Robert C. Spear

Parameter uncertainty and interaction in complex environmental models

Recently developed models for the estimation of risks arising from the release of toxic chemicals from hazardous waste sites are inherently complex both structurally and parametrically. To better understand the impact of uncertainty and interaction in the high-dimensional parameter spaces of these models, the set of procedures termed regional sensitivity analysis has been extended and applied to the groundwater pathway of the MMSOILS model. The extension consists of a tree-structured density estimation technique which allows the characterization of complex interaction in that portion of the parameter space which gives rise to successful simulation. Results show that the parameter space can be partitioned into small, densely populated regions and relatively large, sparsely populated regions. From the high-density regions one can identify the important or controlling parameters as well as the interaction between parameters in different local areas of the space. This new tool can provide guidance in the analysis and interpretation of site-specific application of these complex models.

Environmental effects on parasitic disease transmission exemplified by schistosomiasis in western China

Environmental effects on the transmission of many parasitic diseases are well recognized, but the role of specific factors like climate and agricultural practices in modulating transmission is seldom characterized quantitatively. Based on studies of Schistosoma japonicum transmission in irrigated agricultural environments in western China, a mathematical model was used to quantify environmental impacts on transmission intensity. The model was calibrated by using field data from intervention studies in three villages and simulated to predict the effects of alternative control options. Both the results of these interventions and earlier epidemiological findings confirm the central role of environmental factors, particularly those relating to snail habitat and agricultural and sanitation practices. Moreover, the findings indicate the inadequacy of current niclosamide-praziquantel strategies alone to achieve sustainable interruption of transmission in some endemic areas. More generally, the analysis suggests a village-specific index of transmission potential and how this potential is modulated by time-varying factors, including climatological variables, seasonal water-contact patterns, and irrigation practices. These time-variable factors, a villages internal potential, and its connectedness to its neighbors provide a framework for evaluating the likelihood of sustained schistosomiasis transmission and suggest an approach to quantifying the role of environmental factors for other parasitic diseases.

Correlation of urinary pesticide metabolite excretion with estimated dermal contact in the course of occupational exposure to guthion

Exposure to and absorption of Guthion 50 W.P. (azinphos-methyl) were estimated in orchardists from the Okanagan Valley in British Columbia who were involved in mixing, loading, and application with ultra-low volume air blast equipment. Air monitoring and patch techniques were used to estimate exposure, and alkyl phosphate excretion and cholinesterase inhibition were measured to estimate absorption. All workers were issued with standardized cotton shirts, trousers, and long-sleeved coveralls. All wore half-face respirators, gloves, boots, and hats. Eight wore rubberized protective clothing in addition. The indirect method of measuring urinary metabolites appeared to be the most sensitive. All workers had quantifiable levels of alkyl phosphates following exposure, and 24-h urine samples provided a more reliable estimate than first morning voids. A high correlation was observed between 48-h alkyl phosphate excretion and amount of active ingredient sprayed. A fluorescent tracer was added to the tank along with the Guthion. The finding of Guthion on patches beneath the clothing was confirmed by the presence of the tracer on the skin. With the ultralow-volume application used in this study, the rubberized clothing did not appear to be significantly more protective than the heavy coverall. There was no significant depression of either red blood cell or serum cholinesterase activity in any workers.

An analysis of the Milwaukee cryptosporidiosis outbreak based on a dynamic model of the infection process.

We combined information on the temporal pattern of disease incidence for the 1993 cryptosporidiosis outbreak in Milwaukee with information on oocyst levels to obtain insight into the epidemic process. We constructed a dynamic process model of the epidemic with continuous population compartments using reasonable ranges for the possible distribution of the model parameters. We then explored which combinations of parameters were consistent with the observations. A poor fit of the March 1–22 portion of the time series suggested that a smaller outbreak occurred before the March 23 treatment failure, beginning sometime on or before March 1. This finding suggests that had surveillance systems detected the earlier outbreak, up to 85% of the cases might have been prevented. The same conclusion was obtained independent of the model by transforming the incidence time series data of Mac Kenzie et al.1 This transformation is based on a background monthly incidence rate for watery diarrhea in the Milwaukee area of 0.5%.1 Further analysis using the incidence data from the onset of the major outbreak, March 23, through the end of April, resulted in three inferred properties of the infection process: (1) the mean incubation period was likely to have been between 3 and 7 days; (2) there was a necessary concurrent increase in Cryptosporidium oocyst influent concentration and a decrease in treatment efficiency of the water; and (3) the variability of the dose-response function in the model did not appreciably affect the simulated outbreaks. (Epidemiology 1998;9:255–263)

A Video Imaging Technique for Assessing Dermal Exposure I. Instrument Design and Testing

A new method for measuring dermal exposure has been developed which employs video imaging technology and fluorescent tracers. Workers are examined under long wave ultraviolet light following exposure, and fluorescence is quantitated by a television camera interfaced to a microcomputer. This instrument demonstrates high stability over extended periods of use. Through proper calibration it can quantitate a wide range of fluorescent tracer densities on the skin accurately. The lower limit of detection is 100 ng/cm2. Quenching occurs when very high levels of fluorescence are detected, creating an upper limit to quantitation.

A Video Imaging Technique for Assessing Dermal Exposure II. Fluorescent Tracer Testing

Laboratory and field evaluations were conducted to determine the suitability of employing a fluorescent tracer in conjunction with video imaging analysis to measure dermal exposure during pesticide applications. The Fluorescent Whitening Agent 4-methyl-7-diethylaminocoumarin and the organophosphate malathion were highly correlated (r = .985) when sprayed under controlled conditions. Deposition levels during field studies were correlated similarly (r = .942); however, variability in deposition ratios requires that field sampling be conducted to determine the ratio for a particular application. Penetration of the two compounds through cotton/polyester workshirt material demonstrated a high correlation (r = .979), whereas penetration of cotton/polyester coverall material was more variable (r = .834). The slopes of the regression lines for the two materials were not significantly different. The ratio of pesticide and tracer recovered from targets was consistently higher than the initial tank ratio due to differences in solubility and mixing.

A quantitative framework for a multi-group model of Schistosomiasis japonicum transmission dynamics and control in Sichuan, China

A quantitative framework is presented for the site-specific characterization of schistosomiasis transmission with the object of developing local control strategies. Central to the framework is a worm-burden model using ordinary differential equations of disease transmission in risk groups defined by residence and occupation. The model incorporates temperature- and precipitation-dependent seasonality of infectious stages, snail population dynamics, and seasonal patterns of human water contact specific to the local agricultural setting. The models parameters are separated into two main subsets, those associated with the general biology of the parasite and its life cycle in the human and the snail and those associated with directly measurable features of disease status in the local population or relevant aspects of the local environment. In this regard, the model is structured and parameterized to take maximum advantage of data that can be collected in rural China by conventional methods. For example, it includes a statistical model for egg excretion to the environment by each risk group which is based on local population surveys of the prevalence and intensity of infection. The second element of the framework of analysis relates to the strategy for parameter estimation and calibration to local conditions. We propose a Bayesian approach in which parameter estimates are refined over time by methods employing extensive computer simulations. An early analysis of data collected between 1987 and 1989 in endemic villages near Xichang City in southwestern Sichuan provides encouragement that parametric uncertainty can be reduced to levels adequate to explore effective control strategies.

Large simulation models: calibration, uniqueness and goodness of fit

Abstract Large simulation models of environmental systems which are based on biological and physical mechanisms are useful because of their ability to integrate diverse types of information relevant to the problem under analysis. Inherent in such models is a high degree of both structural and parametric complexity. A number of modeling and simulation studies have demonstrated that there are many parameter sets which produce good fits to calibration data. This lack of uniqueness requires a different perspective on parameter estimation which can be usefully addressed employing computer-intensive methods of multivariate statistical analysis. Regardless of how parameters are estimated, there appears to be a strong case to be made that the ‘best’ parameter estimate is an extended and complex region in a high dimensional space.

Geographic and ecologic heterogeneity in elimination thresholds for the major vector-borne helminthic disease, lymphatic filariasis

BackgroundLarge-scale intervention programmes to control or eliminate several infectious diseases are currently underway worldwide. However, a major unresolved question remains: what are reasonable stopping points for these programmes? Recent theoretical work has highlighted how the ecological complexity and heterogeneity inherent in the transmission dynamics of macroparasites can result in elimination thresholds that vary between local communities. Here, we examine the empirical evidence for this hypothesis and its implications for the global elimination of the major macroparasitic disease, lymphatic filariasis, by applying a novel Bayesian computer simulation procedure to fit a dynamic model of the transmission of this parasitic disease to field data from nine villages with different ecological and geographical characteristics. Baseline lymphatic filariasis microfilarial age-prevalence data from three geographically distinct endemic regions, across which the major vector populations implicated in parasite transmission also differed, were used to fit and calibrate the relevant vector-specific filariasis transmission models. Ensembles of parasite elimination thresholds, generated using the Bayesian fitting procedure, were then examined in order to evaluate site-specific heterogeneity in the values of these thresholds and investigate the ecological factors that may underlie such variabilityResultsWe show that parameters of density-dependent functions relating to immunity, parasite establishment, as well as parasite aggregation, varied significantly between the nine different settings, contributing to locally varying filarial elimination thresholds. Parasite elimination thresholds predicted for the settings in which the mosquito vector is anopheline were, however, found to be higher than those in which the mosquito is culicine, substantiating our previous theoretical findings. The results also indicate that the probability that the parasite will be eliminated following six rounds of Mass Drug Administration with diethylcarbamazine and albendazole decreases markedly but non-linearly as the annual biting rate and parasite reproduction number increases.ConclusionsThis paper shows that specific ecological conditions in a community can lead to significant local differences in population dynamics and, consequently, elimination threshold estimates for lymphatic filariasis. These findings, and the difficulty of measuring the key local parameters (infection aggregation and acquired immunity) governing differences in transmission thresholds between communities, mean that it is necessary for us to rethink the utility of the current anticipatory approaches for achieving the elimination of filariasis both locally and globally.

Comparison of three physiologically-based pharmacokinetic models of benzene disposition

We assess the goodness of fit of three physiologically based models of benzene pharmacokinetics to experimental data in Fischer-344 rats. These models were independently developed and published. Large differences in the quality of the fit are observed. In addition, the parameter values leading to acceptable fits are spread over the entire range of physiologically plausible values and can be quite different from average or standard values. On the other hand, choosing standard values for the parameters does not ensure good predictions of all tissue levels. These results emphasize the difficulty of a rigorous calibration of physiological models, and the need for further research in this area, including precise experimental determination of parameter values. Physiological models are powerful tools, but for risk assessment purposes simpler models, making equivalent use of the crucial data, are probably preferable.