We probably know pediatrics drugs just as much as we do the Mariana Trench, yet, the medical opportunities behind generating medications for younger patients are also the buttons which deter us from doing so.
In US and EU, most pediatrics drugs are not tested on children, even if they are, clinical trials recruit individuals who are presently receiving medical care, instead of healthy volunteers as in the case of adults. Thus, it is hard for clinical pharmacologists to establish an understanding of the initial dose-exposure relationship.
In Japan, the struggle is still on the use of off-label drugs, as pointed out by Dr. Masakazu Hirata, review director at the office of cellular and tissue-based products and pediatrics drug working group member of the Pharmaceuticals and Medical Devices Agency (PMDA), on the 2017 Drug Information Association Annual Meeting. Although World Health Organization (WHO) had warned of the same concern nearly 11 years ago.
In China, according to the National Center for Adverse Drug Reaction (ADR) Monitoring, there are about eight pharmaceutical companies producing children medication; 0.1% of the entire market. Every year, inappropriate medicine dosage is causing 30,000 Chinese children to go deaf and 7000 death, these figures are two to four times higher than adults.
Technology is available but not targeted at children
Artificial intelligence has been widely employed to reposition existing drugs, as a way to omit Phase I trial and accelerate the entire medicine development process. Nevertheless, AI has not been considered when it comes to relabeling of adult drugs and administering off-label drugs or giving a reduced dosage of adult medication to children, remain common.
Recently, a group of Swiss researchers had successfully make use of an advanced simulation platform, to duly controlled epidural electrical stimulation (EES), so that patients can regain part of their walking capability. In terms of pediatrics drug development, will it be viable to recruit young, healthy volunteers to upload their body details onto a similar platform, so that researchers can better observe drug pharmacokinetics (i.e., absorption, metabolism and elimination) in children, without having them to actually ingest the drug?
Lastly, machine learning had found to predict side effects of various drug combinations, so will the same technique be able to predict how the five pediatrics sub-population will probably react to the drug and effective ways for the drug to reach the targeted sites, as an alternative to extrapolation from well control clinical trials involving adults?
Beyond Artificial Intelligence in pediatrics
Unfortunately, the above is still a fantasy at the moment and there remains an invisible wall between technology and pediatrics drug development. AI runs on data but gathering sufficient data is equally difficult, if not, ethically challenging as to involving healthy children in clinical trials.
Making use of past data coming from published studies and clinical trials may be a solution, but the formed picture may not give us a full understanding of a dose-to-drug relationship.
Children are not “little adults” but a vulnerable, heterogeneous group, this mere fact not only limits the development of one whole industry but makes us wonder if we are near to the margin of technology.