Proof of concept and dose estimation with binary responses under model uncertainty

This article suggests a unified framework for testing Proof of Concept (PoC) and estimating a target dose for the benefit of a more comprehensive, robust and powerful analysis in phase II or similar clinical trials. From a pre‐specified set of candidate models, we choose the ones that best describe the observed dose–response. To decide which models, if any, significantly pick up a dose effect, we construct the permutation distribution of the minimum P‐value over the candidate set. This allows us to find critical values and multiplicity adjusted P‐values that control the familywise error rate of declaring any spurious effect in the candidate set as significant. Model averaging is then used to estimate a target dose. Popular single or multiple contrast tests for PoC, such as the Cochran–Armitage, Dunnett or Williams tests, are only optimal for specific dose–response shapes and do not provide target dose estimates with confidence limits. A thorough evaluation and comparison of our approach to these tests reveal that its power is as good or better in detecting a dose–response under various shapes with many more additional benefits: It incorporates model uncertainty in PoC decisions and target dose estimation, yields confidence intervals for target dose estimates and extends to more complicated data structures. We illustrate our method with the analysis of a Phase II clinical trial. Copyright © 2008 John Wiley & Sons, Ltd.     

Species differences in contaminants in fish on and adjacent to the Oak Ridge Reservation, Tennessee

Risks to humans and other organisms from consuming fish have become a national concern in the USA. In this paper, we examine the concentrations of 137Cs, arsenic, beryllium, cadmium, lead, mercury, and selenium in three species of fish from two river reaches adjacent to the US Department of Energy's Oak Ridge Reservation in Tennessee. We were interested in whether there were species and locational differences in radiocesium and metal concentrations and whether concentrations were sufficiently high to pose a potential health risk to humans or other receptors. Striped bass (Morone saxatilis) were significantly larger than white bass (M. chrysops), and crappie (Pomoxis spp.) were the smallest fish. Lead was significantly lower in striped bass, mercury was significantly higher in striped bass, and selenium was significantly higher in white bass compared to the other species. There were no other species differences in contaminants. White bass, the only species that was sufficiently abundant for a comparison, had significantly higher concentrations of cadmium, lead, and selenium in fillets from the Clinch River and significantly higher concentrations of mercury in fillets from Poplar Creek. The low concentrations of most contaminants in fish from the Clinch River do not appear to present a risk to humans or other consumers, although mercury concentrations in striped bass ranged as high as 0.79 ppm, well above the 0.5-ppm action level for human consumption of some US states.     

Age-specific uncertainty in particulate deposition for 1 microm AMAD particles using the ICRP 66 lung model

The ICRP 66 lung model may be used to determine age-specific dose estimates for members of the public via the inhalation pathway. A significant source of uncertainty in internal dosimetric modeling is due to particulate deposition in regions of the respiratory tract. Uncertainties in estimates of particulate deposition are present because of the inherent variability in the deposition model and its various input parameters. An improved understanding of the uncertainty in particulate deposition will further guide research efforts and improve our ability to quantify internal dose estimates. The ICRP 66 lung deposition model is most sensitive to breathing rate when 1 microm AMAD particles are inhaled by members of the public. Uncertainties in deposition fractions are shown to span an order of magnitude with their distributions varying by age and sex for a particular lung region. Age-specific uncertainties of deposition fraction demonstrate increased estimates of extrathoracic deposition in younger age groups due to the decreased size of the respiratory tract airways. This age-specific trend becomes more pronounced for very young children and infants. The largest fractional deposition occurs in the alveolar and extrathoracic regions, regardless of age and sex.     

A possible model for speciation by cytoplasmic isolation in the Culex pipiens complex

Evidence so far available indicates that in the Culex pipiens complex there exists a genetic system with unique properties not comparable with those of the chromosomal gene system and probably involved in a mode of speciation peculiar to this complex. In a long series of backcrosses it has been shown that probably every Culex pipiens population contains a factor that is inherited through the cytoplasm, the only phenotypic expression of which is the crossability or non-crossability (incompatibility) of a given strain with other strains. That there is a whole series of such cytoplasmic factors is indicated by the twenty or so crossing types of Culex strains found throughout the world.     

Estimation of markov chain transition probabilities and rates from fully and partially observed data: uncertainty propagation, evidence synthesis, and model calibration.

Markov transition models are frequently used to model disease progression. The authors show how the solution to Kolmogorov's forward equations can be exploited to map between transition rates and probabilities from probability data in multistate models. They provide a uniform, Bayesian treatment of estimation and propagation of uncertainty of transition rates and probabilities when 1) observations are available on all transitions and exact time at risk in each state (fully observed data) and 2) observations are on initial state and final state after a fixed interval of time but not on the sequence of transitions (partially observed data). The authors show how underlying transition rates can be recovered from partially observed data using Markov chain Monte Carlo methods in WinBUGS, and they suggest diagnostics to investigate inconsistencies between evidence from different starting states. An illustrative example for a 3-state model is given, which shows how the methods extend to more complex Markov models using the software WBDiff to compute solutions. Finally, the authors illustrate how to statistically combine data from multiple sources, including partially observed data at several follow-up times and also how to calibrate a Markov model to be consistent with data from one specific study.     

A model for deposition of fibers in the human respiratory system.

A mathematical model for estimating lung deposition of fibers is described. It is based on the aerodynamic behavior of thin straight rods. Predicted deposition values for rods of various equivalent aerodynamic diameters and lengths, for each of three tidal volumes and for each of three respiratory system compartments are reported.     

A model for deposition of stable and unstable aerosols in the human respiratory tract

A model has been developed for deposition of stable and unstable aerosols in the human respiratory tract. This model has the advantages of employing a realistic asymmetric lung model and utilizing actual breath curves. Results from the model are presented for normal and abnormal breath patterns for both stable and unstable particles. Insofar as comparison is possible, agreement between model calculations and experiment appears to be reasonable.     

Cancer in the mass print media: fear, uncertainty and the medical model

Cancer is increasing in incidence and prevalence in North America and around the world. The mass print media play an important role in information provision about prevention, diagnosis and treatment of this disease, as well as informing health policy and personal experience. This paper reports on a content analysis of the portrayal of cancer in the highest circulating magazines available in Canada and published in Canada or the USA in 1991, 1996, 2001. It includes both manifest and latent analysis of the framing and content of cancer stories. Manifest analysis documented the dominance of the medical as compared to the lifestyle and political economy frames and the predominance of articles on breast as compared to other cancers. Latent themes included: an emphasis on fear of cancer in that: (1) cancer and fear are frequently conflated; cancer is said to grow outside of awareness; cancer is portrayed as (almost) inevitable; cancer is associated with normal experiences; early detection is associated with diagnosis; and scary statistics are emphasized; (2) contradictions and confusion exist within and between articles; and (3) metaphors of war and battle are used frequently. The paper concludes with a discussion of the implications of the linking of fear with cancer in the context of medicine as the solution.     

Food-chain and dose model, CALDOS, for assessing Canada's Nuclear Fuel Waste Management concept

The food-chain and dose model, CALculation of DOSe (CALDOS), was developed for assessing Canada's concept for nuclear fuel waste disposal in a vault deep in crystalline rock of the Canadian Shield. The model is very general and based on the Shield as a whole. The critical group is totally self-sufficient and represented by ICRP (1975) Reference Man for dose prediction. CALDOS assumes steady-state conditions and deals with variation and uncertainty through Monte Carlo simulation techniques. Ingrowth of some radioactive daughters is considered during food-chain transfer. A limit is set on root uptake to avoid unrealistic plant concentrations. Integrated ingestion and inhalation rates of man are calculated in a unique way, based on energy needs. Soil ingestion by man and external exposure from building material are unique pathways considered. Tritium, 129I, and 222Rn are treated through special models, and 14C and 129I involve unique geosphere dose limits. All transfer coefficients are lognormally distributed, and the plant/soil concentration ratio is correlated with the soil partition coefficient. Animals' ingestion rates are normally distributed and correlated with each other. Comprehensive sets of internal and external dose conversion factors were calculated for CALDOS. Sample calculations show that dose distributions tend to be strongly right-skewed. Many features of CALDOS are relevant for environmental assessment in general.     

Ingestion pathway model developed for use with an acute atmospheric dose model at the Savannah River Site

AXAIRQ is a straight-line Gaussian plume dose model used for prospective accident assessment at the Savannah River Site and currently includes the following dose pathways: inhalation, ground shine, and plume shine. In the event of an accident, another possible pathway for dose would be through ingestion of locally produced contaminated foodstuffs. A model, AXINGST, has been developed that would incorporate this pathway. Currently available ingestion models were referenced as a basis for AXINGST. The model calculates an ingestion dose following an atmospheric release of radionuclides and includes site-specific variables where applicable. AXINGST estimates tritium vegetation concentrations that are within a factor of 20-40 of measured data during previous accidental releases at SRS.     

A dermal model for spray painters. Part II: estimating the deposition and uptake of solvents

The contribution of dermal exposure to the total body burden of a given chemical is difficult to assess. However, it is possible that as regulatory pressures lead to reductions in inhalation exposure, the proportion of uptake via the dermal route will increase. This study brings together recent work in the field of dermal exposure to provide a model to estimate both exposure and uptake of solvents through the skin. Using spray painters as an example, the process of modelling exposure is described from identifying the determinants of exposure through to calculating the flux of solvent through the stratum corneum and thus the total dermal uptake. Results from a range of exposure scenarios are presented and areas requiring further research and validation are highlighted. The model should allow the estimation of combined dermal and inhalation exposure to solvents in both current work situations and retrospective epidemiological studies. Furthermore, it is envisaged that the model may provide the basis for generic assessment of dermal exposure and uptake.     

A multi-product multi-period stochastic model for a blood supply chain considering blood substitution and demand uncertainty

Health Care Management Science - This paper presents a multi-product multi-period stochastic program for an integrated blood supply chain that considers red blood cells and platelets while...     

Multicellular model for contaminant dispersion and ventilation effectiveness with application for oxygen deficiency in a confined space

Multicellular models were developed to predict contaminant dispersion in a three-dimensional (3-D) space. The method utilized in this study was patterned after models which have been developed for water pollution in lakes and streams. This approach involved (1) design of cell structures to accommodate the geometry and mass flow characteristics of the 3-D space and (2) approximation of dispersion coefficients to describe contaminant transport in addition to that resulting from mass flow between cells. The dispersion model utilized a mass balance equation to predict contaminant dispersion as a function of time. The computer model was evaluated against experimental data for oxygen deficiency inside a ventilated confined space (CS) model. Eight test cases of ventilation design for the CS model were tested for each of three contaminant release (nitrogen to cause oxygen deficiency) characteristics: (1) purging (oxygen recovery from an initial deficiency); (2) steady state; and (3) variable rate. The dispersion model did a reasonably good job of predicting oxygen concentrations for different locations (cells) in the CS model. The primary limitations of this multicellular method are associated with experimental approximations of flow patterns and dispersion coefficients.     

Verification and modification of the ICRP-67 model for plutonium dose calculation

On the basis of the available data and empirical expressions for the plutonium excretion after injection, an age-related compartmental model has been developed. It provides a better agreement with measured urinary excretion data than the current ICRP 67 model. Moreover, the revised model avoids unphysiological assumptions such as the transfer of activity from soft tissue to urinary bladder, that were part of the ICRP model. The new predictions of the activity in feces and in blood after an injection are closer to the available data than the ICRP 67 estimations and there is also a good agreement with the partitioning of plutonium between skeleton and liver obtained from different autopsy studies. Furthermore, the urinary excretion estimated by the improved model has been checked using some data from occupationally exposed individuals. As the plutonium uptake in these workers occurred by inhalation, the improved model and the ICRP 67 model were compared by connecting them to the ICRP 66 respiratory tract model. The improved model consistently yields a better agreement with the measured excretion and higher estimations of intake than the ICRP 67 model.     

Impact of model, methodological, and parameter uncertainty in the economic analysis of vaccination programs.

Guidelines for economic evaluations insist that the sensitivity of model results to alternative parameter values should be thoroughly explored. However, differences in model construction and analytical choices (such as the choice of a cost-effectiveness or cost-benefit framework) also introduce uncertainty in results, though these are rarely subjected to a thorough sensitivity analysis. In this article, the authors quantify the effect of model, methodological, and parameter uncertainty, taking varicella vaccination as an example. They used 3 different models (a static model, a dynamic model that only looks at the effect of vaccination on varicella, and a dynamic model that also assesses the implications of vaccination for zoster epidemiology) and 2 forms of analysis (cost-benefit and cost-utility). They also varied the discount rate and time frame of analysis. Probabilistic sensitivity analyses were performed to estimate the impact of parameter uncertainty. In their example, model and methodological choice had a profound effect on estimated cost-effectiveness, but parameter uncertainty played a relatively minor role. Under cost-utility analysis, the probabilistic sensitivity analysis suggested that there was a near certainty that vaccination dominates no vaccination, or the other way around, depending on model choice and perspective. Under cost-benefit analysis, vaccination always appeared to be attractive. Thus, the authors clearly show that model and methodological assumptions can have greater impact on results than parameter estimates, although sensitivity analyses are rarely performed on these sources of uncertainty.     

Correction factors in the EPR dose reconstruction for residents of the Middle and Lower Techa riverside.

During 1949-1956, the first Soviet nuclear weapons plant, Mayak, released about 7.6 x 10(7) m(-3) of liquid radioactive waste with a total activity of 10(17) Bq into the Techa River (Southern Urals, Russia). 90Sr contributed 11.6% to the total waste radioactivity. As a result of these radioactive discharges, about 28,000 local residents were exposed to ionizing radiation, and some of them received relatively high doses. Internal exposure of the population residing at the Middle and Lower Techa riverside was mostly from 90Sr deposited in bone and tooth tissues. In order to reconstruct radiation doses to this population group, a study of 35 teeth extracted from local residents was carried out using electron paramagnetic resonance measurements. A total of 73 samples from these 35 teeth (tooth enamel, 33; crown dentin, 20; and root dentin, 20) were prepared and measured with electron paramagnetic resonance. The study revealed high doses (up to 15 Gy) absorbed in tooth enamel of the individuals born during 1945-1949, which was attributed to very high local 90Sr concentration in tooth enamel of this particular age group in the population. The analysis presented here takes into account (a) the time courses both of the release/intake of 90Sr and of the tooth formation, and (b) expected variations in measured absorbed doses due to differing geometric sizes of tooth structures. This methodology enables a more consistent picture to be developed of the 90Sr intake by the Middle and Lower Techa riverside population, based on electron paramagnetic resonance tooth dosimetry.     

Correlation of lung dose with Rn concentration, potential alpha-energy concentration and daughter surface deposition: a Monte Carlo analysis

In the evaluation of lung cancer risk from Rn daughters in the home environment, measurements of Rn concentration or potential alpha-energy concentration (PAEC) have served as proxies for actual radiation dose to the lungs. This paper uses a single-compartment room model, model parameters from a number of recent studies, and Monte Carlo analysis to explore the relative value of Rn concentration, PAEC, and room surface deposition, or plate-out, as estimators of lung dose. Results indicate that Rn concentration and PAEC are fairly good estimators, explaining roughly 50-70% of the variation in dose for the conditions studied, over the range of Rn concentrations considered (74-555 Bq m 3). The relative advantage of one measure over the other depends upon the variability in a number of factors across the houses being evaluated. Room surface activity deposition was not found to be superior to the traditional measures of Rn concentration and PAEC. Significant room for improvement remains in the development of simple home monitors providing an improved estimation of lung dose.     

微尺度辐射环境影响评价软件系统开发

在城市街道及设施厂区等微尺度范围,由于大量的建筑物存在,其流场特征与扩散过程将会直接受到建筑物几何特征的影响,直线高斯烟羽模式将不再适用。基于雷诺平均的欧拉方法计算风场耦合拉格朗日粒子随机游走模式,建立放射性污染物的大气扩散模式,开发微尺度辐射环境影响评价软件系统,可用于厂区内或城市街道等放射性污染物的扩散行为。该系统的开发为今后从事微尺度辐射环境影响评价工作提供技术工具。     

Relating the sediment phase speciation of arsenic, cadmium, and chromium with their bioavailability for the deposit-feeding polychaete Nereis succinea

We studied the influence of sediment geochemistry on bioavailability of As, Cd, and Cr in deposit-feeding polychaetes. Metal phase speciation in sediments was determined with a sequential extraction scheme, and assimilation efficiencies (AEs) of ingested metals were determined by pulse-chase feeding experiments using γ-emitting isotopes. Worms were fed sediments collected from geochemically diverse estuaries that were labeled by sorbing dissolved radiotracers or mixing with radiolabeled algae. Uptake of sediment-bound metals was compared with that from labeled algae or goethite. Metal AEs showed a positive relationship with the exchangeable and carbonate sedimentary fractions, whereas metals in iron and manganese oxides and acid-volatile sulfides, or in pyrite and other refractory material, were inversely correlated with AEs. Arsenic was most bioavailable from algae (72%), less from sediments mixed with algae (24-70%) and least from sediments labeled directly (1-12%). Arsenic AEs in sediments labeled directly showed a positive correlation with sedimentary Mn and Al and negative correlation with Fe. Cadmium AEs were positively correlated with salinity and negatively correlated with sedimentary organic C. The AEs of Cr from sediments or algae were less than 5%, but they were 34% from pure goethite. By quantifying the relationship of metal speciation in sediments with their bioavailability for deposit-feeding polychaetes, the present study provides new insight into understanding metal bioaccumulation in benthic invertebrates.