Preterm delivery (PTD) is a leading cause of infant death, and surviving infants are at risk for poor health. Data from the Iowa Health in Pregnancy Study, a case-control study of maternal stress on risk of PTD and small for gestational age (SGA) deliveries were used to address three aims: 1) develop a method to correct for error in ultrasound measurement among suspected SGA infants, 2) estimate the association of occupational stress on risk of PTD, and 3) examine injury-related risk factors for PTD. Estimates of gestational age using ultrasound can be biased if the fetus is growth-restricted, yielding underestimates due to the small stature of the fetus. Multivariate linear regression modeling was used to estimate and correct for this bias among subjects with a suspected SGA infant who 1) began prenatal care in the first trimester, 2) reported a last menstrual period and 3) had an ultrasound examination between 7-21 weeks. To correct for this bias, an average of 1.5 weeks was added to the ultrasound gestational age. Following the correction, the proportion of PTD cases decreased from 29.1% to 26.5% while SGA cases increased from 23.7% to 31.3%. Using this PTD classification, occupational physical and psychosocial stressors were studied. Continuous employment over the first 20 weeks of pregnancy was associated with a 30% increased risk of PTD versus not working. Working women reporting highly repetitive tasks (aOR=1.47(1.10-1.98)) or inadequate breaks (aOR=1.67(1.03-2.73)) were at increased risk of PTD. Working women who reported high lifting in the home had double the risk of PTD. Over 5% of control subjects reported an injury during pregnancy, and injured women tended to be younger, unmarried, less educated, and have lower incomes. Women with injuries involving >1 body part (aOR=2.50(1.14-5.49)), or injuries to the abdomen and other regions of the body (OR=1.75(0.59-5.23)) were at increased risk of PTD. Our findings provide a statistical approach to assess and correct for underestimates of ultrasound gestational age in case-control studies of PTD and SGA. The analyses of occupational exposures and injury during pregnancy indicate the need for studies that incorporate specific and standardized assessments of these exposures.

Dextromethorphan and guaifenesin are the main active components in over-the-counter cough medications. Prenatal exposure to dextromethorphan has been shown to be teratogenic in animal models. Data from human studies for either dextromethorphan or guaifenesin are limited and inconclusive. We used data from the population-based National Birth Defects Prevention Study (NBDPS) to examine associations between maternal periconceptional (one month before through three months after conception) use of cough medications containing dextromethorphan, with or without guaifenesin, and isolated neural tube defects (NTDs). We also used NBDPS data to explore associations between such exposures and other isolated major birth defects, as well as associations between maternal periconceptional use of cough medications containing guaifenesin alone and isolated major birth defects. Enrolled cases comprised 19,538 live births, still births, and elective terminations with isolated major birth defects, and enrolled controls comprised 10,200 live births without defects delivered from October 1997 through December 2009. Telephone interview reports of pregnancy exposures, including periconceptional use of cough medications, were obtained from mothers of case and control infants. Two approaches were used to build multivariable models: backward model selection and comparing the change in odds ratios (ORs) by adding each covariable individually into the main effect models. Adjusted ORs (aORs) and 95% confidence intervals (CIs) for maternal periconceptional exposure of cough medications containing dextromethorphan, with or without guaifenesin, or guaifenesin alone, respectively, and 22 types of birth defects were estimated using multivariable logistic regression analysis. Applying our first multivariable model building approach, we observed that maternal periconceptional use of dextromethorphan, with or without guaifenesin, was marginally significantly associated with an increased risk of all NTDs combined (aOR=1.7, 95%CI=1.0-2.9), or spina bifida alone (aOR=1.9, 95%CI=1.0-3.5). Applying our second model-building approach confirmed the associations with all NTDs combined and spina bifida alone (aOR==1.9, 95% CI=1.2-3.0 and aOR=2.1, 95% CI=1.2-3.8; respectively). For other isolated birth defects, a positive, marginally significant association was observed for maternal periconceptional use of dextromethorphan, with or without guaifenesin, and cleft lip with or without cleft palate (aOR=1.3, 95%CI=1.0-1.7) applying our first model building approach. Our second model-building approach produced significant associations between such exposure and gastroschisis (aOR=1.8, 95%CI=1.2-2.8). With regard to maternal periconceptional use of cough medications containing guaifenesin alone, we observed marginally significant associations with all NTDs combined (aOR=1.9, 95%CI=1.1-3.5), spina bifida alone (aOR=2.3, 95%CI=1.1-4.4), or anorectal atresia (aOR=1.9, 95%CI=1.0-3.6) applying our first model-building approach. Applying our second model-building approach did not suggest significant associations between such exposure and birth defects. Our findings suggest that maternal periconceptional use of cough medications containing dextromethorphan or guaifenesin may produce selected major structural birth defects in offspring. These findings provide important evidence to better understand the safety of these cough medications and informs about the potential of this modifiable exposure to cause isolated birth defects. Additional population-based research is necessary to validate these positive associations between cough medications and selected isolated birth defects identified in our study.

Gestational diabetes mellitus (GDM) is the most common metabolic complication in pregnancy and is associated with substantial maternal and neonatal morbidity. The standard of care for GDM in most developed countries is universal mid- to late- pregnancy (24-28 weeks gestation) glucose testing. While earlier diagnosis and treatment could improve pregnancy outcomes, tools for early identification of risk for GDM are not commonly used in practice. Existing models for predicting GDM risk within the first trimester of pregnancy based on maternal risk factors perform only modestly in the clinical setting. Heavy reliance on history of GDM to predict GDM development in the current pregnancy prevents these tools from being applicable to nulliparous women (i.e., women who have never given birth). In order to offer timely preventive intervention and enhanced antenatal care to nulliparous women, we need to be able to accurately identify those at high risk for GDM early in pregnancy. Data from the California Office of Statewide Health Planning and Development Linked Birth File was used to address three aims: 1) improve early pregnancy prediction of GDM risk in nulliparous women through development of a risk factor-based model, 2) conduct a systematic review and meta-analysis assessing the relationship between first trimester prenatal screening biomarker levels and development of GDM, and 3) determine if the addition of first and second trimester prenatal screening biomarkers to risk factor-based models will improve early prediction of GDM in nulliparous women. We developed a clinical prediction model including five well-established risk factors for GDM (race/ethnicity, age at delivery, pre-pregnancy body mass index, family history of diabetes, and pre-existing hypertension). Our model had moderate predictive performance among all nulliparous women, and performed particularly well among Hispanic and Black women when assessed within specific racial/ethnic groups. Our risk prediction model also showed superior performance over the commonly used American College of Obstetricians and Gynecologists (ACOG) screening guidelines, encouraging the prompt incorporation of this tool into preconception and prenatal care. Biomarkers commonly assessed in prenatal screening have been associated with a number of adverse perinatal and birth outcomes. However, reports on the relationship between first trimester measurements of prenatal screening biomarkers and GDM development are inconsistent. Our meta-analysis demonstrated that women who are diagnosed with GDM have lower first trimester multiple of the median (MoM) levels of both pregnancy associated plasma protein-A (PAPP-A) and free β-human chorionic gonadotropin (free β-hCG) than women who remain normoglycemic throughout pregnancy. Findings from our meta-analysis suggested that incorporation of prenatal screening biomarkers in clinical risk prediction models could aid in earlier identification of women at risk of developing GDM. Upon linkage of California Office of Statewide Health Planning and Development Linked Birth File and California Prenatal Screening Program records, we found that decreased levels of first trimester PAPP-A, increased second trimester unconjugated estriol, and increased second trimester dimeric inhibin A were associated with GDM development in nulliparous women. However, the addition of these biomarkers in clinical models did not offer improvements to the clinical utility (i.e., risk stratification) of models including maternal risk factors alone. Our findings demonstrate that incorporation of maternal risk factors in a clinical risk prediction model can more accurately identify nulliparous women at high risk for GDM early in pregnancy compared to current standard practice. The maternal characteristics model we developed is based on clinical history and demographic variables that are already routinely collected by clinicians in the United States so that it may be easily adapted into existing prenatal care practice and screening programs. Future work should focus on evaluating the clinical impact of model implementation on maternal and infant outcomes as well as financial costs to the health care system.

Few investigators have used robust analytic methods to assess risk factors and outcomes for surgical site infections (SSIs) after craniotomy and craniectomy (CRANI) procedures. We performed a retrospective study among patients undergoing CRANI procedures between 2006 and 2010 at the University of Iowa Hospitals and Clinics (UIHC) to assess the effect of an intervention (e.g., limiting Gliadel wafer use among patients with malignant brain tumors) on the trend of SSI rates, to identify independent risk factors for SSIs, and to evaluate one-year postoperative patient outcomes associated with these SSIs. We abstracted demographic data and clinical data from medical records or from the UIHC's Health information Management System. We identified 104 patients with SSIs and selected 312 controls. Of SSIs, 88% were deep incisional or organ space infections, 70% were identified after patients were discharged from their initial hospitalizations, 32% were caused by Staphylococcus aureus alone or in combination with other organisms, and 27% were caused by Gram-negative organisms alone or in combination with other organisms. Significant independent risk factors for SSIs were: previous chemotherapy (odds ratio [OR], 10.0; 95% confidence interval [CI] 1.1, 92.1), preoperative length of stay ≥ 1 day (OR, 2.1; 95% CI 1.3, 3.5), preoperative serum glucose ≥ 100 mg/dL (OR, 1.7; 95% CI, 1.0, 3.0), Gliadel® wafer use (OR, 8.6; 95% CI 3.2, 23.1), and postoperative cerebrospinal fluid leak (OR, 4.0; 95% CI, 1.6, 10.3). Gliadel® wafer use was the strongest risk factor; however, limiting Gliadel® wafer use did not decrease SSI rate significantly among patients with brain tumors. Perioperative ventricular drains or lumbar drains were not independently associated with an increased risk of SSIs, but drains may have clinical significance. An SSI risk index that included the significant preoperative patient-related risk factors had a better predictive power than the National Healthcare Safety Network (NHSN) risk index. After adjusting for preoperative length of stay, age, comorbidity score, severity of illness score, the reason for the procedure, and procedure month, patients with SSIs were hospitalized longer postoperatively than were controls during their readmissions (2.3 days; P < 0.0001). After controlling for the same covariates and treating SSI as a time-varying factor, patients with SSIs were more likely than controls to: die (hazard ratio [HR], 3.3; 95% CI, 1.8, 5.8), be readmitted (HR, 4.1; 95% CI, 2.9, 5.8), and have reoperations (HR, 56.6; 95% CI, 38.1, 84.0). In conclusion, surgeons could predict patients' risk of SSIs based on their preoperative risk factors and surgeons could modify some processes of care to lower the SSI risk. Preventing SSIs after CRANI procedures could improve patient outcomes and decrease healthcare utilization.

Preeclampsia is a leading cause of maternal and infant morbidity and mortality worldwide. For the first aim, a systematic literature review and meta-analysis was performed to examine the associations between maternal lipid concentrations during pregnancy (HDL-C, LDL-C, non-HDL-C, total cholesterol, and triglycerides) with subsequent risk of preeclampsia. Data from the Study of Pregnancy-induced Hypertension in Iowa, a population-based, case control study of preeclampsia and gestational hypertension, was used to address the other two aims: 1) elucidate the independent contribution of physical activity, an important modifier of dyslipidemia, on the risk of preeclampsia, and 2) to evaluate the association between the genetic susceptibility for dyslipidemia and the risk of preeclampsia. Published reports examining lipid levels during pregnancy and preeclampsia have been inconsistent. Meta-analyses demonstrated that preeclampsia was associated with elevated total cholesterol, non-HDL-C, and triglyceride levels, regardless of gestational age at time of blood draw, and lower levels of HDL-C in the third trimester. A marginal association was also found with LDL-C levels. Statistical heterogeneity was detected in all analyses. Physical activity has been hypothesized to reduce the risk of preeclampsia, but previous studies have had a range of limitations resulting in mostly suggestive, though nonsignificant findings. After adjustments, increasing levels of LTPA (trend, p=0.02) and increasing amounts of time spent active each day (trend, p=0.03) were significantly associated with a reduced risks of preeclampsia (trend, p=0.02). Increasing amounts of time spent sitting per day was marginally associated with the risk of preeclampsia (trend; p=0.10), and those women who were active an average of more than 8.25 hours per day had the most significantly reduced risk of preeclampsia (adjusted OR 0.58, 95% CI 0.36, 0.95). In examining the effects of a woman's genetic susceptibility to dyslipidemia as a risk factor for preeclampsia, we found that the more risk alleles a woman has for dyslipidemic levels of HDL-C, the greater her risk for developing preeclampsia. While this relationship was of marginal statistical significance, these results are suggestive of a relationship that may help elucidate the pathogenesis of preeclampsia. Our findings demonstrate the contribution of maternal lipid concentrations during pregnancy on preeclampsia risk, and suggest a behavioral modification that could help counter that increased risk. Additionally, implications of HDL-C genetics in preeclampsia risk identify potential genetic and biologic pathways to be explored in the pathogenesis of preeclampsia.

Urinary incontinence, one of the most prevalent conditions in elderly living in nursing homes (NH) was shown to significantly impact patient's quality of life (QOL) and health outcomes. Bladder antimuscarinics (BAM), the main drug class to treat urinary incontinence, have limited effects in managing the condition; however, given their anticholinergic properties and the characteristics of those living in NH, BAM could potentially lead to serious health consequences in this population. We conducted a retrospective cohort study with a new-users design by linking existing Veterans Affairs (VA) data (inpatient, outpatient, pharmacy administrative files, and Minimum Data Set- MDS) between fiscal years 2003 and 2009. Potential risks (i.e. fractures and negative impact on cognitive performance) and benefits (i.e. improvement in urinary incontinence, social engagement and overall QOL) associated with initiation of a BAM were assessed in elderly (65+) admitted for long-term care in the VA Community Living Centers. Descriptive statistics were used to compare BAM new-users and non-users at baseline; in addition, logistic regression was used to identify important predictors of BAM initiation. Treatment selection bias was addressed by using the propensity score matching method. After balancing the groups on baseline characteristics, the risk of fractures (hip fracture, any fracture) in relationship with BAM initiation was evaluated using Cox proportional hazard analysis. BAM impact on the cognitive status measured by the MDS-Cognitive Performance Scale (CPS) was evaluated through generalized estimated equations (GEE) method. Similarly, possible benefits measured through MDS were assessed via GEE. The final cohort included 1195 BAM new-users (with the majority being prescribed Oxybutynin immediate-release) and 22,987 non-users. Predictors of BAM initiation included demographic characteristics, bladder and bowel continence status, comorbidities, medication use, cognitive performance and functional status. Our study showed that BAM improved urinary continence (OR=1.27, 95%CI: 1.07-1.50) in those treated; social engagement as measured by MDS-Index of Social Engagement also improved in users, although at a level that is not clinically significant (difference in mean MDS-ISE=0.2074, 95%CI: 0.0550-0.3598). However, BAM initiation increased the risk of fractures (hip: HR=3.69, 95% CI: 1.46 - 9.34, p=0.0059; any fracture: HR=2.64. 95% CI: 1.37 - 5.10, p=0.0039). Our results showed no difference between new-users and non-users with regard to mean CPS and overall QOL. The purpose of the study was to clarify the proper role of medication use in the management of urinary incontinence in elderly in the VA CLC. The results raise questions about the continued use of Oxybutynin IR, the main BAM prescribed in this population. Given the increased risk for fractures in the context of potential improvement in urinary continence with no clinically significant improvement in social engagement, a wiser step might be to investigate the safety profile for newer BAM for situations when an addition to non-pharmacologic management for urinary incontinence is desired for elderly in long-term care.

Background. Surgical site infections (SSIs) cause many adverse outcomes for patients including increased length of hospital stay, hospital costs, morbidity, and psychological distresses. Staphylococcus aureus is one of the most common causes of SSIs in the United States. Objective. Identify risk factors for Staphylococcus aureus SSIs following breast operations. Design. Retrospective nested case-control study of SSIs among women undergoing breast operations. Setting. An academic health center. Patients. We studied patients undergoing breast operations at the University of Iowa Hospitals and Clinics from 7/1/2004 through 9/30/2015. Cases were patients who acquired SSIs meeting the National Healthcare Safety Network definition and whose SSIs were caused by S. aureus. We randomly selected two controls for each case from patients who had breast operation during the study period and did not meet the SSI definition. Controls were selected randomly from uninfected patients whose operations occurred during the same month and year as a case. Results. Forty two (1.2%) patients acquired S. aureus SSIs after 3494 breast operations. SSIs were identified a mean of 27.8 days after the breast operations; 54.76% were deep incisional infections. Poisson regression analysis revealed that S. aureus SSIs following breast operations at UIHC have been increasing at a statistically significant rate. Bivariable analysis identified several patient and procedure related risk factors that increased the risk for S. aureus SSIs. Patient-related factors included a diabetes mellitus, active skin disease, prior chemotherapy, breast cancer, hypertension, and preoperative hemoglobin. Procedure-related factors included ASA score > 2, a mastectomy followed by immediate reconstruction, sentinel lymph node biopsy (SLN), drain placement, procedure time, and estimated blood loss. A multivariable analysis of patient factors found only breast cancer maintained significance. A similar analysis of procedure factors found that drain placement remained significant. The combined model contained breast cancer, drain placement, and mastectomy followed by immediate reconstruction as significant variables. Conclusions. S. aureus SSIs following breast operations have been increasing at UIHC. Possible remediable risk factors include blood glucose levels, blood pressure, timing of chemotherapy, and drain placement and care. These results will help doctors at UIHC design interventions to prevent S. aureus SSIs following these procedures.

Objective: To assess the ability of propensity score methods to maintain covariate balance and minimize bias in the estimation of treatment effect in a time-to-event setting. Data Sources: Generated simulation model Study Design: Simulation study Data Collection: 6 scenarios with varying covariate relationships to treatment and outcome with 2 different censoring prevalences Principal Findings: As time lapses, balance achieved at baseline through propensity score methods between treated and untreated groups trends toward imbalance, particularly in settings with high rates of censoring. Furthermore, there is a high degree of variability in the performance of different propensity score models with respect to effect estimation. Conclusions: Caution should be used when incorporating propensity score analysis methods in survival analyses. In these settings, if model over-parameterization is a concern, Cox regression stratified on propensity score matched pairs often provides more accurate conditional treatment effect estimates than those of unstratified matched or IPT weighted Cox regression models.

A better understanding of the association of the environment to young children's physical activity is needed to design effective behavioral interventions and to improve communities' infrastructures in ways that support the development of healthy behaviors. This thesis used data from the Iowa Bone Development Study (IBDS) to examine activity behaviors of young children in urban and rural Iowa. Mean daily minutes of moderate through vigorous physical activity (MVPA) and vigorous physical activity (VPA), as recorded by accelerometry-based physical activity monitoring in the IBDS, were compared by level of socioeconomic status (SES) and census block type (urban or rural). Media use (television and videogames), in hours per day by parental report, was similarly compared. Logistic regression was used to assess the association of census block type (urban or rural) with lower categories of MVPA and VPA, and with a high category of media use (exceeding the American Academy of Pediatrics (AAP) recommendation for ≤ 2 hours/day of media). Mean age of the 400 participants included in this thesis was 5.65 years (SD = 0.53), females comprised 53.5%. Children's mean daily minutes of MVPA and VPA were not significantly different among SES levels nor between urban and rural census blocks. Rural children had decreased odds for lower categories of MVPA (Boys' OR 0.91; 95% CI: 0.51, 1.60) (Girls' OR 0.89; 95% CI: 0.53, 1.50) and VPA (Boys' OR 0.84; 95% CI: 0.48, 1.48) (Girls' OR 0.90; 95% CI: 0.54, 1.52), but not significantly so. Lower SES boys and girls engaged in more daily media use (hours/day) than higher SES children, and a higher proportion of low SES children exceeded the AAP recommendation than did middle-level or high SES children. Rural boys' media use (2.7 hours/day) was higher than that of urban boys (2.3 hours/day) (95% CI: 0.06, 0.80 hours/day), whereas urban and rural girls' media hours/day were not significantly different. Odds for excess media use were higher for rural boys (OR 2.11; 95% CI: 1.14, 3.11) and for low SES boys (OR 2.99; 95% CI: 1.34 6.68), as well as for low SES girls (OR 4.12; 95% CI: 1.95, 8.71). Although rural and urban children's MVPA and VPA did not differ in this thesis, rural and low SES boys had both higher daily media hours and increased odds for exceeding AAP media recommendations. Odds for excess media also were increased for low SES girls. Interventions to improve healthy behaviors of young children by limiting screen-based recreation and/or supplying means for alternate activity and play opportunities may be especially beneficial when focused on lower SES and rural areas.

This dissertation focused on examining murder-suicide in the United States through descriptive, time-series and spatiotemporal analyses using a self-created and herein verified national database that spanned the years 1999 through 2009. Chapter 2, "Establishment and validation of a national database for murder-suicide in the United States: 1999-2009," describes the methods and sources used in the creation of a national database of murder-suicide. The database was validated using less geographically and/or temporally expansive databases through the use of capture-recapture methods in two ways: the number of events identified in specified space and time was compared and cases were matched using the perpetrator's name. Victim and perpetrator characteristics were then described as compared to previous studies. Chapter 3, "A time-series analysis of murder, suicide and murder-suicide in the United States, 1999-2007" utilized time-series analysis techniques to investigate the impact of time varying covariates on murder, suicide and murder-suicide. Analyses were conducted in the United States at the national level from January 1999 to December 2007. Johansen's multivariable cointegration analysis showed that two-month time lagged murder was positively associated with murder, suicide and murder-suicide. Two-month time lagged suicide was negatively associated with murder, suicide and murder-suicide. Two-month time lagged murder-suicide was not related to any of the three events. Chapter 4, "Spatiotemporal relationships among murder, suicide and murder-suicide in the United States: 1999-2008" examined space, time, and spatiotemporal relationships among murder, suicide and murder-suicide using a spatiotemporal scan statistic from SaTScanTM. Thirty-five temporal and spatiotemporal clusters of murder, suicide, murder/murder-suicide, suicide/murder-suicide and murder/suicide/murder-suicide were identified. No purely spatial clusters, clusters of murder/suicide without murder-suicide, or purely murder-suicide were identified. The murder-suicide database, that will be made public in 2012, will be a novel source of information for investigators interested in studying murder-suicides with the inclusion of date, place, perpetrator and victim characteristics. Its validation along with the time-series and spatiotemporal analyses provides greater understanding of murder-suicide by itself and compared to murder and suicide.