Launched in 2022, the SC&I Scholarly Futures Program supports the development of Rutgers partnerships and research to gain a better understanding of how healthcare and health communication at the individual, organizational/clinical, community, and populations levels are being transformed by Information and Communication Technologies (ICTs), Artificial Intelligence (AI), and Machine Learning (ML) and how this process of change can magnify and reduce health inequalities.
The program was organized by SC&I health communication scholars Associate Professor of Communication Matthew Matsaganis, Associate Professor of Communication Lisa Mikesell, and Associate Professor of Library and Information Science Vivek Singh.
“The impetus for the three of us coming together,” Matsaganis said, “was created during a conversation we had around the great but also largely unrealized potential of ICTs in general, and advancements involving AI and ML more specifically, to improve patient, healthcare professional, and health organization-level outcomes, and to mitigate related health inequities.”
Mikesell added, “the recent COVID-19 pandemic has amplified interest in the adoption of ICTs and their impact in healthcare and on health disparities, as various forms of technology were employed to address pandemic-related challenges, such as the impact of public orders for physical distancing on healthcare seeking. The pandemic also strengthened concerns that ICTs can exacerbate health inequalities and generate new ones.”
“The three of us agreed,” said Singh, “that tackling this issue from multiple angles would be necessary in order to articulate recommendations for how the potential of ICTs could be more fully realized in the future. More specifically, our research activities have unfolded both in clinical settings and with an emphasis on the architecture of ML- and AI-powered digital health tools. And we have done so in collaboration with colleagues, including healthcare and public health professionals from across the university.”
Together with their partners, the team of faculty from the Communication and Library and Information Science Departments of the School of Communication and Information have set out to:
- Investigate how digital health tools are currently used by health professionals and patients in local New Jersey communities and how they envision using them in the future.
- Identify the individual, organizational, and community-level factors that enable and limit the use of ML- and AI-powered digital health tools.
- Propose new ways to ensure equity and fairness in the development of digital health tools.
In pursuing these goals, Matsaganis, Mikesell, and Singh are also examining best practices for ensuring equity and fairness in the development of ICTs for health.
The team’s collaborators include faculty from the School of Health Professions (Associate Dean for Global Affairs and Professor Riva Touger-Decker Ph.D. and Assistant Professor Ryan White Ph.D.) and from the School of Public Health (Professor Vincent Silenzio, MD, MPH). The SC&I research team underscored that, “such a multidisciplinary team is necessary to develop a robust theoretical framework, ensure methodological rigor, and guarantee analytical validity grounded in the team’s deep understanding of ML and AI, as well as healthcare practices.”
The broader research team also includes a number of Rutgers master’s and Ph.D.-level graduate students.
Describing their research methods, the team said they are combining established qualitative and quantitative social scientific methods with newer computational methods. For example, they said, in order to better understand how patients from diverse backgrounds living in diverse urban New Jersey communities and primary healthcare providers are currently using health ICTs and what shapes this use, they are conducting interviews with patients and healthcare providers in healthcare facilities, including university-affiliated and federally qualified health centers.
At the same time, other members of the team studied the equity of Machine Learning based health prediction algorithms across gender and noticed differences in the performance for men and women, and designed computational mechanisms to make the algorithm performance more equitable across genders.
The team has already published key and surprising findings, Matsaganis said. “Several sociodemographic predictors that the ICT adoption literature suggests are emerged in our analysis (e.g., age, socio-economic position, formal education). But even among patients with considerable experience with ICTs, use of such technologies to pursue health-related goals and particularly to facilitate communication with healthcare providers is highly episodic.”
He added, “additionally, interviews with healthcare professionals, so far, have highlighted that the rapid adoption of ICTs during the COVID-19 pandemic transformed patient-provider communication across clinical sites significantly, primarily through deployment of telemedicine/telehealth applications. However, this adoption was forced by the emergent nature of the crisis, and clinical sites did not have the time to develop the necessary organizational complements (e.g., structures, processes) to ensure maximum payoff. Our data suggest that one of the key challenges to realizing the potential of ICTs is to be found in the seeming inability of patients and healthcare professionals to imagine how ICTs could improve healthcare and patient-provider communication in the future.”
Mikesell added, “one way to see this is that when ICT adoption in healthcare is widespread it almost has to be ‘forced,’ as was the case during the pandemic. Without this external force, it doesn’t often happen because there is inertia in clinical practice settings and lack of vision/imagination regarding the value of tech adoption. It tends to be perceived as a burden rather than a benefit.”
That said, as Matsaganis and Mikesell note in a forthcoming publication they co-authored with two SC&I Ph.D. students on the impact of ICT adoption on patient-provider communication and inequalities, “many historically marginalized patients are especially likely to view ICTs as important for addressing patient-provider communication that they perceive to be of poor quality. This is in part because ICTs can support patient agency and self-advocacy. As such, our team really wants to explore ways to foster imagination among various stakeholders so that both the benefits and burdens are sufficiently weighed, particularly when it comes to supporting minoritized patient populations navigate healthcare encounters that may not best support their needs.”
Singh noted the progress the team has made on a different front. “We found that while the mental health prediction algorithms provided reasonable accuracy overall, there were noticeable differences in the performance across genders. However, it is feasible through various approaches to help counter such bias.”
The SC&I team noted that they hope to report on these findings at conferences this coming spring. Their work has implications for existing theories of technology adoption, particularly in healthcare, and for interventions promoting ICT adoption for improved patient-provider interactions.
Finally, Matsaganis noted that the project is also informing curriculum development at SC&I. In parallel with the research project, the team has also leveraged partnerships he and Dr. Riva Touger-Decker have built with universities in Greece, in the context of the IAPP-Greece Initiative (https://iapp-greece.rutgers.edu/), to design and implement a course on “Health and Communication Technologies” at the School of Communication & Information, in collaboration with colleagues from Harokopio University of Athens. The course is open to Rutgers and Harokopio University students and is an example of how SC&I faculty research informs the courses offered to budding researchers and professionals that will work in fields related to health communication and technology.