Visualizing Invisible Data

Once a drug is approved for its efficacy and effects, it enters the market, and the number of patients taking it increases.

As the number of cases of drug use grows and data accumulates, it becomes possible to evaluate the positive and negative effects of the drug by examining usage-related data.

This is where "pharmacoepidemiology" comes in, evaluating safety and effectiveness while considering differences in drug effects among patients with varying disease severity and physical constitutions.

What methods does Professor Hisashi Urushihara, a leading expert in this field, use to advance his research?

(As of December 2023)

Photo: Keio University Faculty of Pharmacy

After a drug is approved, a vast amount of clinical data, such as prescription history and disease progression, is accumulated. For a long time, this data was not utilized for research purposes. "It's incredibly rewarding when you can eliminate various invisible influences, focus only on what you want to see, and find that the results match your predictions." Professor Urushihara is one of the pioneers in Japan who sought to apply such large-scale clinical data to epidemiology. The catalyst was a conversation with a pharmacist about "wanting to verify the side effects of a certain drug." After that, he undertook one effectiveness verification study after another using clinical data. He joyfully recounted feeling like a child before a treasure chest. In pharmacoepidemiology, we develop methods to properly deduce a drug's effects by standardizing the conditions for comparison. When you catch a cold, not everyone experiences the same effect from a drug; individual conditions like age and health status differ, and many factors influence a drug's action.

The effect of a drug becomes clear only when you compare people who take the drug with those who do not, under similar physical conditions and constitutions. So, how should we group similar individuals? For example, when studying the effectiveness of the influenza vaccine, we calculate the probability of receiving the vaccine. We then compare those who received the vaccine with those who did not, among people with an equivalent probability, say 50%. By comparing among individuals with similar physical characteristics who have the same vaccination probability, the universal effect of the vaccine, independent of individual traits, becomes visible. "This 'propensity score' method has become widespread in Japan," says Professor Urushihara. In the future, he hopes to predict the side effects of all kinds of drugs. As times change, science advances, and the diseases affecting people also change. To deliver safer and more effective drugs as quickly as possible, he continues to tackle research on the evaluation methods themselves.

Q. What does pharmacy mean to you?

A. To scientifically explore the usefulness of drugs to contribute to human health.