College of

Engineering, Architecture and Technology

Three faculty members in the School of Industrial Engineering and Management (IEM) and one in the Department of Statistics (STAT) have secured a major grant from the National Institutes of Health (NIH) worth $1.194 million to study machine-learning algorithms to detect diabetic retinopathy (DR) in the early stages with routine lab results.

 

DR is a microvascular complication of diabetes, and it is the most common cause of vision loss among diabetic patients and the leading cause of blindness among American adults. Many diabetic patients do not comply with CDC’s recommendation for annual ophthalmic exams because DR is asymptomatic in the early stages, and thus they miss the most effective period to halt DR progression and prevent vision loss. Moreover, ophthalmic equipment for DR exams is predominantly limited to urban areas, restricting access by patients in rural communities with limited incomes.

 

The OSU team led by Dr. Tieming Liu (IEM) aims to develop a non-image-based, artificial intelligence (AI) tool for primary care physicians to assess patients' risk for DR using comorbidity data and routine lab results, which are widely available. This tool will help physicians recommend ophthalmic exams and individual screening frequency for at-risk patients confidently. The training data is provided by Cerner Corporation and the Center for Health Systems Innovation (CHSI) at the Spears School of Business. However, similar to other electronic-health-record (EHR) databases, the quality of this dataset suffers from missing values, imbalanced and unlabeled data.

 

The grant, which is spread over four years, will support the research to improve the performance of the DR screening tool by ameliorating the data quality issues in EHR. Specific aims of the project include designing tensor-enabled data imputation, augmentation, and prediction algorithms led by Drs. Tieming Liu, Chenang Liu, and Bing Yao in IEM and a Bayesian hierarchical modeling approach to classify the unlabeled patients' trajectories led by Dr. Ye Liang in STAT.

 

This grant will enhance OSU’s research in both healthcare sciences and AI. Improving EHR data quality offers a great promise for AI-based prediction and personalized medicine of other chronical diseases.  This project will also help CHSI to achieve its mission to transform rural and Native American healthcare. AI techniques offer a realistic and effective workaround to handle the dilemma of delivering a higher quality of healthcare at reduced costs to the disadvantaged communities who have limited access to healthcare resources. EHR-based health analytics offers vast potential for improving healthcare, e.g., prediction of adverse clinical events, more accurate patient triaging, disease progression prediction, and treatment optimization.