iSchool Professor Presents Health Informatics Research at MEDINFO 2019

Over the summer, Dr. Zhe He of the iSchool had four papers, four posters, and 1 panel accepted to MEDINFO 2019, the world’s largest conference in medical informatics. These presentations centered around Dr. He’s eHealth Lab. He explains the research as:

“The research in the eHealth Lab is centered around a core question: ‘How can we find ways to improve healthcare with the use of big health data?’ We have been developing novel data-driven methods to mine useful information from the electronic health data in clinical trial registries, public patient databases, and clinical data warehouses. We are using machine learning and deep learning to predict health outcomes such as mortality and readmission for patients with cardiovascular diseases. We are also using both qualitative and quantitative methods to understand the barriers for clinical trial participation. We are developing methods to assess and bridge the vocabulary gap between health information consumers and health professionals to better engage patients in their healthcare.

This year, our eHealth Lab had a strong presence in MEDINFO 2019, the 17th World Congress of Medical and Health Informatics. MEDINFO is the world’s largest academic conference on health and biomedical informatics organized by the International Medical Informatics Association.

The overarching theme of MedInfo2019 was Health and Wellbeing: E-Networks for all, stressing the increasing importance of networks in healthcare on the one hand, and the patient-centered perspective on the other hand. Combining and integrating cross-institutional data remains a challenge for both patient care and research. Patient-generated data, e.g. originating from mobile apps, sensor-based wearables, and smart home environments, contribute precious information to health care professionals. Not only do these data enable rigorous patient monitoring under normal life conditions, but they also help patients play a more active role in the care process.

MEDINFO 2019 received over 1100 submissions from 62 countries across all IMIA regions – the largest number of submissions in the more recent history of MedInfo conferences. With support from four track chairs, 51 track members, and 990 active reviewers, they have conducted a thorough review process. Virtually all submissions were reviewed by at least three reviewers and assessed by one SPC track member. Based on these recommendations, final decisions were made by the SPC track chairs and the SPC co-chairs during a three-day face-to-face meeting in Paris, France. Finally, 285 full papers/student papers, 47 podium abstracts, 296 posters, seven demonstrations, 45 panels, 21 workshops and nine tutorials were accepted.”

Dr. He’s eHealth Lab had four papers and four posters presented in the conference. Dr. He also organized a panel on the career of health informatics with four other panelists.

Dr. He explains two of the papers below:

In paper #1, Neelufar and Laura are both doctoral students in FSU iSchool. Daniel Bis is an undergraduate student in FSU Computer Science. Jiang Bian is an Associate Professor at University of Florida. In this paper, we seek to use machine learning and deep learning over electronic health records data to predict the one-year mortality among patients with myocardial infarction diagnoses when they were admitted to the emergency room.

Seyedeh Neelufar Payrovnaziri, Laura A. Barrett, Daniel Bis, Jiang Bian, and Zhe He. Enhancing Prediction Models for One-Year Mortality in Patients with Acute Myocardial Infarction and Post Myocardial Infarction Syndrome. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. 2019 Aug 21;264:273-277. doi: 10.3233/SHTI190226. PMID: 31437928.


Predicting the risk of mortality for patients with acute myocardial infarction (AMI) using electronic health records (EHRs) data can help identify risky patients who might need more tailored care. In our previous work, we built computational models to predict one-year mortality of patients admitted to an intensive care unit (ICU) with AMI or post myocardial infarction syndrome. Our prior work only used the structured clinical data from MIMIC-III, a publicly available ICU clinical database. In this study, we enhanced our work by adding the word embedding features from free-text discharge summaries. Using a richer set of features resulted in significant improvement in the performance of our deep learning models. The average accuracy of our deep learning models was 92.89% and the average F-measure was 0.928. We further reported the impact of different combinations of features extracted from structured and/or unstructured data on the performance of the deep learning models.

In paper #2, I worked with Dr. Zhan Zhang, Yu Lu from Pace University, Yubo Kou from Penn State University, Danny Wu from University of Cincinnati, and Dr. Jina Huh-Yoo from Drexel University to understand patient information needs about their laboratory test results using their questions posted on a social Q&A website.

Zhan Zhang, Yu Lu, Yubo Kou, Danny Wu, Jina Huh-Yoo, and Zhe He. Understanding Patient Information Needs about their Clinical Laboratory Results: A Study of Social Q&A Site. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. 2019 Aug 21;264:1403-1407. doi: 10.3233/SHTI190458. PMID: 31438157.

Abstract: Clinical data, such as laboratory test results, is increasingly being made available to patients through patient portals. However, patients often have difficulties understanding and acting upon the clinical data presented in portals. As such, many turn to online resources to fill their knowledge gaps and obtain actionable advice. In this work, we present a content analysis of the questions posted in a major social Q&A site to characterize lay people’s general information needs concerning laboratory test results and to inform the design of patient portals for supporting patients’ understanding of clinical data. We identified 15 information needs related to laboratory test results, and clustered them under four themes: understanding the results of lab test, interpreting doctor’s diagnosis, learning about lab tests, and consulting the next steps. We draw on our findings to discuss design opportunities for supporting the understanding of laboratory results.

Dr. He also organized and coordinated the career development panel titled “Developing Careers in Biomedical and Health Informatics” in MEDINFO 2019. “We invited four leaders in informatics, including Director of the National Library of Medicine Lister Hill National Center for Biomedical Communications Dr. Olivier Bodenreider, Amsterdam UMC’s Associate Professor Dr. Ronald Cornet, Executive Director of Yale Center for Biomedical Data Science Dr. Xinxin Zhu, and Dr. Songmao Zhang, Professor of Chinese Academy of Sciences, to share with us their career trajectories, experiences, and tips. You can find the slide deck of this panel here:

The other two papers presented at MEDINFO 2019:

1) Rubina F. Rizvi, Yefeng Wang, Thao Nguyen, Jake Vasilakes, Zhe He, and Rui Zhang. Analyzing Social Media Data to Understand Consumers’ Information Needs on Dietary Supplements. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. 2019 Aug 21;264:323-327. doi: 10.3233/SHTI190236.  PMID: 31437938.

2) Francisco Modave, Yunpeng Zhao, Janice Krieger, Zhe He, Yi Guo, Jiang Bian. Understanding Perceptions and Attitudes in Breast Cancer Discussions on Twitter. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. 2019 Aug 21;264:1293-1297. doi: 10.3233/SHTI190435. PubMed PMID: 31438134.

The other four posters:

1) Zhe He, Laura A. Barrett, Rubina F. Rizvi, Seyedeh Neelufar Payrovnaziri, and Rui Zhang. Exploring the Discrepancies in Actual and Perceived Benefits of Dietary Supplements Among Obese Patients. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. In press. (Poster).

2) Lynette Hammond Gerido and Zhe He. Improving Patient Participation in Cancer Clinical Trials: A qualitative analysis of HSRProj & RePORTER. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. In press. (Poster).

3) Nur Hafieza Ismail, Ninghao Liu, Mengnan Du, Zhe He, Xia Ben Hu. Identification of Cancer Survivors Living with PTSD on Social Media. MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. In press. (Poster).

4) Zhan Zhang, Yu Lu, Caleb Wilson, and Zhe He. Making Sense of Clinical Laboratory Results: An Analysis of Questions and Replies in a Social Q&A Community.  MEDINFO 2019. August 26-30, 2019. Lyon, France. Studies in Health Technology and Informatics. In press. (Poster).