Date Approved


Degree Type

Open Access Thesis

Degree Name

Master of Science (MS)

Department or School

Health Sciences

Committee Member

Irwin G. Martin

Committee Member

Stephen A. Sonstein


Drug-induced liver injury (DILI) is a rare cause of liver disease that is difficult to diagnose due to the lack of a confirmatory test. The Drug-Induced Liver Injury Network (DILIN) was developed to enroll patients with DILI to understand the clinical outcomes of the disease. The present study aims to develop a tool to identify patients with DILI for enrollment into the DILIN study. The use of ICD-10 codes and liver injury terms were used to search the electronic medical records (EMRs) of patients at the University of Michigan. It was hypothesized that free-text searching using key words would be superior to the use of ICD-10 codes. Two million patient encounters within a six month window were searched using ICD-10 codes related to toxic liver disease. A total of 489 patients were identified and after manual review 32 cases were confirmed. Using the medical record numbers from the patients identified in the initial search, six liver injury terms were used to search free-text in the EMRs. Twelve cases of DILI were identified using “drug-induced liver injury” and none were found using “hepatotoxicity.” The results of this study showed that it is feasible to identify prospective DILI cases using ICD-codes and free-text searching of liver injury terms.