Christian Hardahl

On the Human Health, Impact and Technology webinar series on April 29th 2025, Christian Hardahl, Advisory Business Solutions Manager, EMEA Healthcare Lead at SAS spoke to host Professor Patricia Maguire about "AI Hospital; How interoperability and precision health will redefine patient care". In case you missed it, here are the top takeaways.

Career Snapshot

Christian holds a master's degree in biomedical engineering and health informatics and began his professional journey over 15 years ago at SAS, where he has remained ever since. His career started in the consulting department, working closely with healthcare clients to implement data and analytics solutions across Nordic hospital systems. Over the years, he has witnessed the full evolution of data maturity in the healthcare sector, gaining firsthand experience with the challenges organizations face in adopting data, Business Intelligence (BI), and Artificial Intelligence (AI)—regardless of their good intentions. After several years in consulting, Christian transitioned into a broader industry advisory role covering EMEA and Asia Pacific. In this position, he has focused on guiding healthcare organizations through health technology transformations, research initiatives, and data-driven innovation projects. Throughout his career, Christian has remained driven by a passion for improving patient outcomes through technology, finding deep fulfilment in seeing his work positively impact patients, healthcare professionals, and support staff alike.

AI Hospital Concept

SAS, a global leader in analytics for nearly 50 years, operates across industries like healthcare, finance, and government, providing data-driven solutions that help organizations uncover insights, optimize operations, and drive innovation. In healthcare and life sciences, SAS partners closely with clients to improve patient outcomes, streamline processes, and support digital transformation through advanced analytics.

One emerging vision is the concept of the "AI Hospital." As Christian explains, this idea involves using data, technology, and automation across the entire healthcare organization to enhance every aspect of patient care—from diagnostics and treatment planning to administrative efficiency and resource management. While not yet a global standard, some European healthcare systems are beginning to embrace more holistic and data-driven models. However, realizing the AI hospital vision will require a strong foundation of secure, interoperable data systems—ensuring that diverse healthcare technologies can work seamlessly together to support this future-forward model of care.

 

What Interoperability Really Means in Healthcare

In simple terms, interoperability in healthcare means that all IT systems within the healthcare landscape should be able to communicate with each other using a shared language and structure. This ensures that data entered into one system can be understood and used across others—for purposes like patient care, administration, and even research, such as the work done at institutions like UCD. Achieving this level of seamless data exchange is a significant challenge, requiring ongoing efforts not only from technology companies but also from regulators and public institutions. It's about creating a common framework so that health data becomes truly accessible, meaningful, and actionable across the entire ecosystem.

A Nordic Model for Data-Driven Health Innovation

Denmark is widely recognized as one of the most digitally advanced healthcare systems in Europe, benefiting from a long-standing tradition of digitalization across public and government institutions. A key enabler is the national use of social security numbers, which allows for seamless data linkage across various systems, from electronic medical records to lab and clinical data warehouses. This infrastructure supports extensive research and innovation, including access to high-quality registries and real-world clinical data. For professionals working in healthcare data and analytics, Denmark—and the broader Nordic region—offers a uniquely rich environment. Over the past 15 years, groundbreaking projects have emerged, including advanced AI models for early cancer detection and a major research initiative analysing data from over 9,000 patients across 240 biomarkers. These kinds of initiatives highlight Denmark's role at the forefront of data-driven healthcare innovation. You can find an in-depth look at the use of AI in the Danish healthcare system in (opens in a new window)Artificial Intelligence in Health - an English version will be available soon.

Revolutionizing Emergency Care with AI and Machine Learning

In a groundbreaking project on machine learning for diagnostic support in medical emergency departments, they worked with an extensive set of data to improve patient care. Starting from scratch, they built a cohort of 9,000 patients from acute departments, collecting blood samples to analyse 240 biomarkers per patient. This data, combined with vital parameters recorded within the first 60 minutes of an acute admission, was then processed in an optimized lab. The lab, once filled with people, is now largely automated, with robots transporting samples and coordinating analysis. Once the biomarkers were analysed, the results were integrated with AI technology, which provided real-time risk assessments for 14 potential diagnoses and 5 outcomes. While working with all 240 biomarkers was cost-prohibitive, the team streamlined the process by reducing the analysis to 60 biomarkers, making it more affordable and feasible for widespread implementation. This innovation is helping clinicians quickly assess patient risk, enabling faster and more accurate decision-making, ultimately improving outcomes for patients in emergency care. The promising results we've already achieved have been published in Nature - (opens in a new window)Machine learning in diagnostic support in medical emergency departments.

A Groundbreaking RCT Study

The team is now preparing to test their AI-driven diagnostic model in a large-scale randomized controlled trial (RCT), which will involve between 16,000 - 18,000 patients. In this trial, every second patient will receive the AI-generated risk score, allowing researchers to compare the outcomes for those using the AI tool versus those who don't. The platform, designed to run in a live production environment, will help assess the effectiveness of the AI in real-time clinical settings. The trial aims to evaluate key metrics, such as the area under the curve for both doctors and AI, providing insights into how well the system can support decision-making. With a focus on reducing mortality rates, particularly in high-risk cases and ICU admissions, the trial hopes to demonstrate significant improvements in patient outcomes, including lower re-admission rates.

Building on this innovation, the AI model is designed to detect 14 major diagnoses early—including serious infections like sepsis—allowing for faster, more accurate intervention. One standout feature is an algorithm that can safely identify patients eligible for discharge within 60 minutes, with strong predictive confidence that they will not be readmitted or face serious complications within 7 to 30 days. This tool acts as a valuable decision-support ally for clinicians, enhancing—but not replacing—their expertise. While the model currently covers 14 diagnoses, it already accounts for nearly 45–50% of all cases in acute departments. However, physicians must be trained to interpret the AI’s output effectively, as many conditions beyond those covered by the model still require clinical judgment and diagnostic oversight.

 

Balancing AI Innovation with Ethical Patient Care

As AI becomes increasingly integrated into clinical care, ethical considerations remain essential—especially when patients are hesitant about machine involvement in their treatment. At SAS, a strong focus is placed on trustworthy AI, ensuring transparency and interpretability of models. In the emergency care project, for instance, physicians are not only given risk scores similar to familiar biomarker results but can also explore the underlying algorithms to see which variables most influenced the outcome for each individual patient. This level of transparency helps build trust in the system and empowers clinicians to override or question AI outputs when they conflict with their own clinical judgment. Importantly, the AI is designed to support, not replace, human expertise. While it may streamline diagnostics and reduce the need for repeated blood draws—delivering comprehensive results within 60 minutes—the decision-making remains firmly in the hands of medical professionals. Ultimately, while such systems offer clear benefits in efficiency and patient care, they must be implemented with careful attention to ethics, patient autonomy, and the clinician’s role in guiding informed consent.

The Future of AI in Healthcare: Technology with a Human Touch

The future of AI in healthcare holds immense promise, but it will likely remain a hybrid model—combining advanced technology with essential human judgment. Around the world, organizations are at different stages of adopting AI, from early exploration to leading-edge implementations. While the idea of a fully autonomous AI hospital is compelling, it may not be desirable—or necessary. The real strength of AI lies in its ability to process vast amounts of data quickly, consistently, and around the clock, making it invaluable in areas where speed, scale, and decision-making reliability are critical. Tools like AI agents will likely become more common, enhancing workflows and supporting clinicians. However, the irreplaceable value of human intuition, empathy, and clinical insight means that AI should remain a tool to augment—not replace—the healthcare professional. The future, then, is not about removing the human element, but about using AI where it truly adds value.