powder-free chlorinated (CL) gloves from the same manufacturer. Biaxial flexing of oven-aged glove specimens showed a marginal increase in fatigue life for the PF gloves, but no increase for the PD gloves. The fatigue life of the CL gloves was observed to increase significantly after oven aging. However, this appears to be due to a design feature of the test apparatus, wherein peak volume displacement of the worked specimen is held constant. An aging-induced change in the viscoelastic properties of the CL gloves-permanent deformation of the specimens early in the fatigue test-relieves the stress magnitude applied as the test progresses. Thus, permanent deformation acts as a confounding factor in measuring durability of latex gloves by fixed displacement flex-fatigue.