May 19, 2021 | Erin Bluvas, firstname.lastname@example.org
Four researchers from the Department of Health Services Policy and Management (HSPM) and Department of Health Promotion, Education, and Behavior (HPEB) have been awarded funding from the Health Sciences Center at Prisma Health to conduct research aimed at improving health system performance, population health, or biomedical science that translates to clinical practice and improved patient outcomes. The Prisma Health Research Seed Grant Program has allocated up to $20K for each of the five projects led by HSPM assistant professors Melanie Cozad, Chen Liang and Nabil Natafgi and HPEB associate professor Caroline Rudisill.
The initiative is open to Prisma’s university partners and is intended to facilitate collaborative pilot research that will be disseminated locally, regionally and/or nationally. Each project includes an investigative partnership between the faculty member and a clinical collaborator from Prisma Health.
“We’re excited to continue building on our successful partnership with Prisma Health,”
says Rudisill, Arnold School Director of Population Health Sciences in Greenville.
“These seed grants allow us to conduct world-class research in the Prisma Health environment
that aims to improve the health of the state of South Carolina and inform national
discussions on health care and public health best practices.”
Educating and Enabling Patients through mHealth Technology – Shared Decision Making in Action
Cozad, a health economist based at the Arnold School satellite campus in Greenville who is interested in patient-centered care, will lead a project focused on identifying patient therapeutic goals and preferences through mHealth technology. Using their mobile app to enhance communication between physicians and patients with rheumatoid arthritis as a model, the team’s goal is to create a similar tool to educate patients with metastatic breast cancer and enable them to participate in shared decision-making about their treatment.
“Treatment for metastatic breast cancer is highly patient preference sensitive,” Cozad
says. “As we learned for rheumatoid arthritis, a shared decision-making tool embedded
within a mobile application is meaningful and will be used by patients if there is
a feature that incorporates personal goals for therapy and the patients’ and their
caregivers’ preferences for treatment vis-à-vis the benefits and harms posed by each
Leveraging Electronic Health Records to Assist with Identifying Patients with Opioid Use Disorder
With interests in the development and application of information technologies, data science, and artificial intelligence, Columbia-based Liang will leverage electronic health records to assist with the diagnosis of individuals who have opioid use disorder – an epidemic that impacts approximately two million Americans. Many of these individuals have high rates of healthcare and emergency department utilization. This makes hospitals a critical site for opioid use disorder identification and treatment referral yet up to 70 percent of patients are missed or delayed in diagnosis.
“Electronic health records hold great promise in assisting providers to identify individuals
with possible opioid use disorder; however, the temporal patterns of these records
are often not considered and clinical notes are underutilized,” Liang says. “We will
develop a deep learning model to harness individuals’ chronological records and employ
Natural Language Processing algorithms to identify and extract subtle cues related
to opioid use from clinical notes.”
Reducing Medication-related Harm through Personalized Discharge Summaries
Liang will lead a second project aimed at creating customized discharge summaries to decrease the prevalence of medication-related harm, which occurs at high rates during care transitions. An estimated 26 percent of hospital readmissions are associated with preventable medicine-related harms – leading to worse outcomes for patients and a higher economic burden on hospitals. While discharge summaries are currently required, they do not effectively reflect patients’ personalized needs and context-specific information.
“Some of the recurring factors that contribute to these preventable incidents include
poor health literacy, lack of transportation, lack of support to obtain medications,
mental health conditions, and substance use, and these factors vary by setting, clinic
site and type, patient demographics, social determinants of health, and communities,”
Liang says. “With this study, we will develop machine learning models to learn from
electronic health records and clinical decision rules to suggest personalized information
for discharge summaries.”
Enhancing Teach-Back Methods in Virtual Care Visits
Natafgi’s work also focuses on patient-centered care, with specific interests in quality of care and patient safety, performance measurement and reporting, and telehealth effectiveness and evaluation. With the rapid shift to telemedicine – accelerated by the COVID-19 pandemic – researchers and clinicians recognize the importance of ensuring enhanced physician-patient communication. With this project, Natafgi will design and evaluate an innovative platform that will provide health literacy training for primary care medical residents.
“The COVID-19 pandemic has changed the landscape of primary care visits, particularly
for chronic conditions, and telemedicine health literacy requires specific communication
skills that are not currently a part of medical student and/or residents training
or institutional continuing education,” Natafgi says. “Improving the healthcare provider’s
ability to effectively communicate with patients during a virtual visit can ensure
patients’ understanding of treatment and/or diagnosis and improve the patients’ ability
to find, understand and apply health information and services regarding their chronic
Screening and Intervening on Social Determinants of Health-related Needs
In addition to funding from the UofSC Big Data Health Science Center, Rudisill, a health economist who studies individual decision-making regarding health-related behaviors, will lead the evaluation of a technology and electronic health record-based strategy to screen for and intervene on social determinants of health-related needs. The researchers will use three data sets to build links between survey responses, social determinants of health-based referrals and resource use. The project will investigate characteristics of the individuals making and receiving referrals and understand which parts of the community are experiencing the highest levels of referrals.
“Social determinants of health impact up to 80 percent of health outcomes,” Rudisill says. “We are excited to be part of Prisma Health’s work on identifying and intervening on social determinants of health-related needs in our communities.”