A Real-World Biotech Detective


A Real-World Biotech Detective
Take your curiosity and just observe what's happening around you to analyze what it might mean. For any kind of science, using an analytical mindset and training yourself to question how things work is invaluable.
—Marieke Schoonen
female detective looking at suspicious footprints

Detective work is not found only in law enforcement. Epidemiologists and data scientists all investigate what happens when a medicine enters the real world, “deciphering” everything from who receives the medicine and how it’s prescribed to unanticipated side effects and long-term benefits.

“It's all about observing the real patients out there,” says Marieke Schoonen, senior director of observational research at Amgen. In contrast to clinical studies, in which researchers randomize patients into a treatment protocol to study safety and efficacy, observational research is down to interpretation, like a detective piecing together clues from different datasets. 

Marieke Schoonen

“In observational research, there are a lot of different shades of gray,” she explains. “You have to always be asking questions, like ‘how can my results have been affected by potential systematic differences between the groups that I'm comparing or the way the data were recorded?’”

Much of the observational data Schoonen and her team works with comes from large healthcare databases, such as anonymized electronic health records from hospitals or anonymized claims data from insurance companies, in which physicians and healthcare workers record information about the patients and their care. Sometimes her team even needs to collect data that do not exist elsewhere, for example, through questionnaires. Using statistical principles, the researchers will look for patterns to better understand how certain drugs are working in patients and thus inform future healthcare decision-making.

“A clinician’s records are a nice diary, if you will, where you can go back in time and extract bits of information that are relevant to you to answer your research question,” Schoonen says. “Provided that consent has been provided, in the context of a formal study protocol, we are then allowed to go and extract that information and analyze it.”
Sometimes, like a detective, Schoonen comes across pieces of data that just don’t fit together, like a time when she saw a female patient who was noted as having prostate cancer.  “It can be very messy,” she says. “Sometimes you have to go back to the original records and ask, ‘Was there a transcribing error’? So, you are always applying logic to find the likely truth, which is very exciting.”

Schoonen’s background is in biomedical health sciences. While a secondary student in her home country of The Netherlands, she loved math—“everything with numbers.” When looking at university majors, she wanted something that would tap into her numerical sense while also helping people; biology did not feel like a right fit, so she instead pursued a health sciences master’s degree that allowed her to try different tracks, including epidemiology. “I love epidemiology because it gives you a toolbox of methods and techniques to analyze data, and that toolbox can be applied to any disease area you like,” she says. 

After her master’s program, Schoonen worked briefly at a cancer research center, which is where she first got to work with paper-based questionnaires, like the ones captured in the datasets she works with now. That work spurred her interest in pursuing a PhD at the London School of Hygiene and Tropical Medicine in “pharmacoepidemiology”—studying the effects of drugs in large populations. 

For the past 17 years working at Amgen, Schoonen has served as epidemiologist, closely working with colleagues across statistics, programming, medicine, regulation, economics, and policymaking to design, conduct, and interpret observational research studies that complement the evidence generated by clinical trials. The team she manages aims to push the boundaries of how real-world data can inform drug development. Schoonen says she loves the interpretational side of the work, putting their results into the context of what other evidence is out there. 

Success stories from her department include a peer-reviewed study that investigated risk of osteoporosis-related fractures. In the absence of a clinical trial that compared two commonly used osteoporosis medicines directly, the team applied advanced analytical methods to observational, real-world data from almost 500,000 postmenopausal women and found that one medicine was more effective than the other at reducing major osteoporotic fractures. That information can now inform physicians, insurers, and the broader osteoporosis community about treatment options. “It’s nice to recognize that our patients ultimately do benefit from the work we do,” she says. 

Schoonen’s biggest piece of advice to those who may want to pursue observational research work is to “not take everything at face value.” Rather than trying to influence the data, she says, “take your curiosity and just observe what's happening around you to analyze what it might mean. For any kind of science, using an analytical mindset and training yourself to question how things work is invaluable.”

Blog archives