The evolution of clinical intelligence
Artificial intelligence (AI) is seen as the future of healthcare. But what can it already do for patients today? This PhD research explored how AI can address a broad range of medical challenges, demonstrating how it can help doctors make faster, better-informed decisions while ensuring that human expertise remains central to patient care.
The findings show that AI can improve healthcare in many different ways. It can help hospitals identify where closer collaboration between medical specialties could benefit patients with multiple chronic conditions. In cancer care, AI can reduce the need for additional medical scans, identify tumors on medical images, and predict which patients are likely to experience side effects after radiotherapy. Beyond cancer, AI can track tumors in real time during treatment, estimate lung blood pressure in children without invasive procedures, and provide early warnings of serious infections in newborns.
The research also reveals that successful AI is more than developing increasingly complex algorithms. The quality of the medical data used to train AI models proved to be even more important. Another key challenge is that models developed in one hospital often perform less well in another, highlighting the need for closer collaboration between hospitals and access to more diverse medical datasets.
Ultimately, this thesis shows that AI is not designed to replace healthcare professionals, but to support them. By helping doctors make sense of complex medical data, AI has the potential to make healthcare safer, more effective, and better suited to the needs of patients.