Elizabeth Campbell, a student in the PhD in Information Science program at Drexel’s College of Computing & Informatics (CCI), has dedicated her research to the application of data analytics and machine learning methods to electronic health records (EHRs) to study medical diagnoses, including pediatric obesity.
Campbell, who also serves as a graduate research assistant at The Children’s Hospital of Philadelphia (CHOP), will be moderating a May 9 Executive Panel on Technology in Healthcare which will feature perspectives from healthcare technology leaders on the impact that technology has for patients, care providers, and healthcare organizations.
Campbell describes her path to technology as non-traditional, but no doubt her broad range of experience in public health and extensive field work in communities across the world have all helped to inform her research goals in effecting positive changes in the healthcare IT field.
Her additional research interests include complementary uses of data science methods and non-traditional data sources to conventional epidemiological approaches in public health research, with a focus on health disparities outcomes.
In January, she received the prestigious American Health Information Management Association Merit Scholarship which recognizes outstanding students in health information technology (HIT). She earned a Master of Public Health in Health Policy from the Johns Hopkins Bloomberg School of Public Health and a Bachelor’s in Public Health Studies from The Johns Hopkins University.
Drexel CCI recently sat down with Campbell to learn more about her background, research interests and thoughts on the future of technology in the healthcare field:
CCI: How did you decide to pursue your doctorate in information science?
EC: I have a background in public health, which I studied as both an undergraduate and for my Master’s. During my studies, I was a research assistant at the Johns Hopkins Center for Human Nutrition (where I completed my bachelor’s and master’s degrees in 2014 and 2016 respectively). I worked on data management for two public health nutrition interventions and became fascinated with the role that data played in those research projects: it connected the diverse moving parts of the studies and could be analyzed to tell the story of the intervention’s implementation and effectiveness. The experience sparked a passion for academic research and a curiosity to explore how data can be used to benefit human health. I knew that I wanted to do a PhD and pursue a career in research, and I found that the field of information science and Drexel’s degree program to be the right fit to develop my skills and knowledge in data science and to train in the methods that I wanted to apply in my research.
CCI: How did you become interested in researching Electronic Health Records (EHRs)?
EC: After completing my master’s, I was an ASPPH/CDC Allan Rosenfield Global Health Fellow in Strategic Information (now the PHI/CDC Global Health Fellowship) with the CDC’s Office of Health Informatics in Lusaka, Zambia. During my fellowship, I worked to support the CDC’s re-engineering and implementation efforts for SmartCare, Zambia’s national electronic health record (EHR) system. Although not without its challenges, SmartCare has been instrumental in efforts to combat the HIV/AIDS epidemic in Zambia by streamlining patient information for clinical care provision, acting as a source of data for HIV/AIDS surveillance locally and regionally, and by providing data to inform public health policy development at the national level. I felt really inspired by the impact that EHR data can have on clinical and population health outcomes. My training in health informatics and experience in working with EHRs helped me to focus my research interests as I segued into my PhD from the Fellowship.
CCI: What has your experience been like as a Drexel Women in Computing Society (WiCS) mentor?
EC: I became involved with the Drexel Women in Computing Society’s (WiCS) Mentorship Program during the first year of my PhD program. The program has provided me with the opportunity to connect with female, undergraduate students at CCI, which has been very meaningful for me. I think it’s important for undergraduate and graduate students at CCI to develop a strong community, particularly for women who are traditionally under-represented in the tech sector. Graduate students can be great sources of advice at both a personal and professional level for undergrads and I recognize and value the benefits that such mentoring relationships can have for the mentor and mentee alike. As someone who had a non-traditional path to computing, I am eager to share those experiences and encourage other women to take intellectual risks and explore their curiosities and passions. I will be serving as the President of CCI’s Doctoral Student Association in the coming school year and was also recently elected as the Graduate Liaison for WiCS. I’m really excited to assume these roles and foster collaboration between the two organizations. I hope to build even stronger relationships between our undergrads and PhD students at CCI in this capacity.
CCI: Could you also tell us more about your work at The Children’s Hospital of Philadelphia? How has that complemented your studies?
EC: In my first year in graduate school, my PhD advisor Ellen Bass and my CHOP mentor Aaron Masino helped me to be selected as part of the CHOP-Drexel Research Fellowship Program: Informatics and Analytics Collaborative Research. This allowed me to be mentored by my advisor and also researchers at CHOP. After the first year, my CHOP mentor identified a part-time student internship position in the Department of Biomedical and Health Informatics at CHOP. This support allowed me to continue with my research in which I apply machine learning methods to EHR data to study health outcomes and disparities. My current focus is pediatric obesity incidence and associated comorbidities. The ability to utilize CHOP data and receive mentorship at CHOP and at Drexel has helped me to increase my knowledge about healthcare and to synthesize my passion for public health with my graduate training in informatics. I have learned so much and am truly privileged to develop under the mentorship of my colleagues at both Drexel and CHOP.
CCI: What do you see as some of the biggest challenges and/or opportunities surrounding the use of EHRs, or even healthcare organizations adopting or implementing new technology in general?
EC: I think that integrating EHR systems into clinical workflows as well as ensuring privacy and security for personal health information are some of the biggest challenges surrounding the use of EHRs. The challenges that healthcare workers (doctors, nurses, etc.) have faced in training to use EHRs and modify their clinical practice as a result, speak to challenges that healthcare organizations face when adopting or implementing any new technology.
Additionally, from a research perspective, working with data generated from EHRs is extremely difficult. Its heterogeneity, high levels of missingness (missing data), and the presence of irregular time intervals for variable measurements create numerous challenges in obtaining and transforming relevant data for meaningful analysis. Utilizing EHR data in health informatics research requires tremendous time, effort, and computational resources.
CCI: How can machine learning and data analytics aid in helping us track/gain more insight from diagnoses?
EC: There’s a lot of buzz surrounding current and potential applications of machine learning and artificial intelligence in the healthcare sector, and I think there’s good reason. Predictive analytics have myriad potential benefits, including helping to reduce hospital readmissions, aiding in disease and adverse event detection, and assessing chronic disease risk. Additionally, AI has the potential to contribute to remote patient monitoring and improve efficiency in numerous tasks such as appointment scheduling. However, it’s important to note that such methods have the potential to enhance the work of and not replace clinical practitioners.
I also think it’s important to discuss potential challenges to integrating machine learning and other non-traditional analytical techniques into medicine. Tools that utilize such methods should work equally for all patients and avoid improved predictive outcomes for one group above others (famous examples outside of healthcare include racially biased recidivism prediction among AI-powered tools used by law enforcement or incidents of racial and gender bias that have occurred in facial recognition system development). In practice, AI has proved very difficult to scale in healthcare. Developing analytical tools that provide useful and interpretable information for clinicians and that clinicians can actually use in their everyday practice is hugely challenging although I am optimistic that these challenges can be overcome in time.
What are you looking forward to learning at the May 9 Executive Panel on Technology in Healthcare?
EC: We have a diverse and experienced group of panelists assembled, who can provide valuable insights on the role of technology in healthcare from unique perspectives. Their knowledge and experience can help us to understand the evolution of technology in healthcare, as well as its potential future directions. I am especially excited to learn about issues of technology in healthcare from an executive standpoint, and to better understand how healthcare data and technology influences organizational decision-making and efficiency.