Evaluating the Effect of Cochlear Implants and Real Life Listening Environments|
Ecological Momentary Assessment (EMA) methodology involves collecting self-reported data through repeated surveys to describe respondents’ current or very recent (i.e., momentary) experiences and related contexts in their natural (i.e., ecological) environments. In this study we collect EMA data using smart phones where the participant can respond to multiple surveys throughout the day. EMA can provide a wealth of information on moments in a respondent’s life without the distortions caused by recalling memories and delayed evaluation of experiences associated with retrospective self-reports. EMA can allow researchers to examine the interaction between context and experience by collecting detailed contextual information (e.g., characteristics of the listening situation) with self-reported experience. Although EMA is a useful tool in audiology research, analyzing EMA data is often challenging due to the complex structures it exhibits. Our EMA survey asked a series of questions regarding participants’ listening environments, listening activities, and feelings/experiences. We wish to compare listening environment, activities, and feelings of our participants before they receive a cochlear implant to after they receive their cochlear implant.
Understanding Paths to Graduation: Student Migration between Majors at a Big 10 University Navigating a bachelor’s degree is challenging, and many students face difficult decisions in choosing which major(s) to pursue. Correspondingly, student success professionals want to better understand their student populations to identify ways to help students navigate the myriad options available to them. At the University of Iowa, a team of Biostatisticians works with a group of admissions, student advising, and information technology professionals to produce forecasts for a variety of outcomes directly focused on recruiting, retaining, and better serving the undergraduate student body. This project will build on that collaboration to try and understand the different paths students take through the University of Iowa system, and how that relates to successful graduation outcomes. This project will focus on computational problems in large datasets, data visualization, and will use the cutting-edge Julia scientific computing environment to do so.
Effects of Hospitals as Reservoirs for Disease in the Community Clostridiodes difficile is a significant cause of morbidity and mortality, and in 2017 has led to an estimated 12,800 deaths and 223,900 hospitalizations in the U.S. alone. While C. difficile infections (CDIs) are highly common to occur in healthcare settings, community onset CDIs are also on the rise. There is thus a growing interest in understanding the relationship between healthcare associated and community onset CDIs. Our research group has previously shown that the more CDI cases in the hospital, the more likely there is to be a higher number of healthcare associated CDIs (HA-CDIs). We have also determined that prior hospital exposure to family members increases one’s risk for a CDI. However, it remains an open question as to how HA-CDIs, even when treated fully in the hospital, may yet lead to further community onset CDIs (CO-CDIs). This project will look at over 21 million inpatient visits across 7 U.S. states to estimate the attributable number of CO-CDIs due to hospitals serving as a C. difficile reservoir. Specifically, we will use supervised machine learning techniques and a simulation framework to estimate the proportion of CO-CDIs due to individuals recovering from a HA-CDI being discharged into the community.
Predicting Renal Failure in Patients with Kidney Disease
C3 glomerulopathy (C3G) is a group of related conditions that cause renal disease. The kidney problems associated with C3 glomerulopathy tend to worsen over time. About half of affected individuals develop end-stage renal disease (ESRD) within 10 years after their diagnosis. Kidney disease researchers at the University of Iowa have been following a cohort of C3G patients in the hopes of identifying biomarkers capable of providing advance warning that an individual may be at high risk of ESRD in the near future. This project involves analyzing those biomarkers and developing predictive models for which patients are most likely to progress to renal failure in the next year.
Sensitivity Assessment of a Two-Step Method in Skin Image Identification Image processing is a field of an increased interest in the scientific/medical community. The use of Red Green Blue (RGB) decomposition to extract image information has proven to be valuable, although other decompositions and additional input variables should not be ruled out. It has been shown that skin identification can be achieved by a two-step process: (1) partitioning the input data into clusters, and (2) estimating a binary predictor for each cluster. This project will focus on the problem of skin identification through cluster analysis, predictive modeling, and classification. A variety of skin images with contrasts will be used and noise will be superposed on the images for sensitivity analysis. The ultimate goal is to be able to contour images for skin lesions in cancer imaging analysis.