Researchers at Seoul National University Bundang Hospital have confirmed they can know the frequency of drugs’ side effects resulting from their long-term use, faster than before by analyzing big data.

The research team used the Common Data Model (CDM), a standard analysis method for big data. It analyzed the prevalence of side effects based on blood test results of pediatric patients. They have been taking anticonvulsants for the long-term treatment of epilepsy.

Epilepsy is one of the chronic diseases of the nervous system, in which convulsions and seizures occur repeatedly, and up to 1.2 percent of the population suffers from it. Patients have to take anticonvulsants for years, as the primary treatment is to prevent epilepsy attacks with the drug.

The information on the appearance and frequency of side effects is essential for all drugs, especially those used for an extended period. Although patients generally experience relatively mild abnormalities, sometimes they can suffer life-threatening severe side effects. However, side effects are assessed only in a very limited number of patients in most cases, such as pre-marketing clinical trials or post-marketing investigations. It is nearly impossible to investigate all patients who use drugs.

The research team, led by Professors Hwang Hee, Kim Hun-min, and Yoo Soo-young, analyzed the big data of medical information systems that have already been de-identified and structured, the hospital said in a news release on Wednesday.

They used the Observational Medical Outcomes Partnership (OMOP)-CDM database of about 170 million patients for the study. OMOP-CDM is a data model that transforms electronic medical records information, such as different terms and formats for each medical institution, into a standardized structure.

The study used data from blood tests conducted during the period of drugs and anticonvulsants used by 1,344 out of around 5,000 patients treated at the hospital’s Department of Pediatric Neurology, specializing in epilepsy, from 2003 to 2017.

Based on the results of blood tests conducted during the five most commonly used anticonvulsants, the research team confirmed abnormalities such as anemia, thrombocytopenia, leukopenia, hyponatremia, thyroid dysfunction, and liver dysfunction.

Using the CDM data, the researchers could analyze the overall information of abnormalities in blood tests due to anticonvulsants in all children with epilepsy, the hospital said. Furthermore, they confirmed the exact prevalence of side effects that could be caused by the drugs already known, as well as the previously unknown side effects.

"We were able to complete the study in months with the CDM model while similar observation of drug side effects typically usually requires more than a year," Professor Hwang said.

The study was meaningful as it showed the possibility of replacing some of the existing post-marketing surveys in a short period with less expense if it is spread to multi-agency research in the future due to the nature of CDM, he noted.

Professor Kim also said, “CDM model is fast and accurate, but careful design is also significant because there are some points that can be missed out in the process of specifying search conditions.”

This study was published in Epilepsia, the official journal of the International League against Epilepsy, as the world's first study of anticonvulsant side effects using the CDM model.