Often when people get sick or need health information, they turn to the Internet for answers, with health searches being one of the most popular uses of the Internet today. This high demand for health information, coupled with the rising popularity in social networking, has resulted in the creation of many online health communities in recent years. At the same time, the number of people participating in these health-related communities is steadily rising.
Online health communities offer users the opportunity to interact with others who are dealing with the same disease or health issues. These peer-patient conversations can be helpful for learning about patients’ personal experiences managing their illness, but what they currently lack are clinical moderators and the insight from clinical experts on more clinically oriented information.
Jina Huh, Assistant Professor in the Department of Media and Information, is working to solve this issue and has been awarded a National Institutes of Health (NIH) National Library of Medicine Career Development Award (K01) to develop a semi-automated system that would infuse clinical expertise into peer-patient conversations in online health communities. The system will aid patient self-management by delivering balanced information from patients’ personal experiences and clinical expertise. The three-year, $467,801 grant began in October 2014.
“In face-to-face patient support groups, clinicians moderate patients who share their experiences so they can clarify clinical questions and answer any questions patients might have,” Huh said. “From the patient’s perspective, they not only get other peer-patient experiences, but also clinical expertise from the clinical moderators. But in the online community setting, that sort of support group dynamic between patients and moderators is missing.
“My proposal is to develop a system that would enable that social dynamic present in face-to-face settings to be augmented in online settings.”
Huh already has conducted a number of studies with health professionals looking at when such a system should step in.
“They just want the system to interject when there are medical terms being used that indicate any symptoms or treatments, such as ‘numbness’ or ‘severe pain,’” Huh said.
In creating the system, Huh will use the UMLS (unified medical language system), which lists all medical terms and categorizes them.
“Using that system, I can check the text to see whether those terms occurred in each specific post,” Huh said. “So when patients are talking to one another if there is a time when the machine determines they need clinical expertise, the machine will detect it and will bring in existing sources on the Web or from our curated database and link in that expertise onto the peer-patient conversation.”
During the first year of the grant, Huh will develop the training data set to train the machine to learn how to classify the patients’ posts and find out when the machine can intervene. The second year, she will work on the user interface and develop a feedback mechanism so users can give feedback to the machine’s results and the way the machine works with users. Over the third year, Huh plans to deploy a working prototype.
For this study, Huh is working with the diabetes community and partnering with Alliance Health Network and its online diabetes community, DiabeticConnect.com. Alliance Health Network plans to give Huh access to the users of this online community who will test the system.
“We will have a control group and a test group,” Huh said. “The control group will only look at the regular online health community while the test group will look at the system added on top of the regular online health community and will be able to see the clinicians expertise weaved on top of their conversations.”
Huh then will measure behavioral and mental outcome changes, including self-efficacy, diabetes management behavior, quality of life, social support and stress.
Huh’s mentors on the project include Professor Wanda Pratt, the Information School, University of Washington (mentor chair); University Distinguished Professor Barbara Given, MSU College of Nursing; and Professor Joyce Chai, Department of Computer Science and Engineering, MSU College of Engineering. Consultants on the project are Associate Professor Marianne Huebner, Department of Statistics and Probability, MSU College of Natural Science; Associate Professor of Nursing Amber Vermeesch, University of Portland, Oregon; and John Crowley, Vice President of Consumer Marketing and Online Communities, Alliance Health.
Earlier research by Huh that laid the groundwork for this project has received praise as one of the best articles published in 2013 in biomedical informatics by the annual International Medical Informatics Association (IMIA) Yearbook of Medical Informatics. The article, “Text classification for assisting moderators in online health communities,” was published by the Journal of Biomedical Informatics and cowritten by Pratt and Assistant Professor Meliha Yetisgen, Department of Biomedical Informatics and Medical Education, School of Medicine, University of Washington. That work explored low-cost text classification methods to determine whether a thread in an online health forum needs moderators’ help.
“I am just really glad this line of work is being recognized by the biomedical informatics community,” Huh said. “I feel like my work is getting acknowledged and my hard work is paying off, and I am happy about that.”
Huh serves as a Trifecta Intellectual Leader (TIL) for the Trifecta initiative, which fosters interdisciplinary research by building relationships among experts in three MSU colleges – Communication Arts and Sciences, Engineering, and Nursing – to collaboratively advance the delivery of health services for underserved populations.
“The grant reviewers really liked that MSU offers this Trifecta initiative and that I am a Trifecta faculty. They noted this as part of the positive reviews for the grant,” Huh said. “Because of this, they thought that MSU is a very good environment for me to conduct this research.”
Huh manages the iCARE Lab and is part of the BITLab (Behavior Information Technology Lab). She also is an Adjunct Assistant Professor in MSU’s College of Engineering and the College of Nursing.