Lessons learned from implementing OMOP CDM in Japanese oncology​

In the vast landscape of modern medicine, a quiet revolution is taking place not in the operating room, but in the digital archives of hospitals. For decades, doctors have relied on clinical trials to prove that a new drug works, but these trials often involve carefully selected patients in controlled settings that do not always reflect the messy reality of everyday life. To understand how treatments truly perform for the general population, researchers need to study "real-world data"—the millions of records generated every day as people visit doctors, receive diagnoses, and take medications. The challenge has always been that every hospital records this information differently, using its own unique codes and formats, making it nearly impossible to combine data from different places to see the bigger picture. To solve this, the global research community developed a universal translator for medical information called the OMOP Common Data Model. Think of it as a standard language that allows a hospital in Tokyo to speak directly with a hospital in New York, translating their specific records into a shared format so they can be analyzed together. This approach promises to accelerate the discovery of effective treatments and improve patient care on a global scale, but it requires a massive, precise effort to convert local records into this new standard.

A team of researchers at the National Cancer Center Hospital East in Japan set out to test whether this translation process could work for the complex world of Japanese oncology. They focused on the electronic medical records of nearly 8,400 patients treated for breast cancer between 2015 and 2022. Their goal was not to treat patients or discover a new cure, but to see if the existing digital records from their hospital could be successfully transformed into the OMOP format and then checked for accuracy using a specialized quality dashboard. This was a critical test because, while other countries have successfully built large networks of such data, Japan had not yet demonstrated that its specific medical records could be converted and used for international research. The team took the raw text files containing patient histories, lab results, and treatment details and began the arduous task of mapping every single term to the global standard. They had to ensure that a Japanese diagnosis code for a specific type of cancer matched the correct international concept, and that a drug prescribed in Tokyo was recognized as the same substance used in Europe or the United States.

The results of this conversion were largely encouraging. The researchers found that they could successfully standardize more than 80 percent of the clinical information across major areas of patient care, including diagnoses, laboratory tests, and medication prescriptions. They processed millions of individual data points, converting over 39 million laboratory test results and nearly 300,000 prescriptions into the new format. This demonstrated that the Japanese system is capable of joining the global network of observational research. However, the process was not without its hurdles. The team discovered that surgical procedures were the most difficult area to translate; while they could map most other data, the specific terms used for surgeries in their system did not align well with the international standards, leaving that portion of the data incomplete. Additionally, they encountered gaps in the records, such as missing dates for certain tests or unclear units of measurement for vital signs like blood pressure, which had to be manually addressed or excluded.

To ensure the converted data was trustworthy, the researchers ran nearly 2,000 automated checks to verify its validity and completeness. Only 24 of these checks flagged an error or a failure, a very small fraction that suggests the overall quality of the converted dataset is high. The few errors that did appear were mostly related to missing information in specific tables, such as details about patient deaths or drug dosages, which the team identified as issues that could be fixed by refining their conversion software or improving the source records. The study also highlighted a significant cultural and administrative barrier: the Japanese government uses its own set of standard medical codes that are not yet linked to the international vocabulary used by the OMOP system. This means that every time a Japanese hospital wants to join this global network, they must manually create a bridge between their local codes and the international ones, a time-consuming task that requires significant human effort.

Ultimately, this study proves that Japanese medical records can be transformed into a format that allows them to be analyzed alongside data from the rest of the world. It confirms that the infrastructure exists to support large-scale international studies on cancer treatment effectiveness using real-world data from Japan. However, the path forward requires more than just technical conversion; it demands a coordinated effort to link Japanese medical terminology with global standards so that the translation process becomes smoother and less reliant on manual work. As more hospitals in Japan adopt these methods and as government initiatives move to standardize medical data nationally, the potential for Japan to contribute to a global understanding of cancer treatment grows stronger. The work done here serves as a blueprint, showing that while the journey to a unified global medical database is complex, it is entirely possible, paving the way for future research that could benefit patients everywhere.