5 steps to put healthcare on the AI fast-track

AI and digital technology will have a greater impact on the entire value chain of the healthcare industry than any previous developments. We are witnessing how this technology is currently being implemented in the healthcare sector and how powerful AI is and will be. The coming decade will be pivotal, for the future of AI in the healthcare sector is being shaped today.

AI can improve clinical outcomes and support medical professionals in decision-making, benefitting patients with more personalized treatment and better health. AI can make the professions in healthcare more attractive by improving job conditions. Furthermore, AI and digital innovations can help boost efficiency in spending across the overall healthcare system, which will be important for societies already burdened by rising healthcare costs. The global healthcare services market is currently valued at around $8 trillion and is continuing to grow.

Right now, the controversy surrounding AI can be (over-)simplified to the opposing positions of “decelerate” versus “full speed ahead”. I would lean towards "accelerate" because there are numerous unmet medical needs – some very challenging – that AI could address. Some basic requirements are essential for AI to be successfully harnessed:

The data clearly show that only around half of all treatments offered to patients are actually effective. The quality of diagnosis is highly variable. Medicine is still often too imprecise and not sufficiently personalized. This is where AI comes into play. Improved diagnosis is one place where AI could benefit patients soonest – for example, in earlier cancer detection in mammograms, better interpretation of complex medical data and reduced medical errors.

Just like medications and medical devices, medical AI can and should be required to present high-quality scientific evidence of safety and efficacy. AI necessitates clear rules. This was also made clear at the AI safety summit in London and in the new EU AI Act. Mistakes in governance such as those made in the development of social networks must not be repeated. Health is a matter of trust at all levels.

Many job profiles will change significantly (in the pharma and medtech industries as well as in hospitals), and change requires leadership. If AI is perceived as competition in the battle for jobs and not as a support in day-to-day work, acceptance will fail to materialize. AI can make the profession more attractive by improving job conditions, taking over repetitive tasks and giving employees more time to work with patients. Accepting AI also includes among patients. They need to be convinced of the benefits of AI and need to be assured that their data will not be misused.

Most data collected at various points along the patient journey is still not utilized; however, initiatives such as the European health data space are about to change this situation. This pool of data is an enormous treasure trove. However, it's not just about data volumes, but above all about data quality. Medical and demographic data is collected by private companies and state institutions alike. In order to increase the availability and applicability of data, it is imperative that we shift from local to cloud-based solutions in the broadest possible manner.

AI must prove its benefits for costs, the environment and patients. Concrete applications with positive examples are essential for the broader rollout of AI. Here are some case studies from within and outside our company:

According to Stanford’s Artificial Intelligence Report 2023, the focus areas that attracted the most private investment in 2022 were medical and healthcare ($6.1 billion). AI is advancing at an unusually fast pace, and the processes in the healthcare industry are not yet adapted to handle this kind of speed. Players in the healthcare space have to concentrate on their strengths – most notably domain knowledge and proprietary data – while being quick to respond to external developments.

Big tech is entering the scene in the healthcare industry. These technology companies have superior resources and computing power at their disposal. Yet it is important not to relinquish control against this backdrop. This can be accomplished by pooling resources and expertise as well as enforcing transparency requirements. In other words, carefully managed open collaboration between different stakeholders could set the direction for the development and adaptation of language models to the specific needs and requirements of the healthcare sector.

The great advantage of AI is that it can be used in everything from studies and basic research to the development of medicines and medical devices, as well as in everyday clinical practice. Utilized correctly, AI will generate an enormous increase in value along the entire healthcare value chain and benefit the whole patient journey.

Everyone involved should have a common interest in narrowing the gap between the technological and scientific possibilities – and the current services to patients in terms of prevention, diagnosis, treatment and aftercare. The AI revolution notwithstanding, one thing must always be clear: The healthcare industry will continue to be a human-to-human endeavour in the future focused on patient care.