Manual extraction gathered information on the drug treatment for ulcerative colitis

Hospital data lakes are a central source of information in Medaffcon’s studies, and often, data retrieval can be automated. However, what happens when more complex matters need to be extracted from the databases? In such cases, manual extraction is used, says Medaffcon’s Scientific Advisor, Riikka Mattila

When data retrieval cannot be automated due to diverse sources, manual extraction provides the practical solution.

A good example of manual extraction is a Real-World Evidence (RWE) study published last fall, examining the effectiveness of treatment for ulcerative colitis. The study involved 23 Finnish IBD centers specializing in the treatment of ulcerative colitis.

According to Mattila, all the information retrieval was done manually. In each hospital, a clinical specialist extracted data from patient records manually and recorded them on a separate research form. The study sought variables describing symptoms and the severity of the disease.

“In Finland, there are four data lakes: in Helsinki, Tampere, Kuopio, and Turku. If information is needed from other hospitals, databases are not available, and the information must be retrieved manually. That’s how it was in this study,” Mattila explains.

According to Mattila, data lakes allow for the automatic extraction and mathematical analysis of numerous pieces of information, particularly when dealing with straightforward variables like hemoglobin.

“However, when collecting for example complex symptoms, they cannot be automatically extracted from the data lake. For instance, there might be a diagnosis where the doctor has included symptoms. In such cases, a person must retrieve the information,” says Mattila.

She explains that in Finland, this type of registry study is technically similar to other registry studies, even though information is retrieved from patient records. Since the study is a registry study and not a clinical trial, it does not require approval from an ethical board. The approval process is therefore lighter.

“In other countries, retrieving information directly from patient records can be challenging. In Finland, such research only requires permission from Findata or the data controller. This falls under the Act on Secondary Use of Health and Social Data,” Mattila points out.

Medaffcon conducted research on the effectiveness of tofacitinib in the treatment of ulcerative colitis in collaboration with Pfizer and Finnish gastroenterologists. Mattila, along with Biostatistician Iiro Toppila and Data Scientist Olivia Hölsä, comprised the team from Medaffcon. They are part of the Data and RWE teams with a total of 15 experts.

Mattila joined Medaffcon three years ago. In her role, she primarily conducts RWE studies. She has extensive experience in various therapeutic areas, with about 15 years of research experience, including three years at the Max Planck Institute in Germany.

“What’s interesting in my current role is that I get to do science. And there’s a great atmosphere at Medaffcon,” Mattila shares.

Read more about Medaffcon’s teams and key figures from 2023 here.