Skip to main content

Health Sciences

Fellow: Dr. Steve Aragon, Associate Professor, Department of Healthcare Management at WSSU.
Research Theme: The Latent Structure of Provider Patient-Centeredness and its Effects on Patients’ Experience-of-care

A valid and reliable model that reveals the effects of provider’s patient-centeredness on patients’ experience-of-care has the potential to advance knowledge, improve quality, reduce disparities, and increase patient-centered care. Always a latent ability of the best health providers, patient centeredness first gained national prominence when the Institute of Medicine (IOM) and Agency for Healthcare Research and Quality (AHRQ) recommended it as a new model for the delivery of medical care, one that was more focused on the preferences, needs, and values of patients versus those of the health care system [39,40]. Not seeing expected improvement in the quality of care, the Centers for Medicare and Medicaid Services (CMS) issued the most influential national healthcare policy in years, by financially incentivizing hospitals to improve patients’ experience-of-care [41, 42]. Consistent with CADS goals, and grounded in the Primary Provider Theory [43, 44, 45], the proposed large data empirical research will test the omnibus structural hypothesis that Patients’ Experience-of-care = f (Provider Patient-Centeredness). Methodologically, the project will employ large national healthcare datasets, third-generation multigroup structural equation modeling (MSEM), and confirmatory factor analysis (CFA) to determine the reliability and validity of the hypothesized model of patient-centeredness and its effects on patients’ experience-of-care. The model’s empirical fit, stability or invariance across samples (replicability), and its robustness against a competing model will be tested. Potential estimation bias of effects will also be assessed. The investigation will have sufficient power to reject an incorrect model. Model fit will be based on a convergence of evidence, including model covariance matrix fit tests and indices, the size of the model’s standardized residuals, the magnitude and significance of the model’s effects, and it’s coherence with the Primary Provider Theory and the real world. The project’s goals include: 1) empirically determining the underlying structure of patient-centeredness and its effects on patients’ experience-of-care, and 2) evaluating the usefulness of MSEM as a large data analytics methodology in healthcare.

Computer Science

Fellow: Dr. Muztaba Fuad, Professor of Department of Computer Science at WSSU
Research Theme: Reducing Redundancy and Improving Privacy of Mobile Crowdsensing Data

Mobile crowdsensing is an evolving research area with a big potential to improve user experience, facility management, and resource usage. However, the biggest challenge for such crowdsensing is the size of this massive datasets, which poses a challenge to provide real-time analysis of that data to produce a timely response. Traditionally in a university like WSSU, resources are allocated for student success depending on usage data and surveys. Most times, these data collections happen at the end of the academic year and any changes appear in the following year, limiting the effectiveness of those changes in the current student population. However, it can be automated, and the university administration can get a real-time view of resource usage and utilization if mobile crowdsensing techniques are developed and utilized. This proposed research project will investigate the different implementation challenges that recent mobile crowdsensing techniques face. One such challenge is the redundancy of similar data collected over a period of time. We envision a new Cost-aware and Context-sensitive Sensing algorithm to be developed as part of this research which will provide high quality of sensing by having dynamic cost (bandwidth, location, power, etc.) functions and sampling of context (geo-location, signal strength-based location triangulation for indoor location, hardware properties, human behaviors, etc.) aware data to minimize redundancy. To facilitate the later, incorporation of different IoT devices in the environmental context should provide a new dimension with data collection and should minimize the effect of similar data. This should also lessen the challenge of having limited sensing capabilities in mobile devices to collect a wide range of good quality data. 

Biological Sciences

Fellow: Dr. Jill Keith, Professor of Biochemistry in Biological Sciences and Chemistry at WSSU
Research Theme: Using Data Science to Make Therapeutic Predictions

It usually takes an average of fifteen years and nearly one billion dollars to bring a drug to market. Researchers have recently used a computational approach to find therapeutics in less time and in a cost-effective manner. The proposed research aims to use a similar approach and leverage the vast amounts of molecular data generated and perform data-informed research to help cure drug addiction. Specifically, we will mine publicly available gene expression data to conduct research that leads to the repurposing of current Food and Drug Administration (FDA)-approved drugs. It is hypothesized that we can predict relationships between drugs and disease by comparing gene expression data of healthy and addicted individuals and correlate the differences with gene expression data from FDA-approved drugs. Our predictions should give rise to potential therapeutics based on this correlation. In this study, publicly available data sets will be used to extract drug gene expression signature when comparing treated to untreated samples and generate a disease-drug score. Computational tools will be used to compare the aforementioned signatures and generate disease-drug scores. If the disease-gene and drug-gene signatures are opposite (a positive score), the drug has the potential of being a pharmacotherapy for an individual addicted to cocaine. We will then perform in vitro and in vivo studies established in our lab and with our collaborators to uncover if our data-informed predictions are correct.

Physical Sciences

Fellow: Dr. Tennille D. Presley, Assistant Professor of Physics in the Department of Chemistry at WSSU
Research Theme: Using Data Science to Elucidate Music’s Biophysical Influence on the body.

Music is a cross-cultural form of communication that exists throughout the world. Because it has grown to be an integral part of daily life, there has been interest in the biophysical mechanisms of how music affects an individual. Listening to music has proven to be beneficial in various ways such as affecting skin conductivity, lowering blood pressure and stress, improving muscle function, and even altering gene function. Furthermore, music affects cell proliferation and viability; however, the complete mechanisms are not well-understood, and many studies are limited to utilizing jazz or classical music. In this proposed research, we aim to employ DS to develop a predictive model for identifying the body’s response to various genres of music such as hip-hop, gospel, rhythm & blues (r&b), country, reggae, and even “noise”. To achieve our goal, the GTZAN Genre Collection and the NCD-RisC datasets will be used. Distinctions of instrumental versus lyrical music across genres will be addressed, and differentiated based on age, race, gender, etc. This data will be compared to biophysical parameters (i.e. heart rate, blood pressure, skin conductivity, workload, and power output). It is our expectation that this data will provide insight into whether two individuals exhibiting the same blood pressure will elicit the same skin conductivity to a certain genre, or even if the same genre will promote an analogous power output for individuals of a similar weight. This prediction model will be beneficial in aiding a person’s decision to listen to a particular genre (whether lyrical or instrumental) to obtain a desired outcome. Moreover, the next step could be to advance it towards understanding mechanisms of various diseased states such as diabetes, Alzheimer’s and Parkinson’s disease.

Social Sciences

Fellow: Dr. Russell M. Smith, Professor of Geography in the Department of History, Politics & Social Justice and the Lead for the Spatial Justice Studio at the Center for Design Innovation.
Research Theme: Developing Spatial Justice Index for North Carolina

According to Rocco, “Spatial Justice refers to general access to public goods, basic services, cultural goods, economic opportunity and healthy environments”. This research proposal seeks to explore the issue of spatial justice through the development of a Spatial Justice Index (SJI). To date, most work associated with spatial justice has been qualitative and case study based. In this research, the SJI will be created by quantitatively exploring geographic (i.e. land uses, distance to parks, schools, vices, etc.), demographic (i.e. race, ethnicity, age, etc.) and socio-economic variables (i.e. income, education, poverty, etc.) of census tracts in an effort to apply the concept of spatial (in)justice to NC communities. We envision the development of a SJI that can be applied across the entire State to help communities comprehend, accept and potentially combat spatially injustices within their communities. The SJI will identify places in which spatial justice is a reality and places where spatial injustices occur. In order to realize this vision, the proposed research plans to utilize a combination of geographic and quantitative variables to develop an index for cross community comparison. Activities will include exploring features/attributes that are most correlated with spatial justice, identifying/collecting datasets related to the features, building a machine learning model (more specifically clustering model) that can meaningfully provide the insights about the SJI. In a word, the proposed research aims to leverage a combination of geographic information systems technologies and US Census Bureau data to create a model SJI for future application across the country. The research findings will be applied in the context of North Carolina datasets to evaluate its effectiveness.