Dose-response Estimates Research Articles

Bayesian Estimation of Inverse Dose Response

Inverse dose-response estimation refers to the inference of an effective dose of some agent that gives a desired probability of response, say 0.5. We consider inverse dose response for two agents, an application that has not received much attention in the literature. Through the posterior profiling technique (Hsu, 1995, The Canadian Journal of Statistics 23, 399-410), we propose a Bayesian method in which we approximate the marginal posterior distribution of an effective dose using a profile posterior distribution, and obtain the maximum a posteriori (MAP) estimate for the effective dose. We then employ an adaptive direction sampling algorithm to obtain the highest posterior density (HPD) credible region for the effective dose. Using the MAP and HPD estimates, investigators will be able to simultaneously calibrate the levels of two agents in dose-response studies. We illustrate our proposed Bayesian method through a simulation study and two practical examples.

A Stata Package for the Estimation of the Dose-response Function through Adjustment for the Generalized Propensity Score

In this article, we briefly review the role of the propensity score in estimating dose–response functions as described in Hirano and Imbens (2004, Applied Bayesian Modeling and Causal Inference from Incomplete-Data Perspectives, 73–84). Then we present a set of Stata programs that estimate the propensity score in a setting with a continuous treatment, test the balancing property of the generalized propensity score, and estimate the dose–response function. We illustrate these programs by using a dataset collected by Imbens, Rubin, and Sacerdote.

A method to provide improved dose–response estimates for airborne pathogens in animals: An example using porcine reproductive and respiratory syndrome virus

This paper describes a method to provide improved probability estimates that exposure to a specific dose of an airborne infectious pathogen will result in animal infection. Individual animals were exposed to a specific dose of airborne pathogen. Following exposure, animals were individually housed and monitored for evidence of infection. The detection of specific antibodies and/or the pathogen in diagnostic specimens was evidence that the exposure dose resulted in infection. If replicated over a range of doses, the results can be used to derive a dose–response curve for a variety of animal species and infectious pathogens. This information is useful in estimating the likelihood of infection associated with exposure to airborne infectious microorganisms. Applications include predicting the risk of transmission associated with exposure to airborne pathogens, modeling the transmission of airborne pathogens, and determining requirements for effective exposure doses for vaccines delivered in aerosols.

Self-poisoning during pregnancy as a model for teratogenic risk estimation of drugs

The results of animal investigations cannot be directly extrapolated to pregnant women. Clinical and analytical epidemiological studies for human teratogenic risks have many inherent methodological problems and their predictions must be regarded with caution. Evaluation of the potential teratogenicity of extremely large doses of drugs in self-poisoned pregnant women who attempted suicide offers a unique model for finding associations between congenital abnormalities and different drugs. All self-poisoned patients were cared for at a toxicological inpatient clinic in Budapest, between 1960 and 1993. Out of 1044 pregnant women identified from three different periods of the project, 19 died. Women who survived were visited at home to reveal birth outcomes, and their children were evaluated medically to identify congenital abnormalities and to estimate cognitive-behavioral status. The previous or subsequent children of these pregnant women were used as controls. In general, self-poisoned pregnant women were young (peak age was 18 to 20 years) and primiparous; 55% were unmarried and of lower socioeconomic status. Suicide attempts with drugs were most frequent in the fourth postconceptional week and second pregnancy month. Of 1044 self-poisoned pregnant women, 411 delivered live-born babies, of which 367 children (89.3%) were evaluated. The self-poisoning model appears to have several benefits (e.g., dose-response estimation) in comparison with other methods. It is suggested that an international monitoring system of self-poisoned pregnant women should be established to provide a larger database.

Consumption of animal foods and endometrial cancer risk: a systematic literature review and meta-analysis

This article summarizes and quantifies the current evidence relating dietary intake of animal products and endometrial cancer. Literature searches were conducted to identify peer-reviewed manuscripts published up to December 2006. Twenty-two manuscripts from three cohort studies and 16 case-control studies were identified. One of these cohort studies evaluated only fried meat and another only milk consumption; they were not included in our meta-analyses. The third cohort study identified did not present exposure levels and could not be included in dose-response meta-analysis. This cohort study did not show an association with meat or red meat consumption. Random-effects dose-response summary estimates for case-control studies evaluating these foods were 1.26 (95% CI: 1.03-1.54) per 100 g/day of total meat, 1.51 (95% CI: 1.19-1.93) per 100 g/day of red meat, 1.03 (95% CI: 0.32-3.28) per 100 g/day of poultry, 1.04 (95% CI: 0.55-1.98) per 100 g/day of fish, and 0.97 (95% CI: 0.93-1.01) per serving of dairy. Our meta-analysis, based on case-control data, suggests that meat consumption, particularly red meat, increases endometrial cancer risk. The current literature does not support an association with dairy products, while the evidence is inconsistent for poultry, fish, and eggs. More studies, particularly prospective studies, are needed.

Solid Cancer Incidence in Atomic Bomb Survivors: 1958–1998

This is the second general report on radiation effects on the incidence of solid cancers (cancers other than malignancies of the blood or blood-forming organs) among members of the Life Span Study (LSS) cohort of Hiroshima and Nagasaki atomic bomb survivors. The analyses were based on 17,448 first primary cancers (including non-melanoma skin cancer) diagnosed from 1958 through 1998 among 105,427 cohort members with individual dose estimates who were alive and not known to have had cancer prior to 1958. Radiation-associated relative risks and excess rates were considered for all solid cancers as a group, for 19 specific cancer sites or groups of sites, and for five histology groups. Poisson regression methods were used to investigate the magnitude of the radiation-associated risks, the shape of the dose response, how these risks vary with gender, age at exposure, and attained age, and the evidence for inter-site variation in the levels and patterns of the excess risk. For all solid cancers as a group, it was estimated that about 850 (about 11%) of the cases among cohort members with colon doses in excess of 0.005 Gy were associated with atomic bomb radiation exposure. The data were consistent with a linear dose response over the 0- to 2-Gy range, while there was some flattening of the dose response at higher doses. Furthermore, there is a statistically significant dose response when analyses were limited to cohort members with doses of 0.15 Gy or less. The excess risks for all solid cancers as a group and many individual sites exhibit significant variation with gender, attained age, and age at exposure. It was estimated that, at age 70 after exposure at age 30, solid cancer rates increase by about 35% per Gy (90% CI 28%; 43%) for men and 58% per Gy (43%; 69%) for women. For all solid cancers as a group, the excess relative risk (ERR per Gy) decreases by about 17% per decade increase in age at exposure (90% CI 7%; 25%) after allowing for attained-age effects, while the ERR decreased in proportion to attained age to the power 1.65 (90% CI 2.1; 1.2) after allowing for age at exposure. Despite the decline in the ERR with attained age, excess absolute rates appeared to increase throughout the study period, providing further evidence that radiation-associated increases in cancer rates persist throughout life regardless of age at exposure. For all solid cancers as a group, women had somewhat higher excess absolute rates than men (F:M ratio 1.4; 90% CI 1.1; 1.8), but this difference disappears when the analysis was restricted to non-gender-specific cancers. Significant radiation-associated increases in risk were seen for most sites, including oral cavity, esophagus, stomach, colon, liver, lung, non-melanoma skin, breast, ovary, bladder, nervous system and thyroid. Although there was no indication of a statistically significant dose response for cancers of the pancreas, prostate and kidney, the excess relative risks for these sites were also consistent with that for all solid cancers as a group. Dose-response estimates for cancers of the rectum, gallbladder and uterus were not statistically significant, and there were suggestions that the risks for these sites may be lower than those for all solid cancers combined. However, there was emerging evidence from the present data that exposure as a child may increase risks of cancer of the body of the uterus. Elevated risks were seen for all of the five broadly classified histological groups considered, including squamous cell carcinoma, adenocarcinoma, other epithelial cancers, sarcomas and other non-epithelial cancers. Although the data were limited, there was a significant radiation-associated increase in the risk of cancer occurring in adolescence and young adulthood. In view of the persisting increase in solid cancer risks, the LSS should continue to provide important new information on radiation exposure and solid cancer risks for at least another 15 to 20 years.

Methylmercury and IQ: Dose–Response Estimate of Prenatal Effect

Vol. 115, No. 4 EnvironewsOpen AccessMethylmercury and IQ: Dose–Response Estimate of Prenatal Effect Valerie J. Brown Valerie J. Brown Search for more papers by this author Published:1 April 2007https://doi.org/10.1289/ehp.115-a212aAboutSectionsPDF ToolsDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InReddit Methylmercury, the most biologically active mercury compound, is well known to cause serious health effects, particularly to the developing fetal nervous system. Effects can include attention deficits as well as IQ, motor, memory, and language impairment. A new analysis now combines data from three earlier studies to produce an integrated estimate of the dose–response relationship between maternal mercury exposure during pregnancy and lowered childhood IQ [EHP 115:609–615; Axelrad et al.].The authors analyzed combined IQ data from three longitudinal studies conducted in the Faroe Islands, the Seychelles Islands, and New Zealand. These studies measured a variety of neurodevelopmental end points, including IQ, attention, and motor skills. The range of pre-natal exposures in the three populations is comparable to those of some U.S. populations. For example, a 2003 study found the lowest maternal blood mercury level in the Faroe Islands to be 0.53 μg/L, and the CDC reported in 2004 that more than half of U.S. women had blood mercury concentrations higher than this. Geometric mean blood concentrations in the United States from 1999 to 2002 were 0.92 μg/L for women of childbearing age; for children the mean was 0.33 μg/L.The New Zealand and Seychelles studies reported results in terms of ppm of hair mercury, whereas the Faroe Islands study reported effects in terms of ppb of cord blood mercury. So the team converted the Faroe Islands results to their equivalents in units of hair mercury. They found a childhood IQ decrease of 0.18 points for each ppm rise in maternal hair mercury. The team assumed a linear, nonthreshold dose–response curve. However, they noted that if very low exposures produce a steeper curve, as has been found recently with childhood lead exposure, their calculation may underestimate the effects of prenatal mercury exposure. Similarly, certain cognitive abilities such as word retrieval and retention of verbally presented information are not captured by IQ scores, so relying only on IQ as a measure of cognitive function will also underestimate mercury’s effects.Eating fish is the most common route of human exposure to methylmercury. In 2004 the FDA and the EPA issued a joint statement advising women of childbearing age and children to limit their weekly consumption of commercially caught fish to 12 oz (6 oz for locally caught fish) in order to avoid harmful exposure. The EPA has set a reference dose of 0.1 μg/kg/day for methylmercury as an estimate of the daily exposure unlikely to cause harm over a lifetime.Methylmercury’s effect on IQ is separate from its effect on attention and motor skills. But because IQ is a well-established end point used in cost–benefit and economic analyses of the effects of environmental contaminants, establishing the dose–response relationship for IQ is a first step in quantifying the benefits of reducing mercury exposure.Smart moveA new analysis takes a first step toward quantifying the benefits of reducing mercury exposure, which may include avoiding IQ deficits.FiguresReferencesRelatedDetails Vol. 115, No. 4 April 2007Metrics About Article Metrics Publication History Originally published1 April 2007Published in print1 April 2007 Financial disclosuresPDF download License information EHP is an open-access journal published with support from the National Institute of Environmental Health Sciences, National Institutes of Health. All content is public domain unless otherwise noted. Note to readers with disabilities EHP strives to ensure that all journal content is accessible to all readers. However, some figures and Supplemental Material published in EHP articles may not conform to 508 standards due to the complexity of the information being presented. If you need assistance accessing journal content, please contact [email protected]. Our staff will work with you to assess and meet your accessibility needs within 3 working days.

Monotone Smoothing with Application to Dose-Response Curve

Motivated by problems that arise in dose-response curve estimation, we developed a new method to estimate a monotone curve. The resulting monotone estimator is obtained by combining techniques from smoothing splines with nonnegativity properties of cubic B-splines. Numerical experiments are given to exemplify the method.

Environmental Burden of Disease from Exposure to Lead in Ireland

P-709 Introduction: The WHO Environmental Burden of Disease (EBD) methodology was used to assess burden of disease from lead exposure in Ireland. The aim was to produce a national baseline assessment and to evaluate whether intervention was required. Methods: The outcomes used for children were IQ loss and resulting rates of mild mental retardation (MMR), rates of anaemia, and gastrointestinal symptoms were calculated for children. For adults, raised blood pressure as an indicator for risk of cardiovascular disease was used. Mean blood lead levels from a survey of a rural population living beside an old mining site were used and assumed to be representative of urban exposures (Adults = 3.0μg/dL, n=341; pre-school children= 2.7μg/dL, n=84). Communities beside mining sites have been found to have similar levels of exposure to urban populations. Urban populations are assumed to have a greater proportion of their exposure from transport sources so estimates were constructed with and without the use of a correction factor to account for annual reductions since the ban on lead in gasoline in 2000. Assuming a normal distribution of the log-transformed means, the proportion of the population at each blood Pb interval was estimated, and rates of disease for each health outcome were mapped to each interval, based on the WHO dose-response estimates. Results: The attributable incidence rates of MMR in children was low (0.25/1000), as were rates of gastrointestinal symptoms (0.06/1000) and anaemia (0.04/1000). The Attributable Fractions for cerebrovascular and hypertensive disease in men was 1% and 0% for ischaemic heart disease and other heart disease in men. The Attributable Fraction for all cardiovascular outcomes in women was 0%. Discussion & Conclusions: The study population used is not representative of the Irish population. Populations adjacent to mining sites do not experience exposures comparable to those observed in communities beside other point sources, such as smelters. The survey population may be more representative of rural exposures, making the current assessment an underestimate. Analysis and comparison with similar urban and rural populations will be conducted to determine whether the current assessment should be assumed as a low, medium or high estimate. In countries with long-standing legislative controls on the use of lead, exposure of the general population will be low. Locally high levels due to point sources and exposure of children living in degraded housing are likely to be more important. Targeted surveillance of identified ‘at risk’ groups as recommended by the CDC in the US may be more practical for informing public health than the wider EBD assessment. The study highlighted the difficulty in conducting epidemiology in the absence of any biomonitoring in Ireland.

Reducing Uncertainty in Estimating Complex Exposure for Integrated Health Risk Assessment

MM2-O-04 Present salient problems in modeling exposure to chemical stressors include the multitude and complexity of pathways, lack of data and explicit mechanistic understanding for identical dose-response assessment, and the diversity of phenotypic consequences of exposure, among others. Joint interaction of modifying factors including genetics, diet, and age window of exposure makes the articulation of an effective framework for complex exposure modeling in support of integrated human health risk assessment even more challenging. This work introduces the concept of a multi-layered computational platform supported by a profound knowledge base enabling rapid advances in toxicology and environmental health. This platform should comprise links with environmental and health monitoring systems at the regional and global scales taking into account the spatial and temporal variation of information from monitoring networks. It includes multimedia modeling capabilities to capture the fate of chemicals across environmental media and micro-environments that could significantly alter actual exposure profiles to toxicants. Combining the information from environmental fate modeling, epidemiologically (experientally) derived data, and toxicokinetic/dynamic models for the estimation of internal exposure can give powerful support to advanced human health risk assessment. The output of this computation layer can be further coupled with gene and protein expression profiles as signatures of exposure to classes of toxicants to derive biologically based dose-response estimates adopting a systems biology approach to risk assessment. An additional dimension of complexity for total exposure and integrated risk modeling is the interactivity of components of chemical mixtures at the molecular level. Toxicants have the potential for additivity, synergism but also for antagonism of toxicity, as exposure-response pathways interact to enhance cellular defense mechanisms and limit metabolism that leads to cell damage. Furthermore, individual susceptibility to chemical toxicity due to varying gene expression has to be taken into account to allow for more targeted human health risk assessment. The answer to these additional challenges lies with the development of homogeneous multimedia models for a wide range of contaminants and socioeconomic/environmental contexts, the extension of toxicokinetic/dynamic models to predict toxicant interactions, and the development of methods for the derivation and explicit consideration of model and data uncertainties both during model construction and during analysis of results. The paper concludes by giving specific examples of successful implementation of the complexity analysis paradigm delineated above and making suggestions for joint action toward the development of the necessary computational infrastructure.

Monitoring Environmental Exposures: Now It’s Personal

Most diseases are thought to arise from the combined effects of genes and the environment. While great strides have been made in our understanding of human genetics, the contribution of environmental exposures to disease remains poorly understood. This lack of understanding impedes real progress in identifying genetically susceptible people whose responses to environmental agents are severe or unique relative to the general population. If identified, targeted prevention and treatment strategies might be applied to these groups, with potentially lifesaving results. Without more personal exposure information, however, scientists have a limited ability to identify pollution-susceptibility genes that elevate disease risk. This mismatch between exposure and genetic research slows the identification of environmental factors that—if altered or removed—could even prevent some diseases from occurring in the first place. “Genes aren’t modifiable,” explains Frederica Perera, a professor of environmental health sciences at the Mailman School of Public Health of Columbia University. “So, these environmental components in diet, food, water, and air are the only ones we can act on for disease prevention.” Exposure is, at a core level, the instance of environmental stimuli such as chemicals, infectious agents, diet, and lifestyle factors interacting with the human body. But studying human exposures poses difficult challenges. Scientists studying the effects of chemical pollutants can’t ethically dose people, so they often base their dose–response estimates on animal models. Further, they base their estimates of human exposure on indirect measures taken from a person’s home or work environment, along with assumptions about individual behaviors that influence the risk of coming into contact with a given pollutant. In a typical study, epidemiologists might use census figures, questionnaires, and general environmental monitoring data (such as samples of water from a household tap or soil in a schoolyard) to estimate how much of a given chemical an individual has come into contact with, and for how long. But while these studies approximate population-level exposures, they provide little information about real exposures to individual people. And because of this, scientists know little about how specific pollutants—particularly in combination with each other, with diet, and with physical activity—affect any individual’s response to the pollutant. With a clear need for progress, exposure assessment has recently come under the spotlight. This year, the NIEHS launched the Exposure Biology Program (EBP), a four-year effort with two overarching goals: to improve exposure assessment technology, and to identify biomarkers for common pathogenic mechanisms that reflect the human response to environmental agents. Such biomarkers could include changes in metabolites, proteins, or DNA that reflect the individual’s genetic susceptibility to environmental harm. “Right now we don’t really understand how exposure levels translate to human health risk, so our goal is to fill that gap,” says EBP coordinator Brenda Weis, a senior scientific advisor at NIEHS. “We need to get a better measure of exposure at the point of human contact, and we need to integrate those measures with biological response measures derived from samples taken directly from exposed people. So, this is a more ‘medical’ approach to exposure assessment—in the sense that measures are taken on a personal level—rather than the broader, ecological approaches we’ve been using so far.”

Environmental Tobacco Smoke Exposure and Risk of Stroke in Nonsmokers: A Review With Meta-analysis

Our objective was to provide a comprehensive review of the epidemiologic evidence relating stroke to exposure to environmental tobacco smoke (passive smoking) in lifelong nonsmokers. In all, 7 prospective, 6 case-control, and 3 cross-sectional studies were identified that provided relevant information. The 16 studies varied considerably in design, exposure indices used, and disease definition. Based on 24 sex-specific relative risk estimates, and using current spousal exposure (or nearest equivalent) as the exposure index, meta-analysis gave an overall estimate of 1.25 (95% confidence interval 1.16-1.36), with no significant heterogeneity. There was no significant heterogeneity by sex, year of publication, fatality, or disease end point. There was some indication that relative risks were less elevated in prospective studies and in US or European studies. No elevation was seen for subarachnoid hemorrhage. The estimate was similar using ever rather than current exposure, or total rather than spousal exposure. Adjustment for risk factors other than age had no material effect on the estimate. Eight studies provided dose-response estimates, the combined relative risk for the highest level of exposure being 1.56 (95% confidence interval 1.34-1.82). Many studies have evident weaknesses, and recall bias and particularly publication bias are major concerns. Currently, the association is only suggestive of a possible causal relationship. A clearer picture could be obtained using data from existing very large prospective studies that have already provided risk estimates for passive smoking and heart disease, and by conducting large, well-designed studies of incident stroke.

Some Statistical Implications of Dose Uncertainty in Radiation Dose–Response Analyses

Statistical dose-response analyses in radiation epidemiology can produce misleading results if they fail to account for radiation dose uncertainties. While dosimetries may differ substantially depending on the ways in which the subjects were exposed, the statistical problems typically involve a predominantly linear dose-response curve, multiple sources of uncertainty, and uncertainty magnitudes that are best characterized as proportional rather than additive. We discuss some basic statistical issues in this setting, including the bias and shape distortion induced by classical and Berkson uncertainties, the effect of uncertain dose-prediction model parameters on estimated dose-response curves, and some notes on statistical methods for dose-response estimation in the presence of radiation dose uncertainties.

A Bayesian analysis of mouse infectivity data to evaluate the effectiveness of using ultraviolet light as a drinking water disinfectant

Modelling disinfectant performance using Bayesian hierarchical methods can overcome problems with traditional methods and lead to improved estimates. Animal and cell-culture assays are used to estimate the degree of inactivation of a microorganism produced by a given disinfectant dose. Assay data traditionally are analyzed with logistic model or most probable number (MPN) method. These methods are limited particularly when assays show all (or no) animals or cells to be infected—estimates are reported as greater than (or less than) a measurement limit (i.e., censored data). The proposed Bayesian approach (1) properly models the propagation of uncertainty through the data analysis/modelling process, resulting in reduced model uncertainty, and (2) uses appropriate probability distribution models for the response variables, avoiding the censored data problem and more accurately describing statistical error when estimating dose–response behavior. This paper applies the Bayesian hierarchical models to logistic and MPN data from published papers for the ultraviolet (UV) inactivation of Cryptosporidium. Results are compared to those from three alternative models. The Bayesian model estimates a significantly lower UV dose for a given level of Cryptosporidium inactivation than the alternative models, due mainly to the reduced model uncertainty.

Associations of pesticides, HCV, HBV, and hepatocellular carcinoma in Egypt

The rate of hepatocellular carcinoma (HCC) is increasing in Egypt where the major risk factors are chronic infections with hepatitis B and C viruses (HBV and HCV). A major segment of the population is employed in agriculture, raising the possibility that exposure to pesticides is an additional risk factor for HCC. The objective of this study is to investigate pesticides as environmental risk factors for HCC while taking into account viral risk factors. We conducted a case-control study of 236 subjects with confirmed HCC recruited from the National Cancer Institute, Cairo University, Egypt, and 236 controls matched on sex, age group and urban–rural status recruited from orthopedic department, Cairo University Hospital, Egypt. Patients who agreed to participate signed a consent form, answered a questionnaire and gave a blood sample for hepatitis virus testing. The manuals of the Ministry of Agriculture for approved use and type of pesticides since 1965 were linked to the questionnaire data for types of crops and pests that the subject had to combat, to attribute specific pesticides that were used by each subject. Subjects also reported duration of the exposure (years). Case-control comparisons in these data were stratified by sex, age group, and urban vs. rural residence. Data were analyzed using unconditional logistic regression models adjusting for age, HCV RNA, and current hepatitis B infection. Among rural males, the adjusted odds ratio (OR) for organophosphorus compounds was 2.7 (95% CI=1.3–5.3) and for carbamates it was 2.9 (95% CI=1.4–5.8). No statistically significant associations between HCC and pesticides were observed for urban males or for females. As expected, the strongest risk factors for HCC in this study were HCV RNA (OR=16–17) and current HBV infection (OR=27–28). This study therefore suggests that exposures to organophophorus and carbamate pesticides are additive risk factors to current HCV and HBV infection among rural males. Future investigation should address the possible hepatocarcinogenicity of pesticides using biomarkers of exposure and other techniques to better estimate dose–response relationships.

Arsenic in groundwater in Bangladesh: A geostatistical and epidemiological framework for evaluating health effects and potential remedies

This paper examines the health crisis in Bangladesh due to dissolved arsenic in groundwater. First, we use geostatistical methods to construct a map of arsenic concentrations that divides Bangladesh into regions and estimate vertical concentration trends in these regions. Then, we use census data to estimate exposure distributions in the regions; we use epidemiological data from West Bengal and Taiwan to estimate dose response functions for arsenicosis and arsenic‐induced cancers; and we combine the regional exposure distributions and the dose response models to estimate the health effects of groundwater arsenic in Bangladesh. We predict that long‐term exposure to present arsenic concentrations will result in approximately 1,200,000 cases of hyperpigmentation, 600,000 cases of keratosis, 125,000 cases of skin cancer, and 3000 fatalities per year from internal cancers. Although these estimates are very uncertain, the method provides a framework for incorporating better data as it becomes available. Moreover, we examine the remedy of drilling deeper wells in selected regions of Bangladesh. By replacing 31% of the wells in the country with deeper wells the health effects of drinking groundwater arsenic could be reduced by approximately 70% provided that arsenic concentrations in deep wells remain relatively low.

Improving Risk Assessment: Research Opportunities in Dose Response Modeling to Improve Risk Assessment

Substantial improvements in dose response modeling for risk assessment may result from recent and continuing advances in biological research, biochemical techniques, biostatistical/mathematical methods and computational power. This report provides a ranked set of recommendations for proposed research to advance the state of the art in dose response modeling. The report is the result of a meeting of invited workgroup participants charged with identifying five areas of research in dose response modeling that could be incorporated in a national agenda to improve risk assessment methods. Leading topics of emphasis are interindividual variability, injury risk assessment modeling, and procedures to incorporate distributional methods and mechanistic considerations into now-standard methods of deriving a reference dose (RfD), reference concentration (RfC), minimum risk level (MRL) or similar dose-response parameter estimates.

Improving Risk Assessment: Priorities for Epidemiologic Research

The Epidemiology Work Group at the Workshop on Future Research for Improving Risk Assessment Methods, Of Mice, Men, and Models, held August 16 to 18, 2000, at Snowmass Village, Aspen, Colorado, concluded that in order to improve the utility of epidemiologic studies for risk assessment, methodologic research is needed in the following areas: (1) aspects of epidemiologic study designs that affect doseresponse estimation; (2) alternative methods for estimating dose in human studies; and (3) refined methods for dose-response modeling for epidemiologic data. Needed research in aspects of epidemiologic study design includes recognition and control of study biases, identification of susceptible subpopulations, choice of exposure metrics, and choice of epidemiologic risk parameters. Much of this research can be done with existing data. Research needed to improve determinants of dose in human studies includes additional individual-level data (e.g., diet, co-morbidity), development of more extensive human data for physiologically based pharmacokinetic (PBPK) dose modeling, tissue registries to increase the availability of tissue for studies of exposure/dose and susceptibility biomarkers, and biomarker data to assess exposures in humans and animals. Research needed on dose-response modeling of human studies includes more widespread application of flexible statistical methods (e.g., general additive models), development of methods to compensate for epidemiologic bias in dose-response models, improved biological models using human data, and evaluation of the benchmark dose using human data.There was consensus among the Work Group that, whereas most prior risk assessments have focused on cancer, there is a growing need for applications to other health outcomes. Developmental and reproductive effects, injuries, respiratory disease, and cardiovascular disease were identified as especially high priorities for research. It was also a consensus view that epidemiologists, industrial hygienists, and other scientists focusing on human data need to play a stronger role throughout the risk assessment process. Finally, the group agreed that there was a need to improve risk communication, particularly on uncertainty inherent in risk assessments that use epidemiologic data.

Statistical models of health risk due to microbial contamination of foods

Between 6 million and 33 million cases of food-related illness are estimated to occur in the United States each year, with about 5000 episodes resulting in death. Growing concerns about the safety of food prompted the National Food Safety Initiative of 1997, the goal of which is to reduce the incidence of illness caused by food-borne pathogens. A key component of the food safety initiative is the improvement of farm-to-table risk assessment capabilities, including the development of improved dose-response models for estimating risk. When sufficient data are available, allowable contamination levels of specific micro-organisms in food are established using dose-response models to predict risk at very low doses based on experimental data at much higher doses. This necessitates having reliable models for setting allowable exposures to food-borne pathogens. While only limited data on relatively few micro-organisms that occur in food are available at present for dose-response modeling and risk estimation, still none of the two-parameter models proposed so far, including the popular Beta-Poisson (BP) model, appears to be completely satisfactory for describing and fitting all of the present data (Holcomb et al., 1999). The Weibull–Gamma (WG) model is the only three-parameter model that has been proposed to date. In this paper, new competitive three-parameter models are derived, using a formulation that can be parameterized to represent statistical variation with respect to the dose of micro-organism received by the host and the host’s susceptibility to infection. Parameters of the models are estimated using the maximum likelihood method. Experimental data on several common microbial contaminants in food are used to illustrate the methodology.

Thyroid Cancer Following Scalp Irradiation: A Reanalysis Accounting for Uncertainty in Dosimetry

In the 1940s and 1950s, over 20,000 children in Israel were treated for tinea capitis (scalp ringworm) by irradiation to induce epilation. Follow-up studies showed that the radiation exposure was associated with the development of malignant thyroid neoplasms. Despite this clear evidence of an effect, the magnitude of the dose-response relationship is much less clear because of probable errors in individual estimates of dose to the thyroid gland. Such errors have the potential to bias dose-response estimation, a potential that was not widely appreciated at the time of the original analyses. We revisit this issue, describing in detail how errors in dosimetry might occur, and we develop a new dose-response model that takes the uncertainties of the dosimetry into account. Our model for the uncertainty in dosimetry is a complex and new variant of the classical multiplicative Berkson error model, having components of classical multiplicative measurement error as well as missing data. Analysis of the tinea capitis data suggests that measurement error in the dosimetry has only a negligible effect on dose-response estimation and inference as well as on the modifying effect of age at exposure.