DEVELOPMENTAL AND REPRODUCTIVE TOXICITY OF ENVIRONMENTAL CONTAMINANTS

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: HATCH
• Reporting Frequency: Annual
• Accession No.: 0223790
• Project Start Date: Oct 1, 2010
• Project End Date: Sep 30, 2015
• Program Code: [(N/A)]- (N/A)

Recipient Organization
MICHIGAN STATE UNIV
(N/A)
EAST LANSING,MI 48824

Performing Department
Animal Science

Non Technical Summary
The long-term goal of the project is to achieve a better understanding of the impact of animal agriculture on the modern society by integrating the risks and benefits in three aspects: economics, environmental protection, and human health. At this stage of the project, we will focus on the potential health risks associated with animal agriculture operations. Objectives of the project are (I) to identify animal production practices associated with low human health risk and (II) to use mathematical to examine the possible health risks associated with animal production. The specific aims are (1) to identify environmental hazards that may impact on infant health and assess possible relationships to animal agricultural, (2) to conduct toxicity assessment of hazardous agents that are associated with animal operations, (3) to examine infant health factors that may be directly or indirectly impacted by animal operations, (4) to construct and validate quantitative models for the risk analyses, and (5) to identify production methods associated with low health risks and high benefits, and to encourage animal agriculture to adopt low-health-risk production practices. The expected outputs will be information on the potential hazards and health risks of substances from animal production and assessment models that may be used to identify low risk production methods. In Specific Aims 1, we will determine whether a particular chemical is or is not causally linked to particular negative outcomes of pregnancy. For Specific Aim 2 The toxicity assessment will be conducted for birth-outcomes using methods described in the Guidelines for Reproductive Toxicity Risk Assessment (EPA, 1996) and Guidelines for Developmental toxicity Risk Assessment (EPA, 1991). For Specific Aims 3 and 4, types and numbers of livestock in Michigan from 2001 to 2010 will be obtained from the USDA National Agricultural Statistics Service (NASS) Michigan Field Office. Locations and types of other potential sources of contaminants will be obtained from Michigan Department of Natural Resources & Environment (MDNRE), US EPA Superfund Program, and the Toxics Release Inventory Program. Michigan birth and infant death data, from 2001 to 2010, will be obtained from the Michigan Department of Community Health (MDCH). Exposure The model will be constructed with the consideration of hazardous substances potentially harmful to maternal and infant health. All births and deaths in the State of Michigan from 2001 to 2010, about 130,000 per year, will be used in this study to maximize the quality of the estimates. Over the 10 years of the proposed study period, 2001-2010, we estimated that there will be approximately 1,300,000 newborns and 800,000 women who gave birth. For the livestock data files, there will be about 19,000 principal livestock operators. For Specific Aim 5, the risk and benefit models will be constructed using results obtained in studies conducted under Specific Aims 1-4.

Animal Health Component

(N/A)

Research Effort Categories

Basic

(N/A)

Applied

(N/A)

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
723	3910	1170	40%
903	3910	1150	15%
901	3910	1170	30%
723	3910	1150	15%

Knowledge Area
901 - Program and Project Design, and Statistics; 903 - Communication, Education, and Information Delivery; 723 - Hazards to Human Health and Safety;

Subject Of Investigation
3910 - Cross-commodity research--multiple animal species;

Field Of Science
1150 - Toxicology; 1170 - Epidemiology;

Keywords

animal production

environmental contaminants

production methods

pregnancy outcomes

infant mortality

risk analysis

risk and benefit analysis

Goals / Objectives
The long-term goal of the project is to achieve a better understanding of the impact of animal agriculture on the modern society by integrating the risks and benefits in three aspects: economics, environmental protection, and human health. At this stage of the project, we will focus on the potential health risks associated with animal agriculture operations. Objectives of the project are (I) to identify animal production practices associated with low human health risk and (II) to demonstrate a scale and space relevant quantitative approach to examine the possible health risks associated with animal production. The specific aims are (1) to identify environmental hazards that may impact on birth-outcomes and assess possible relationships to animal agricultural, (2) to conduct toxicity assessment of hazardous agents that are associated with animal operations, (3) to examine explanatory variables, i.e. predictors, of birth outcomes at the individual and county levels and identify those that may be directly or indirectly impacted by animal operations, (4) to construct and validate quantitative models for the risk analyses, and (5) to identify production methods associated with low health risks and high benefits, and to encourage animal agriculture to adopt low-health-risk production practices. The expected outputs will be information on the potential hazards and health risks of substances from animal production and assessment models that may be used to identify low risk production methods.

Project Methods
In Specific Aims 1, we will determine whether a particular chemical is or is not causally linked to particular negative outcomes of pregnancy. For Specific Aim 2 The toxicity assessment will be conducted for birth-outcomes using methods described in the Guidelines for Reproductive Toxicity Risk Assessment (EPA, 1996) and Guidelines for Developmental toxicity Risk Assessment (EPA, 1991). For Specific Aims 3 and 4, types and numbers of livestock in Michigan from 2001 - 2010, and the geocoded residences of principal livestock operators will be obtained from the USDA National Agricultural Statistics Service (NASS) Michigan Field Office. Locations and types of other potential sources of contaminants will be obtained from Michigan Department of Natural Resources & Environment (MDNRE), US EPA Superfund Program, and the Toxics Release Inventory Program. Michigan birth and infant death data, from 2001 to 2010, will be obtained from the Michigan Department of Community Health (MDCH). Exposure models will be constructed based on Euclidean distances from the maternal and infant residences to the locations of livestock operations and the amount of contaminants. The model will be constructed with the consideration of hazardous substances potentially harmful to maternal and infant health. (Shih et al., 2006; Von Essen et al., 2010). For data analysis, we will begin by estimating separate models for the relationship between livestock numbers and birth outcomes that include only the exposure to livestock, providing a baseline for the relationship between the exposure and birth outcomes. We will then estimate models for birth outcomes that include the full set of explanatory variables. we will use factor analysis to combine measures into meaningful summary measures (Wang, 2005). Prior to regression analysis, the correlations among the explanatory variables will be examined. Preliminary multiple regression models will be constructed and evaluated before the appropriate models are selected to describe the relationship between the pregnancy outcomes and about the hypothesized predictor variables. All births and deaths in the State of Michigan from 2001 to 2010, about 130,000 per year, will be used in this study to maximize the quality of the estimates. Over the 10 years of the proposed study period, 2001-2010, we estimated that there will be approximately 1,300,000 newborns and 800,000 women who gave birth. For the livestock data files, there will be about 19,000 principal livestock operators. For regression models that include 10 predictor variables in the analysis, using the expected sample size of approximately 130,000 births per year in Michigan, we may be able to detect about 5-10% changes in infant mortality rate in multiple regression analyses, at the alpha level 0.03 and power of 0.8. For Specific Aim 5, the risk and benefit models will be constructed using results obtained in studies conducted under Specific Aims 1-4.

Progress 10/01/10 to 09/30/15

Outputs
Target Audience:The knowledge was delivered through workshops, seminars, publications, and formal classroom instuctin. The audience include agricultural business groups, university students and researchers, extension educators, law professionals, and the general public. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Training activities: Emily Groll, undergraduate student in Biological Science, worked on this project as a research assistant in the area of environmental contaminants. Jessica Henderson, undergraduate student in Human Biology, worked on this project as a research assistant in the area of data management and geocoding. Samuel Nyquist, undergraduate student in Geography, worked on this project as a research assistant in the area of geospatial analysis. Charlene Arney, undergraduate student in Geography, worked on this project as a research assistant in the area of geospatial analysis. Dawn Chi Chang, graduate student in Measurement and Quantitative Methods, worked on this project as research assistant in the area of geocoding and quantitative analysis. How have the results been disseminated to communities of interest?The knowledge gain from this project was deliveredin workshops, seminars, and formal classroom instruction. the audience include extension educators, livestock producers, organic livestock producers, law school faculty and students, environmental activists, faculty and students in MSU's School of packaging, and MSU students in Journalism, Crop and Soil Science, and Environmental Engineering. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? We examined the major maternal factors that are likely to cause or influence infant health, including maternal and paternal education, race, marital status, and socioeconomic factors. We also examined delivery facility and time of birth, infant sex, birth weight, maternal age, prenatal care index, length of pregnancy, and tobacco and alcohol use information. We examined material risk factors, including pre-pregnancy diabetes, pregnancy diabetes, chronic hypertension, pregnancy hypertension, previous preterm birth, pregnancy resulted from infertility treatment, pregnancy complications, delivery method, and infectious diseases that could potential impact on fetal development. Assessing the risk of Michigan infant mortality in relation to living near agricultural animal operations. Using Michigan live births and infant death records and distance-weighted estimation of potential exposure to agricultural animals, our analyses demonstrated no association between infant mortality and exposure to agricultural animals. (1) We identified six types of potentially hazardous substances associated with agricultural animal operations, Ammonia, nitrogen dioxide, nitric acid, hydrogen sulfide, endotoxins, pathogens, and particular matters.(2) We assessed potential health effects of ammonia, nitrogen dioxide, and hydrogen sulfide on human health in the surrounding residents. For ammonia, its concentration in the surrounding area, 50-150 meter away from an operation, is unlikely to exceed the US EPA's reference concentration (RfC) of 0.1 ppm. Nitrogen dioxide concentration in agricultural intensive region could reach 0.01 ppm, which is below EPA's primary and secondary standard of 0.053 ppm. Hydrogen sulfide concentration may reach 1 ppm or greater, which exceeds the odor threshold of 0.00047 ppm, but much lower than the level for human eye irritation (10-20 ppm) or causing lesions of nasal mucosa. (3) For the maternal health and infant mortality databases, appropriate and necessary exclusions, such as deaths caused by definitively non-environmental factors, are identified and coded. In this study, "definitively non-environmental cause of death" is defined by the underline causes of death from accidents, external injuries, hereditary diseases, medical treatments, and congenital defects caused by maternal use of tobacco or maternal sexually transmitted diseases. All other underline causes of infant death, including congenital malformation and infant or maternal infection, are coded as potential deaths from environmental factors. Driving time between maternal residence and the nearest hospital is calculated using ArcGIS. We have built multivariate logistic regression models for each year to examine risk factors associated with infant deaths. The independent variables examined include alcohol and tobacco use, plurality, race, source of payment, sex, driving time to the nearest hospital, month prenatal care began, maternal birth place, marital status, number of prenatal visits, maternal sexually transmitted diseases, and maternal education, age, and body mass index. (4) The Euclidean distances between maternal residents to animal operators are measured using ArcGIS. Maternal/ infant exposure to agricultural animal operations is modeled using the Euclidean distances to animal operators and weighted by the number and type of animals, using the Gaussian models. (5) Because we observed no effect on infant mortality, no intervention recommendation is made based on the results of this project. Data collected: Maternal health, birth, and infant mortality data are obtained from Michigan Department of Community Health (MDCH). For the study period, a total of 1,278,139 records were collect. Operator's locations and numbers of livestock numbers were obtained from USDA National Agricultural Statistic Service (NASS), there are about 19,000 operators for each of the study years. For the livestock operators, we obtained the latitudes and longitudes of animal operators and types and numbers of animals. For each birth record, we obtained latitudes and longitudes of the residences of the mothers. For maternal risk factors, we obtained race, ancestry, education, marital status, source of payment, the month during pregnancy when prenatal care began, number of prenatal visits during pregnancy,calculated and estimated weeks of gestation,maternal tobacco and alcohol use, date quit smoking, others in household smoke, maternal birthday and age, mother transferred prior to delivery, hospital mom transferred from,and other health risks, including anemia, cardiac disease, diabetes prior to this pregnancy, diabetes diagnosis in this pregnancy,genital herpes, oligohydramnios, hemoglobin, chronic and pregnancy hypertension, eclampsia, incomplete cervical dilation, Rh sensitivity, vaginal bleeding during this pregnancy prior to the onset of labor, uterine bleeding, pregnancy resulted from infertility treatment, mother had a previous cesarean delivery, number of previous c-sections, drug abuse, gonorrhea, syphilis, genital herpes, Chlamydia, hepatitis C, Group B streptococcus, and HIV risk. We also obtained data that indicate obstetric procedures performed during pregnancy or labor andcharacteristics of the process of labor. For infant health, we obtained data of sex, race, infant birth and death dates, birth time, birth weights, types of birth defects, child transferred,birth place, county of birth,state and county of maternal residence, infant transferred to neonatal intensive care unit, cause and related causes of death, method of delivery (C-section, forceps, or vacuum), plurality, multiple birth order, and APGAR scores, abnormal conditions of newborn, assisted ventilation required immediately following delivery, assisted ventilation required for more than six hours, NICU admission, newborn given surfactant replacement therapy, antibiotics received by the newborn for suspected neonatal sepsis, seizure or serious neurologic dysfunction, significant birth injury (skeletal fracture(s), peripheral nerve injury, and/or soft tissue/solid organ hemorrhage which requires intervention. Summary statistics and discussion of results: Using logistic regression on infant death, adjusted by race, multiple birth, maternal age, tobacco smoking, payment methods maternal education, and maternal birth place, the odds of infant death range between 0.999 and 1.000, with confidence intervals from 0.996-1.003. None of the results was statistically significant. The results indicate that exposure to agricultural animals has no effect on infant death. We also conduct the same analysis for urbanized counties (Wayne, Oakland, Macomb, Kent, and Genesee) and other less urbanized counties,and obtained similar results. For babies born to black mothers, the odds of infant mortality were 2.66 times greater than the odds for babies born to white mothers, p value <.001, with 95% confidence interval 2.316-3.057. For twin babies, the odds of mortality were 4.92 times greater than the odds of singletons, p. value <.001, with 95% confidence interval 4.081-5.946. For babies born to mothers who smoked, the odds of mortality were 1.565 times greater than the odds for babies born to mothers who did not smoke, p <.001, with 95% confidence interval 1.318-1.858. For babies born to mothers who smoked and quit smoking during pregnancy, the odds of mortality were 0.949, p < .798, with 95% confidence interval 0.730-1.271. Thus, quitting smoking during pregnancy is an effective intervention to minimize infant mortality for smokers. Key outcomes or other accomplishments realized: (a) There is no evidence for any effect of exposure to agricultural animals on infant mortality in Michigan. (b) Tobacco smoking, twin pregnancy, and babies born to black mothers are significant risk factors for infant mortality in Michigan.

Publications

• Type: Other Status: Published Year Published: 2011 Citation: Schweihofer, J and Chou, K. Impacts of livestock chewing on treated wood, Michigan State University Extension News for Agriculture, May 20, 2011
• Type: Other Status: Published Year Published: 2011 Citation: Schweihofer, J and Chou, K. Impacts of livestock chewing on treated wood, Sheep and Goat Guide newsletter, Penn State Extension
• Type: Book Chapters Status: Published Year Published: 2014 Citation: Chou, K. and J. Henderson (2014) Endocrine System. In: Wexler, P. (Ed.) Encyclopedia of Toxicology, 3rd edition vol2. Elsevier Inc., Academic Press, pp. 332-340
• Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: Environmental concerns of poultry confinement in Symposium on Poultry Confinement, Invited Speaker and Panelist, College of Law, Michigan State University, April 4, 2013
• Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: A review of health impacts in communities and relationship to living near CAFOs, Invited Speaker, Co-presenters: Melissa May & Roy Black, Communities and Livestock Conference, Michigan State University, April 23, 2013
• Type: Conference Papers and Presentations Status: Published Year Published: 2015 Citation: Toxins in Risk Assessment, Quantitative Microbial Risk Assessment Interdisciplinary Instructional Institute, Michigan State University, August 9, 2015
• Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: Bisphenol A: Outlook, Invited speaker, School of Packaging, Michigan State University, October 31, 2013
• Type: Conference Papers and Presentations Status: Published Year Published: 2015 Citation: Sugar and transfat -- nutrients or toxicants? Cheng Shiu University, Taiwan, May 21, 2015

Progress 10/01/13 to 09/30/14

Outputs
Target Audience: Nothing Reported Changes/Problems: We did not expect the approximation of using operators’ addresses to estimate the location of animals to exceed the distance from which the animal locations could possibly have any effects on infant mortality. A new set of geocoded databased using the operations’ addresses has been generated. The new dataset will be applied in the exposure models to generate a new set of exposure index. The geocoding process and data analyses took substantially longer time than before, because, for data security reasons, NASS computers for non-USDA guest users can no longer being on-line in 2014. What opportunities for training and professional development has the project provided? Two graduate students participated in the study. This project democrats the state-of-art of using GIS and other geospatial methods for environmental health research. It also provides a platform to demonstrate the needs for and benefits of interdisciplinary studies. How have the results been disseminated to communities of interest? The results of the study has been shared with other faculty members. The methods developed in this project is being considered for developing research proposal for other environmental health studies. What do you plan to do during the next reporting period to accomplish the goals? The exposure models will be further examined may be modified using the datasets of Michigan agricultural animal operations. The associated between animal operation and infant mortality will be examined using logistic regression analyses. A manuscript will be prepared to report the results.

Impacts
What was accomplished under these goals? Progress, Oct 1, 2013-Sept 30, 2014: To study the relationship between infant mortality and exposure to agricultural animals, four distance- weighted models were used to produce potential exposure indexes (PEX). Our results suggested that the errors of using operators’ addresses to estimate the location of animals might exceed the distance from which the animal locations could possibly have any effects on infant mortality. A new set of geocoded databased has been generated to use the operations’ addresses, instead of the operators’ addresses, to calculated the PEX. This effort, in part, contributes to the progress of achieving Specific Aims (3) and (4).

Publications

Progress 01/01/13 to 09/30/13

Outputs
Target Audience: The target audience I reached during the reporting period includes animal agricultural producers, environmental and animal welfare activists, students and faculty of MSU College of Law, other MSU faculty members and students, extension educators, and agricultural industry representatives. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? This project has provided several one-on-one learning opportunities for students in the field of environmental/public health and geospatial analyses. One of the project’s undergraduate research assistant was admitted to medical school and two others are now employed by government agencies as geospatial analysts. How have the results been disseminated to communities of interest? The results are presented in the Symposium on Poultry Confinement, MSU College of Law, on April 4, 2013, and at Communities and Livestock Conference, Michigan State University, on April 23, 2013. What do you plan to do during the next reporting period to accomplish the goals? Manuscripts for publishing the results are in preparation. Additionall hypotheses onenvironmentalimpacts on maternal and infant health will be constructed and tested.

Impacts
What was accomplished under these goals? (a) Identifying known hazardous substances associated with poultry operations and assessing potential impacts of the substances on health for constructing species-specific hypotheses. Two of the substances, ammonia and nitrogen dioxide, are less like to impact on the health of residents in the neighborhood. Their concentrations in air, however, may exceed the US EPA inhalation reference concentrations in boiler houses, depending on temperature, ventilation, production stages, etc. Hydrogen sulfide may exceed the inhalation reference concentration at the property line. This activity meets a part of the specific Aims (1), (3) and (4). (b) Assessing the risk of Michigan infant mortality in relation to living near agricultural animal operations. Using Michigan annual live births and infant death records, 2006-2009, and distance weighted estimation of potential exposure to agricultural animals, logistic regression analyses demonstrated no association between infant mortality and exposure to agricultural animals. This activity meets a part of the specific aims (3) and (4).

Publications

Progress 01/01/12 to 12/31/12

Outputs
OUTPUTS: Michigan hospital, births & death data, agricultural animal numbers and locations for the years 2006-2010 have been collected, compiled, and geocoded. During the five years, there are a total of 605,964 births, and 4,574 infant died under the age of 12 months, with yearly infant mortality rates ranging from 6.9/1000 births to 8.4/1000 births. Each year, there are about 19,000 principal agricultural animal operators in our analyses. Appropriate and necessary exclusions, such as deaths caused by definitively non-environmental factors, are identified and coded. In this study, "definitively non-environmental cause of death" is defined by the underline causes of death from accidents, external injuries, hereditary diseases, medical treatments, and congenital defects caused by maternal use of tobacco or maternal sexually transmitted diseases. All other underline causes of infant death, including congenital malformation and infant or maternal infection, are coded as potential deaths from environmental factors. Driving time between maternal residence and the nearest hospital is calculated using ArcGIS. We have built multivariate logistic regression models for each year to examine risk factors associated with infant deaths. The independent variables examined include alcohol and tobacco use, plurality, race, source of payment, sex, driving time to the nearest hospital, month prenatal care began, maternal birth place, marital status, number of prenatal visits, maternal sexually transmitted diseases, and maternal education, age, and body mass index (BMI). The Euclidean distances between maternal residents to animal operators are measured using ArcGIS. Maternal/ infant exposure to agricultural animal operations is modeled using the Euclidean distances to animal operators and weighted by the number and type of animals, using the Gaussian models. PARTICIPANTS: The collaborators at MSU are Roy black, Agricultural Food & Resource Economics, [e-mail]; Wendy Powers, Animal Science, [e-mail]; and Melissa Millerick-May, Medicine, [e-mail]. The study is also conducted in collaboration with David Kleweno, USDA, [e-mail]. A MSU student, Jessica Henderson of Human Biology, [e-mail], also participated in the study. She is now a medical student of MSU's College of Osteopathic Medicine. Two students from the Department of Geography have worked with us on this project: Sam Nyguist, [e-mail], and Charlene Arney, [e-mail]. TARGET AUDIENCES: Results of this study will be presented at local, national, and international conferences and published in the journals related to animal agriculture and environmental health sciences. The target audiences are agriculture and public health practitioners, policy makers, and environmental health researchers. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
We have examined the association between infant mortality and exposure to swine operators using four different standard deviations (100, 200, 400, and 600 meters) in the Gaussian formula. Using the 2006 to 2009 data files, the preliminary results of our logistic regression analyses indicate no relationship between infant mortality and exposure to swine operations, either before or after adjusting with the identified risk factors, including maternal tobacco use, others smokers in the house, sources of payment, and maternal age , education, and body mass index (BMI).

Publications

• No publications reported this period

Progress 01/01/11 to 12/31/11

Outputs
OUTPUTS: We have imported Michigan agricultural animal datafiles from 2006 to 2010. We have geocoded the 2006 and 2007 datafiles. Currently, we are examining the use of animal conversion units and the uncertainties associated with different conversion approaches. We are also in the process of developing geospace relevant exposure assessment methods using ArcGIS. We are analyzing the association between infant mortality and changes in animal numbers at the county level in Michigan. We have imported Michigan birth and death records from 2000-2009, a total of over one million birth records. The datafiles are purchased from Michigan Department of Community Health. We have also identified the types of 212 Michigan hospitals by the presence or absence of obstetric, birthing, and neonatal intensive care units. We have geocoded the hospital locations. In addition, we have examined the ICD10 codes and re-coded over 11,000 causes of death into 19 categories, based on each ICD code's etiology and their relevance to potential environmental contaminants. We are examining data fidelity and potential confounding variables in the relationship between infant mortality and agricultural animal production. Some of the errors in birth and death records, as we receive from the health department, may be the results of reporting and recording errors, while others may be the results of patients' inattention to their own health conditions. Examples of the latter include the presence or absence of pre-pregnancy hypertension and herpes, the pre-pregnancy body weight, and sexually transmitted diseases. In addition, many variables in the birth datafiles depend entirely on voluntary information provided by patients. Studies have shown that these types of data contain significant amounts of non-response (i.e. response marked as "unknown" in the records) and false information. It is known that a significant number of patients may forge information that supports healthy habits and morally acceptable behaviors, while others may do so to avoid legal consequences. Examples of these types of non-response or false response include alcohol use and drug abuse. PARTICIPANTS: The collaborators at Michigan State University (MSU) are Roy black, Agricultural Food & Resource Economics, [e-mail]; Wendy Powers, Animal Science, [e-mail]; and Melissa Millerick-May, Medicine, [e-mail]. The study is also conducted in collaboration with David Kleweno, USDA National Agricultural Statistic Service (NASS), [e-mail] and Gerald Tillman, Michigan Field Office, USDA NASS, [e-mail]. As for training and professional development, a MSU student, Jessica Henderson of Human Biology, [e-mail], also participates in the study. There are opportunities for additional student participations. TARGET AUDIENCES: Results of the study will be presented at local, national, and international conferences and published in the journals related to animal agriculture and environmental health sciences. The target audiences are agriculture and public health practitioners, policy makers, and environmental health researchers. PROJECT MODIFICATIONS: None.

Impacts
In Michigan, the 15 counties with the highest animal units are Kent, Calhoun, St Joseph, Oceana, Kalamazoo, Hillsdale, Branch, Sanilac, Clinton, Gratiot, Ionia, Huron, Cass, Ottawa, and Allegan. The 15 counties with the lowest animal units are Baraga, Keweenaw, Alger, Gogebic, Crawford, Iron, Kalkaska, Otsego, Oakland, Roscommon, Luce, Wayne, Benzie, Houghton, and Schoolcraft. Preliminary observations indicate that, in the 10 years of 2000 to 2009, infant mortality rates (IMR) in Michigan (average 8.12 deaths per 1000 births) are consistently higher than that of the national average (6.55/1000). While the IMRs in the US improved by decreasing 0.6 deaths per 1000, Michigan improved by decreasing ~0.8 deaths per 1000. Using the 2007 animal numbers, the annual IMR of the top 15 Michigan counties with the highest numbers of animal units are consistently lower than the 15 Michigan counties with the lowest animal units, with the except of 2001. When examining the data fidelity and potential confounding variables, using the 2009 birth and death records, we observed that 0.43% of moms used alcohol during pregnancy and 1.49% moms were marked as unknown. The IMR was elevated for alcohol users (12.0/1000) and unknown usages (15.4/1000). For the non-users, IMR is 7.6/1000. If the unknowns were treated as missing data in the statistic analysis, alcohol use would have no statistically significant impact on IMR. The unknown alcohol usages, therefore, should not be treated as missing values. For the variables of maternal infectious diseases, herpes itself does not appear to be a risk factor for infant mortality. Mothers whose records indicate having herpes, in fact, had lower IMR than the population mean: IMR of mothers with herpes is 6.2/1000 comparing with the population mean, 7.7/1000, or the mean of mothers without herpes, 7.4/1000. On the other hand, the unknown category is associated with elevated IMR, 23.7/1000. These observations indicate that the "unknown" responses of herpes and similar variables should not be treated as missing values in the analysis.

Publications

• No publications reported this period