Developing medicinal product with AI: New white paper from Plattform Lernende Systeme shows potential and challenges

Artificial Intelligence (AI) can accelerate the development of medicines and promote personalised medicine across the board. In this way, better and individualised medicines can be brought to market more cost-effectively. However, before the population can benefit from more favourable drug prices and innovative medicines, sufficient high-quality data from patients must be made available for research in addition to data on active ingredients, and a legally secure regulatory framework must be created. A current white paper from Plattform Lernende Systeme uses practical examples to show the potential of AI in drug development and addresses design options for overcoming the existing hurdles.

In Germany, an average of one medicine is prescribed for every visit to the doctor. Pharmaceuticals are the third largest item in German healthcare expenditure. However, their development is becoming increasingly expensive and time-consuming. It takes around twelve years for a drug to be launched on the market at an average total cost of around 2.8 billion US dollars. The main reasons for this are the increasingly complex products and study designs, the increasing requirements for documentation and safety during development and the costly recruitment of participants for clinical trials. In many cases, such as with antibiotics, the development of new active ingredients is no longer profitable - to the detriment of healthcare.

Personalised cancer therapies and innovative active ingredients

The process of drug development, from the initial idea to authorisation, can be made more efficient with the use of AI and offers the opportunity to save years of work and costly investments, according to the white paper "Developing medicinal product with AI". With the help of AI, huge amounts of data can be systematically analysed and extensive knowledge can be evaluated quickly. In this way, suitable drug targets and candidates can be found in a short time, better predictions can be made about the side effects of the drugs and the chemical synthesis, i.e. the production of the active ingredient, can be optimised. AI can also support the selection and monitoring of test subjects for clinical trials and authorisation. AI-based data analysis also enables the development of personalised therapies, for example for the treatment of cancer, which are tailored to the individual clinical picture of the person affected.

The authors of the white paper also identify the challenges on the path to AI-based drug research. For example, large amounts of high-quality data on active ingredients must be available for the use of AI. However, this requires the willingness of research companies to share data. Gaps exist in particular in the database on human biology, for example on disease mechanisms and the effect of drugs. This could be closed with high-quality data from the population that is made available via electronic patient records or health insurance companies. The AI analysis of patient data allows statements to be made about the effectiveness and side effects of drugs as well as personalised treatment recommendations.

There are currently plans at both European and national level to increase the availability of data, for example in the form of the draft regulation on the European Health Data Space (EHDS) and the Health Data Utilisation Act (GDNG) in Germany. The aim is to make health data accessible to industry for research purposes. The experts from Plattform Lernende Systeme recommend that the increased availability of data should not be torpedoed by regulatory restrictions on AI-supported research.

Authorisation needs binding standards and transparency

The use of AI in the development of active ingredients must also be taken into account in the authorisation and reimbursement of medicinal products. The results of AI data analyses must be comprehensible and AI-based statements on medical aspects must be clearly verifiable. Regulatory processes must be adapted accordingly. AI not only utilises data from traditional studies, but also synthetically generated data. Binding standards are required for the review and validity of AI-based data as part of the authorisation process.

The use of AI methods in drug development also means that more authorisation applications for new drugs are submitted to the authorities in less time. However, the white paper also states that AI can also help regulatory authorities to speed up processes in order to keep up with the increased pace of development.

About the white paper

The white paper "Developing medicinal product with AI: From an idea to clinical licensing. Applications, potential and challenges" was written by members of the Health Care, Medical Technology, Care working group of Plattform Lernende Systeme. It is available to download (in German) free of charge.