Increases in bicycling in the United States results in increased exposure to crashes and injuries. This research focuses on the factors involved in bicycle crashes in the United States and the state of Iowa. Data from the U.S. Nationwide Inpatient Sample and the Iowa Department of Transportation were used to address three aims: 1) estimate the burden and examine the outcomes of bicycle crashes resulting in hospitalizations nationwide by motor vehicle involvement, 2) describe how bicycle motor vehicle crashes vary by intersection and non-intersection in Iowa, and 3) identify the impact of on-road bicycle facilities on bicycle-motor vehicle crashes in Iowa. Using the U.S. Nationwide Inpatient Sample, years 2002-2009, the estimated annual burden of injury from bicycle-related hospitalizations equated to a billion dollars in hospital charges, over 100,000 days in the hospital, and over 300 in-hospital deaths. We also found that bicycling crashes involving motor vehicles had more hospital charges, longer stays, and greater odds of in-hospital death. We also used the Iowa Department of Transportation crash database, 2001 to 2010, to examine risk factors for bicycle-motor vehicle (BMV) crash locations. We found that BMV crashes involve risk factors at person, crash, environment, and population levels and these vary by intersection and non-intersection. Compared to intersections, non-intersection crashes were more likely to involve young bicyclists (0-9 years), locations outside city limits, with driver vision obscured, reduced lighting on the roadway and less likely involve failure to yield right of way. Finally, we conducted a case site-control site study in Iowa, using crash data from 2007 to 2010 to investigate the impact of pavement markings (bicycle lanes and shared lane arrows) and bicycle-specific signage on crash risk. Our results suggest that bicycle facilities are protective against crashes, with the most protective being the combination of both pavement markings and signage, followed by pavement markings alone, and then signage alone. This project shows that bicycling carries a large burden of injury in the United States and that there are many contributing factors to bicycle crashes. It also provides evidence suggesting that infrastructure changes can decrease crash occurrence and there opportunities to intervene at other levels (e.g., person factors) to have an even greater impact overall.

The objective of this research was to evaluate the relationships between greater adiposity and bone development during adolescence. Bone was evaluated from age 11 to 17 years in the Iowa Bone Development Study using peripheral quantitative computed tomography (pQCT). Body composition (fat and lean mass) was estimated by dual energy x-ray absorptiometry (DXA). The first research aim evaluated the associations between greater overall adiposity and subsequent maturation and bone strength in 135 girls and 123 boys. Greater adiposity was defined according to age 8 Body Mass Index (BMI) to categorize participants as overweight (OW) or healthy-weight (HW). Maturation was defined as the age of peak height velocity (PHV). Bone strength was assessed at the radius and tibia (bone strength index, BSI, and strength-strain index, SSI). Differences in bone strength between OW and HW were evaluated with sex-specific multi-level regression models to account for individual growth and correlation between repeated measurements. Analyses were adjusted for centered age (measurement visit age - grand mean age of cohort), change in fat mass, and limb length in Model 1, with additional adjustment for lean mass in Model 2. Analyses were repeated using biological age (visit age - age PHV). BMI was positively associated with age of maturation in girls and boys (p< 0.05). HW versus OW girls had significantly lower BSI and SSI at the radius and tibia (p< 0.05) in Model 1. Results remained significant except for radial BSI in Model 2. HW versus OW boys had significantly lower BSI and SSI (all p< 0.05) at the tibia, but not radius, in Model 1. Significant differences were sustained in Model 2. Analyses were repeated using biological age, which yielded similar results for boys, but reduced parameter estimates were observed in girls, with only tibial SSI significant in Model 2 (p< 0.05). These findings support a stronger role for greater adiposity in the occurrence of earlier maturation and greater bone strength in girls than boys while greater lean mass appeared to play a greater role in boys. The second research aim evaluated associations between abdominal adiposity and bone in 132 girls and 122 boys. Visceral adipose tissue area (VAT, cm2) and subcutaneous adipose tissue area (SAT, cm2) were estimated from DXA scans. Sex-specific analyses evaluated the fat-bone relationship with growth models using biological age as the time variable adjusted for limb length and lean mass. There were no significant associations between bone parameters and VAT or SAT in girls. In boys, greater VAT was associated with lower trabecular bone density (tBMD) and BSI (all p< 0.05) at the tibia, but not radius. Greater VAT and SAT were associated with smaller cortical bone size and thickness (all p< 0.01) at the radius, but not tibia. Analyses limited to overweight participants showed VAT was negatively associated with periosteal circumference at the radius and tibia, cortical bone thickness at the tibia and SSI (all p< 0.05) at the radius in girls. In boys, the results were relatively unchanged for VAT, while SAT was only significantly associated with lower tBMD (p< 0.05) at the tibia. These results suggest the bone-fat relationship may vary depending on adiposity and bone site. The third research aim evaluated the longitudinal association between intramuscular fat and cortical bone at the tibia from age 11 to 17 years in 153 girls and 143 boys. Muscle density (MD) was used to estimate intramuscular fat (IMF). Lower MD indicates greater IMF. The relationships between muscle density and cortical bone parameters were modeled using multi-level regression models adjusted for biological age, limb length and muscle cross-sectional area measured by pQCT. In the adjusted multi-level regression models, MD was positively associated with cortical bone parameters, but only reached statistical significance for BMD, bone mineral content (BMC), bone cross-sectional area, cortical thickness and SSI in girls, while only SSI was significant in boys (all p< 0.05). These results suggest that greater fat content within muscle may be harmful to weight-bearing cortical bone during adolescence. In conclusion, findings from the first aim suggest there are sex- and site-specific differences in the relationship between adiposity and bone during adolescence. Findings from the second and third aims indicate these differences could be explained, in part, by the existence of specific fat depots (abdominal more so than intramuscular fat) that could be harmful to bone and that may be more apparent in boys due to a sex-specific fat distribution pattern that favors accumulation of abdominal rather than peripheral fat.

Introduction: The prevalence of opioid abuse has increased throughout Northeast Tennessee. Subsequently, more infants are born drug-exposed or with Neonatal Abstinence Syndrome (NAS). According to the Tennessee Department of Health, hospitalizations for deliveries with maternal substance abuse tripled in Tennessee between 1999 and 2011. During this period, the inpatient hospitalization rate for NAS increased 11-fold. In 2017, there were 163 NAS cases reported in Northeast Tennessee. Depending on intrauterine and environmental exposures, there may be differences in health, growth, behavior, and development in infants. Our goal was to assess and explore those differences to help update education and care recommendations for pediatric primary care clinics. Methods: This cross-sectional study was set in a Northeast Tennessee pediatric clinic. 120 patients seen for a newborn visit between June 30, 2013 and July 1, 2014 were randomly selected. An additional sample of all infants with suspected drug exposure was identified for this period based on diagnosis codes. In total, 99 infants had no drug exposure and 62 were drug-exposed. An 83-item chart abstraction template was developed. Data was analyzed by SPSS. The chi-squared test and Mann-Whitney U test were used, with a critical value of p<0.05 to determine significance. The Bonferroni correction was applied to account for multiple comparisons. The research protocol was reviewed and approved by the Institutional Review Board of East Tennessee State University. Results: Of the 120 charts initially selected, 3.33% of infants were exposed to buprenorphine, 1.67% to methadone, 0.83% to marijuana, 0.83% to cocaine, and 1.67% to tobacco, 8.33% to benzodiazepine, and 1.67% to barbiturates during gestation. In total, 18.33% of infants had any drug exposure, 10.00% to any opiate, and 3.33% had a documented diagnosis of NAS in their chart. Prenatal drug exposure was significantly associated with multiple demographic factors as well as pediatric respiratory, behavioral, gastrointestinal, infectious disease, and cardiac conditions. Conclusions: Prenatal drug exposure was significantly associated with multiple pediatric complications. This illustrates the significance of addressing the increased incidence of prenatal drug exposure in Northeast Tennessee. Future multivariate analyses will attempt to control for potential confounders. This analysis is retrospective and exploratory, and any associations should be confirmed with a prospective study. A weakness of this study includes potential under-diagnosis of drug exposure and NAS due to lack of documentation in the EHR. Additional research will include further comparison of maternal and infant complications in drug-exposed and non-exposed infants. This will allow for a better understanding of the risks associated with maternal drug exposure. Findings from these research projects will be used to inform clinical initiatives for NAS in Northeast Tennessee.

This study examines the risk of cancer in children of pesticide applicators from the Agricultural Health Study. The study includes 36,537 children of Iowa participants who were evaluated for cancer incidence during 1975 through 2013 from birth through the age of seventeen. Standard incidence rates for any cancer and specific groups of cancers classified using the International Classification of Childhood Cancer was calculated using rates from the general population of Iowa controlling for year of follow, age, sex, and race. Hazard ratios for Group I-III cancers and paternal exposure to specific pesticides were calculated using exposure information collected on 50 pesticides during phase 1 and 2 of the Agricultural Health Study. The exposure information allowed for calculation of intensity-weighted days of exposure to pesticides using the Agricultural Health Study exposure algorithm. Additionally, maternal ever exposure to specific pesticides was used to evaluate risk of childhood cancer. There were 118 cancers identified in children of Agricultural Health Study participants. The all-cancer standardized incidence ratio was significantly elevated (SIR = 1.27 95% CI: 1.04-1.50). The most common groups of cancers were Group I leukemia, myeloproliferative disease, and myelodysplastic disease (n=34) followed by Group III central nervous system (CNS) and miscellaneous intracranial and intraspinal neoplasms (n=25). For paternal intensity-weighted days of exposure, there were 31 of 50 specific pesticides that had sufficient cases of cancer to investigate using Cox proportional hazard models. The herbicide trifluralin significantly increased the risk for Group I childhood cancers for any parental pesticide exposure 2 years before birth through birth when compared to children with no paternal exposure (HR = 2.72 95% CI: 1.15, 6.44). This was consistent with results found from analyzing exposure split into two quantiles based on median exposure of exposed children with a Group I cancer. Parental use of the herbicide S-Ethyl-dipropylthiocarbamate (EPTC) did not result in a sufficient number of Group III cancer cases to look at levels of exposure to EPTC, but ever exposure showed an increased hazard ratio when compared to children with unexposed fathers (HR = 2.56 95% CI: 1.06, 6.20). Other pesticides (dicamba, cyanazine, and terbufos) showed mixed evidence of an association with specific childhood cancers, but were either under powered to evaluate with sensitivity analysis or showed inconsistent risk across exposure levels. Less extensive exposure information was available for mothers of children of the Agricultural Health Study, so analysis was restricted to ever or never exposure to pesticides during a mother’s lifetime. Additionally, there were a limited number of cases of cancer for which maternal exposure to specific pesticides was reported resulting in only 4 pesticides being evaluated for childhood cancer risk (glyphosate, 2,4-dichlorophenoxyacetic acid (2,4-D), carbaryl, and malathion). For these four pesticides, this study did not detect any increased risk of childhood cancer from maternal exposure. In summary, this study provides the first epidemiological evidence of an increased risk of childhood cancer for trifluralin and EPTC. Since this study provides the first evidence of this increased risk, additional analysis is needed to validate the results. This study demonstrates how pesticide exposure information from participants of the AHS can be used in the evaluation of their children’s cancer risk. Additional follow-up and analysis of this cohort beyond the age of 17 would provide further insight into cancer risk during early adulthood from early life pesticide exposure.

Thesis (M.S.C.E.) / OBJECTIVES: Half of all U. S. pregnancies are unintended, and of these, about 6 out of 10 are carried to term. These are known as “unintended births,” and they are at elevated risk for adverse social, economic and health outcomes. The goal of this study was to examine, using the latest data from the National Survey of Family Growth (2006-2010), whether women with unintended births were more likely than their counterparts to use effective postpartum contraception. METHODS: In this retrospective cohort study of U.S. women ages 15-44, all births within three years of the study interview were classified as intended or unintended, and the method and timing of postpartum contraception were ascertained. Our primary outcome was time to initiation of a highly effective postpartum contraceptive method during the first postpartum year. Highly effective contraception was defined as a method with less than one pregnancy per 100 woman-years of typical use. Secondary outcomes included time to initiation of any contraceptive use over the first postpartum year, and postpartum contraceptive non-use. RESULTS: Complete data was available for 2,691 births. Young, black, undereducated, unmarried, and poor women, as well as those who paid for their delivery with Medicaid were more likely than their counterparts to have unintended births (p<0.0001). After adjustment for confounders, women with unintended births were 1.4 (95% CI: 1.2-1.6) times as likely to use highly effective postpartum contraception compared to women with intended births. During the first month postpartum, women with unintended pregnancies were 0.8 times as likely to use no contraceptive method as women with intended births (95% CI 0.7-0.9, p=0.0046); there were no differences between groups for the rest of the first postpartum year. Married women, white women, women with college degrees, and women with who were giving birth for the first time were somewhat more likely than others to initiate effective contraception after an unwanted pregnancy. CONCLUSION: A relatively small difference in time to initiation of highly effective postpartum contraception, along with an almost 1 in 5 prevalence of postpartum contraceptive nonuse for the entire first postpartum year, indicates that women in this study sample are not well-prepared to prevent recurrence of unintended birth. These results add to the body of data supporting effective postpartum contraception as a public health priority.

In 2015, participating US Emergency Departments (EDs) reported approximately 2.8 million visits related to skin and soft tissue infections (SSTIs). Studies indicate that there may be disparities by patient sex in physician treatment guideline adherence rates as a result of a gender bias during physician-patient interactions; however, only two epidemiologic studies have investigated the role of patient sex in guideline adherence rates for SSTIs. These prior studies were limited in size and covariate assessment. Thus, the magnitude and direction of the effect of patient sex is uncertain, warranting further research. Therefore, we conducted a large prospective study to elucidate the role that patient sex plays in guideline adherence rates among physicians for non-purulent cellulitis at two UMass Memorial Health Care Group EDs in 2017. Data on treatment and sex was abstracted from electronic medical records. Compliance with treatment guidelines was based on 2014 Infectious Disease Society of America (IDSA) guidelines. Adjusted multinomial regressions indicated that female patient sex was associated with lower prevalence of overtreatment (POR=0.72, 95%CI: 0.57-0.92). In contrast, female physician sex was significantly associated with higher prevalence of overtreatment (POR=1.48, 95%CI: 1.16-1.87), but did not affect the relationship of patient sex with overtreatment (P-interaction=0.80). Awareness of differential treatment by patient sex may improve physician adherence to guidelines. This study contributes to a growing body of literature elucidating the role of sex in medical decision making and is the first to account for both patient and physician sex as well as relevant covariates in studies regarding cellulitis treatment.

The following graduate thesis aims to identify important clinical variables, including injury, treatment and healing characteristics, that serve as prognostic indicators for complications in patients with fractures of the tibial shaft. In particular, the complications of focus in this thesis are surgical site infections and nonunion. The three analytical studies comprising this thesis were derived from large data sets arising from two randomized controlled trials and an observational cohort study. The first chapter (Open Tibial Fractures: Updated Guidelines for Management) is a published literature review that provides an overall introduction to the thesis. It highlights the paucity of high-quality evidence currently available to inform many of the treatment strategies for patients with open fractures of the tibial shaft. The second chapter (Timing of Irrigation & Debridement and Infection Risk in Severe Open Fractures) is a sub-study of all open fracture patients recruited in the International Orthopaedic Multicenter Study (INORMUS) in Fracture Care. The findings of this study suggest that timing delays to irrigation and debridement for patients with open fracture injuries is associated with an increased risk of surgical site infection. The third and fourth chapters evaluate the association between early healing measures and nonunion in patients with tibial fractures. Specifically, chapter three (Exploring the Association of 3-Month Radiographic Union Score for Tibia Fractures (RUST) with Nonunion in Tibial Shaft Fracture Patients) demonstrates that radiographic healing at three months post-operatively is strongly associated with nonunion at one year. Similarly, chapter four (Nonunion in Patients with Tibial Shaft Fractures—Can Early Functional Status Predict Healing?) demonstrates that functional status at three months post-operatively is also correlated to eventual healing. Both of these studies include patients from the randomized controlled trials, SPRINT (Study to Prospectively Evaluate Reamed Intramedullary Nails in Patients with Tibial Fractures) and FLOW (Fluid Lavage of Open Wounds). / Thesis / Master of Science (MSc) / The enclosed thesis work evaluates outcomes in patients with fractures primarily of the tibial shaft. In particular, the comprised studies assess whether certain characteristics, such as injury factors, treatment variables and early healing progression, are associated with adverse outcomes in these patients. One study found that delays in timing to appropriate surgical care for patients with open fractures (open wound at the site of the fracture) leads to greater risk of infection. Furthermore, two studies found that both the level of radiographic healing and functional status of patients at three months from surgery can help predict if the patient will ultimately heal radiographically at one year from injury. The findings of this thesis work should help orthopaedic care providers identify patients at high risk for infections and nonunions, such that these patients can be closely monitored to minimize the risk of such complications.

<p>There has been much controversy in the cardiovascular literature regarding sex/gender differences in the presentation of coronary artery disease and downstream implications. The aim of this thesis is not to resolve this controversy, but rather to assess and critique potential sex/gender similarities and differences from a variety of perspectives, explored through various methodologies.</p> <p>This thesis contains four main studies, each employing different quantitative and qualitative methods. An overarching framework was developed to contextualise each study presented in this thesis. The first main study entitled, the “<em>RACE CAR”</em> trial assessed physician opinion <em>prospectively</em> observing that women are perceived to benefit less from cardiac catheterization compared to men, while controlling for age, TIMI risk and preference for cardiac catheterization. The <em>“Identifying women with severe angiographic coronary disease</em>” study observed physician referral patterns <em>retrospectively</em> and determined that although women are less likely to have severe angiographic disease compared to men, the traditional risk factors and CCS Class IV angina are significant predictors of severe angiographic disease. This is an important finding to help physicians better identify women at risk.</p> <p>The findings from these two studies identified the need for the cardiovascular research community to better define angina, particularly among women. Using qualitative methodology, a new theory of angina emerged, illustrating symptoms along a <em>gender continuum</em>. Based on the findings from the qualitative study, the final study of this thesis developed an assessment tool to test the symptom parameters along the <em>gender continuum</em>. The findings confirm that the symptoms of women and men represent more <em>shared experiences</em> rather than differences, particularly among patients with obstructive coronary artery disease.</p> <p>These studies collectively address knowledge gaps and add new information to various stages of patient cardiac care within the sex/gender programme of cardiovascular research.</p> / Doctor of Philosophy (PhD)

<p><strong>BACKGROUND</strong>.<strong> </strong>Classically, radiography (x-ray) has been used to visualize the destructive effects of uncontrolled rheumatoid arthritis (RA) on articular bones. Both x-ray and magnetic resonance imaging (MRI) may be used to visualize erosive disease. The multi-slice and multi-planar characteristics of MRI provide greater detail per anatomy imaged than x-ray.</p> <p><strong>OBJECTIVE</strong>. The objective of this dissertation was to compare the relative merits of x-ray and MRI erosion detection.</p> <p><strong>METHODS</strong>. In Chapters 1 through 3, RA, its clinical management, and the role of diagnostic imaging were introduced. In Chapter 4, the overall objective was first investigated by evaluating the current state of knowledge using a systematic review. In Chapter 5, inter-rater reliability across four participating radiologists was investigated. In Chapter 6, reliability-adjusted evaluations were used to directly compare paired x-ray and MR images.</p> <p><strong>RESULTS</strong>. The systematic review indicated that x-ray has low sensitivity and high specificity for MRI erosions. The associations were dependent on RA symptom duration. The findings from the prospective studies conducted were consistent with the literature. In a patient-centred analysis, the proportion of patients with erosive disease detected on either modality was dependent on the anatomy compared. Despite similar proportions of patients with erosive disease detected in comparable diagnostic imaging sittings, the proportions were comprised of markedly different patients.</p> <p><strong>CONCLUSIONS</strong>. The literature comparing x-ray and MRI erosion detection was systematically reviewed. The dependence of the relative diagnostic test accuracy on symptom duration was highlighted. At the patient level of analysis, the greater number of anatomical sites examined on x-ray overcomes the superiority of MRI to detect erosions at the unit of measurement. The two modalities are complementary insofar that they detect erosive disease in distinct patients. Further investigation into the merit of imaging varied anatomy on MR and optimally accounting for inter-rater reliability in such comparisons is warranted.</p> / Doctor of Philosophy (PhD)

<p>We defined hypertension in pregnancy as a composite of gestational hypertension, preeclampsia and eclampsia. The etiology of hypertension in pregnancy remains controversial. The three chapters of this thesis explore the risk of hypertension in pregnancy from various kidney conditions. Chapter 1 introduces the reader to the thesis. Chapter 2 is a systematic review that studied the risk of developing hypertension in pregnant women with chronic kidney disease but not on dialysis. We found that women with chronic kidney disease had at least a twofold higher relative risk of developing hypertension during pregnancy compared with women having no chronic kidney disease. Chapter 3 is a retrospective study looking at the risk of developing gestational hypertension and preeclampsia in women who had symptomatic gastroenteritis after drinking water infected with <em>E. coli</em> O157:H7 during the Walkerton outbreak in May 2000. We conducted this study using linked datasets at the Institute of Evaluative Sciences (ICES) Toronto, Ontario. We observed that there was no increased risk of developing gestational hypertension or preeclampsia among the symptomatic women compared with women from the neighbouring towns who were asymptomatic or did not drink the water. Chapter 4 is a protocol of a prospective cohort study recruiting female kidney donors and healthy non-donors as the comparative group to study pregnancy outcomes in these individuals. This is a multicentre study involving 12 transplant centres throughout Canada. There are 59 participants in this study to date (Feb 28, 2013) of which seven have been pregnant so far. Data collection for this study is ongoing.</p> / Doctor of Philosophy (PhD)