Kathryn Chaloner

OPTIMAL BAYESIAN DESIGN APPLIED TO LOGISTIC REGRESSION EXPERIMENTS

Abstract A traditional way to design a binary response experiment is to design the experiment to be most efficient for a best guess of the parameter values. A design which is optimal for a best guess however may not be efficient for parameter values close to that best guess. We propose designs which formally account for the prior uncertainty in the parameter values. A design for a situation where the best guess has substantial uncertainty attached to itis very different from a design for a situation where approximate values of the parameters are known. We derive a general theory for concave design critria for non-linear models and then apply the theory to logistic regression. Designs found by numerical optimization are examined for a range of prior distributions and a range of criteria. The theoretical results are used to verify that the designs are indeed optimal.

Phase 3 Trial of Transplantation of Human Islets in Type 1 Diabetes Complicated by Severe Hypoglycemia

OBJECTIVE Impaired awareness of hypoglycemia (IAH) and severe hypoglycemic events (SHEs) cause substantial morbidity and mortality in patients with type 1 diabetes (T1D). Current therapies are effective in preventing SHEs in 50–80% of patients with IAH and SHEs, leaving a substantial number of patients at risk. We evaluated the effectiveness and safety of a standardized human pancreatic islet product in subjects in whom IAH and SHEs persisted despite medical treatment. RESEARCH DESIGN AND METHODS This multicenter, single-arm, phase 3 study of the investigational product purified human pancreatic islets (PHPI) was conducted at eight centers in North America. Forty-eight adults with T1D for >5 years, absent stimulated C-peptide, and documented IAH and SHEs despite expert care were enrolled. Each received immunosuppression and one or more transplants of PHPI, manufactured on-site under good manufacturing practice conditions using a common batch record and standardized lot release criteria and test methods. The primary end point was the achievement of HbA1c <7.0% (53 mmol/mol) at day 365 and freedom from SHEs from day 28 to day 365 after the first transplant. RESULTS The primary end point was successfully met by 87.5% of subjects at 1 year and by 71% at 2 years. The median HbA1c level was 5.6% (38 mmol/mol) at both 1 and 2 years. Hypoglycemia awareness was restored, with highly significant improvements in Clarke and HYPO scores (P > 0.0001). No study-related deaths or disabilities occurred. Five of the enrollees (10.4%) experienced bleeds requiring transfusions (corresponding to 5 of 75 procedures), and two enrollees (4.1%) had infections attributed to immunosuppression. Glomerular filtration rate decreased significantly on immunosuppression, and donor-specific antibodies developed in two patients. CONCLUSIONS Transplanted PHPI provided glycemic control, restoration of hypoglycemia awareness, and protection from SHEs in subjects with intractable IAH and SHEs. Safety events occurred related to the infusion procedure and immunosuppression, including bleeding and decreased renal function. Islet transplantation should be considered for patients with T1D and IAH in whom other, less invasive current treatments have been ineffective in preventing SHEs.

Assessment of a Beta Prior Distribution: PM Elicitation

An interactive computer scheme is described for eliciting from an analyst a beta prior distribution on the parameter 7t of a binomial distribution. Information on the analysts beta- binomial predictive distribution is obtained through a questioning and feedback algorithm based on modes.

Effect of early and late GB virus C viraemia on survival of HIV‐infected individuals: a meta‐analysis

To conduct a meta‐analysis to synthesize the evidence regarding the effect of co‐infection with GB virus C (GBV‐C) on survival of HIV‐infected individuals, and to estimate the effect.

Identification, Characterization and Immunogenicity of an O-Antigen Capsular Polysaccharide of Francisella tularensis

Capsular polysaccharides are important factors in bacterial pathogenesis and have been the target of a number of successful vaccines. Francisella tularensis has been considered to express a capsular antigen but none has been isolated or characterized. We have developed a monoclonal antibody, 11B7, which recognizes the capsular polysaccharide of F. tularensis migrating on Western blot as a diffuse band between 100 kDa and 250 kDa. The capsule stains poorly on SDS-PAGE with silver stain but can be visualized using ProQ Emerald glycoprotein stain. The capsule appears to be highly conserved among strains of F. tularensis as antibody 11B7 bound to the capsule of 14 of 14 F. tularensis type A and B strains on Western blot. The capsular material can be isolated essentially free of LPS, is phenol and proteinase K resistant, ethanol precipitable and does not dissociate in sodium dodecyl sulfate. Immunoelectron microscopy with colloidal gold demonstrates 11B7 circumferentially staining the surface of F. tularensis which is typical of a polysaccharide capsule. Mass spectrometry, compositional analysis and NMR indicate that the capsule is composed of a polymer of the tetrasaccharide repeat, 4)-α-D-GalNAcAN-(1->4)-α-D-GalNAcAN-(1->3)-β-D-QuiNAc-(1->2)-β-D-Qui4NFm-(1-, which is identical to the previously described F. tularensis O-antigen subunit. This indicates that the F. tularensis capsule can be classified as an O-antigen capsular polysaccharide. Our studies indicate that F. tularensis O-antigen glycosyltransferase mutants do not make a capsule. An F. tularensis acyltransferase and an O-antigen polymerase mutant had no evidence of an O-antigen but expressed a capsular antigen. Passive immunization of BALB/c mice with 75 µg of 11B7 protected against a 150 fold lethal challenge of F. tularensis LVS. Active immunization of BALB/c mice with 10 µg of capsule showed a similar level of protection. These studies demonstrate that F. tularensis produces an O-antigen capsule that may be the basis of a future vaccine.

Antibody repertoire development in fetal and neonatal piglets XXI. Usage of most VH genes remains constant during fetal and postnatal development

Usage of variable region gene segments during development of the antibody repertoire in mammals is unresolved in part because of the complexity of the locus in mice and humans and the difficulty of distinguishing intrinsic from extrinsic influences in these species. We present the first vertical studies on VH usage that spans the fetal and neonatal period using the piglet model. We tracked VH usage in DNA rearrangements and in VDJ transcripts throughout 75 days of gestation (DG) in outbred fetuses, thereafter in outbred germfree and colonized isolator piglets, isolator piglets infected with swine influenza and in conventionally reared nematode-infected adults. Seven VH genes account for >90% of the pre-immune repertoire which is the same among tissues and in both transcripts and DNA rearrangements. Statistical modeling supports the view that proportional usage of the major genes remains constant during fetal life and that postnatal usage ranking is similar to that during fetal life. Changes in usage ranking are developmental not antigen dependent. In this species exposure to environmental antigens results in diversification of the repertoire by somatic hypermutation of the same small number of VH genes that comprise the pre-immune repertoire, not by using other VH gene available in the germline. Therefore in swine a small number of VH genes shape the antibody repertoire throughout life questioning the need for extensive VH polygeny.

The safety and efficacy of adefovir dipivoxil in patients with advanced HIV disease: a randomized, placebo-controlled trial.

ObjectiveEfficacy and safety of adefovir dipivoxil (adefovir) added to background antiretroviral therapy in advanced HIV disease. DesignRandomized, double-blind, placebo-controlled multicenter trial. SettingFifteen clinical trial units providing HIV primary care. ParticipantsAdults with CD4 cell count ⩽ 100 × 106/l, or 101–200 × 106/l with prior nadir ⩽ 50 × 106/l. InterventionsOral adefovir or placebo 120 mg once daily. Main outcome measuresSurvival, cytomegalovirus (CMV) disease, plasma HIV-RNA, CD4 cell count, grade 4 drug toxicity, permanent drug discontinuation due to toxicity. ResultsAmong the 253 patients assigned adefovir and the 252 assigned placebo, respectively, 17 and 16 died (P = 0.88), and four and eight experienced CMV disease (P = 0.25). Mean change in log10 plasma HIV-RNA in the adefovir and placebo groups, respectively, was 0.09 and −0.03 copies/ml at 6 months (P = 0.22) and 0.06 and −0.02 at 12 months (P = 0.87). Changes in CD4 cell counts were not different between groups. At 12 months the cumulative percent with proximal renal tubular dysfunction (PRTD) was 17% in the adefovir group and 0.4% in the placebo group (P < 0.0001, log rank test). Median time to resolution of PRTD was 15 weeks among patients assigned adefovir, and 16% of patients did not resolve completely 41 weeks after onset. More drug discontinuations occurred in the adefovir group than in the placebo group. ConclusionsNo virologic or immunologic benefit was observed when adefovir was added to background antiretroviral therapy in advanced HIV disease, and adefovir was associated with considerable nephrotoxicity. This study does not support the use of adefovir for treatment of advanced HIV disease in pretreated patients.

The Equivalence of Constrained and Weighted Designs in Multiple Objective Design Problems

Abstract Several competing objectives may be relevant in the design of an experiment. The competing objectives may not be easy to characterize in a single optimality criterion. One approach to these design problems has been to weight each criterion and find the design that optimizes the weighted average of the criteria. An alternative approach has been to optimize one criterion subject to constraints on the other criteria. An equivalence theorem is presented for the Bayesian constrained design problem. Equivalence theorems are essential in verifying optimality of proposed designs, especially when (as in most nonlinear design problems) numerical optimization is required. This theorem is used to show that the results of Cook and Wong on the equivalence of the weighted and constrained problems apply much more generally. The results are applied to Bayesian nonlinear design problems with several objectives.

Bayesian design for accelerated life testing

Abstract In this paper we study experimental design for accelerated life tests where the lifetimes have either lognormal distributions or Weibull distributions. We assume that the quantities of interest are quantiles of the lifetime distribution at a specified stress and that the log lifetime decreases as a function of stress. There is an upper limit on the stress that can be used in the experiment and a fixed period of time in which to perform the experiment so that there is type I censoring when the experiment is terminated. Examples are given and it is shown that locally optimal designs, which correspond to putting a point mass prior distribution on the parameters, are very similar to Bayesian designs, which allow for prior uncertainty in the parameter values.

Bayesian analysis in statistics and econometrics: essays in honor of Arnold Zellner

This book is a definitive work that captures the current state of knowledge of Bayesian Analysis in Statistics and Econometrics and attempts to move it forward. It covers such topics as foundations, forecasting inferential matters, regression, computation and applications.