Determining safe yet effective medication dosages for children is an ongoing challenge for pharmaceutical companies and doctors alike. A new drug is usually tested on adults first, and the results of these studies are used to select doses for pediatric studies. The underlying assumption is typically that children are like adults, only smaller, which is often true, but also overlook the differences that arise from the fact that children’s organs are still developing.
In addition to the problem, pediatric studies do not always shed light on other differences that may affect drug dose recommendations. There are many factors that limit children’s participation in drug trials – for example, some diseases are simply rarer in children – and as a result, the data sets generated are generally very sparse.
To make medicines and their development safer for children, researchers from Aalto University and the pharmaceutical company Novartis have developed a method that makes better use of available data.
This is one method that can help determine safe drug doses more quickly and with fewer observations than before. “
Aki Vehtari, co-author, associate professor of computer science at Aalto University and the Finnish Center for Artificial Intelligence FCAI
In their study, the research team created a model that improves our understanding of organ development.
‘The size of an organ is not necessarily the only thing that influences its performance. Children’s organs are just not as efficient as those of adults. In drug modeling, if we assume that size is all that matters, we can end up over-dosing, ” explains Eero Siivola, lead author of the study and PhD student at Aalto University.
While the standard approach to assessing pediatric data relies on subjective evaluations of model diagnostics, the new approach, based on Gaussian process regression, is more data-driven and therefore less prone to bias. It is also better at handling small sample sizes as uncertainties are taken into account.
The research stems from FCAI’s research program on Agile and probabilistic AI and provides a good example of a method that makes the most of even very scarce datasets.
In the study, the researchers demonstrate their approach by re-analyzing a pediatric study examining Everolimus, a drug used to prevent organ transplant rejection. But the potential benefits of their method are far-reaching.
“It works for any drug we want to study the concentration of,” Vehtari says, such as allergy and painkillers.
The approach can be particularly useful in situations where a new drug is being tested on an entirely new group – children or adults – that is small, potentially making the testing phase much more efficient than it is today. Another promising application involves expanding the use of an existing drug to other symptoms or illnesses; the method could support this process more effectively than current practices.
Siivola, E., et al. (2021) Qualifying Drug Dosage Regimens in Pediatrics Using Gaussian Processes. Statistics in Medicine. doi.org/10.1002/sim.8907.