How a Norwegian hospital unlocked capacity without adding staff
Across Europe's healthcare systems, pressure has been building for years. Waiting lists grow, and patients wait, sometimes months, for procedures that feel within reach. But the challenge is not simple. Running an operating department means coordinating an extraordinary number of variables at once: staff availability, equipment, room schedules, surgical teams, anaesthesia, downstream beds, and more.
Within that complexity, bottlenecks emerge, a single constraint that quietly limits everything else. Managing something this complex, with interdependent moving parts and shifting bottlenecks, is beyond what traditional planning tools were built for.
At Lovisenberg Diaconal Hospital in Norway, however, something shifted. By introducing AI-driven surgery planning, the hospital achieved a 21% increase in completed surgeries and a significant reduction in overtime. But the story behind those numbers raises questions that go well beyond any single technology or institution. The most important of them: what does waiting actually cost?
The hidden cost of waiting times
Long surgical waiting times are rarely discussed in terms of their true cost. The headline figure, weeks or months on a waiting list, understates the cumulative harm. For a patient awaiting shoulder surgery, every passing month may mean more pain, more sick leave, and a worsening condition that will ultimately require more intensive treatment. Multiply that across thousands of patients, and the consequences become significant at both human and societal levels.
Research consistently shows that delayed treatment leads to increased mental stress, physical deterioration, and longer recovery times.
There are also downstream economic effects: reduced workforce participation, greater dependence on social welfare systems, and higher long-term care costs.
In many cases, the capacity is already there, however, buried beneath layers of complexity. The constraint is not surgical skill or physical infrastructure; it is the coordination of everything required for a procedure to happen on time.
From manual burden to measurable impact
Operating rooms are among the most resource-intensive environments in any hospital. For a procedure to proceed, the right surgeon, support staff, equipment, room, and post-operative recovery bed must all be available simultaneously. Patient acuity, clinical priority, staff training requirements, and institutional targets all add further variables. A single unexpected absence or overrun creates a cascade that experienced schedulers know all too well.
This challenge is not new. Hospitals have managed complexity through extensive manual planning processes: spreadsheets, institutional knowledge, and the tacit expertise of experienced planners who carry the scheduling logic largely in their heads. For a long time, it has been enough.
But the scale has shifted. Rising patient volumes, growing complexity, and increasing pressure on staffing have pushed these systems beyond what they were built to handle. When a planner leaves, the knowledge walks out with them, and when volumes increase, the cracks start to show.
Better tools and processes are needed. Tools that can handle the combinatorial complexity of surgical scheduling in a way that no individual planner can practically manage alone. Menon Economics estimates the financial gain from improved operating room scheduling at approximately £170,000 per room in additional annual revenue — reflecting untapped capacity that already exists within many hospitals.
What AI can, and cannot do
This is where artificial intelligence (AI) enters the picture, though it is worth being precise about what that means in this context.
There is a meaningful distinction between tasks that can be automated and tasks that require human judgement. In healthcare, that distinction matters enormously. The ability to assess a situation as a whole, weigh considerations that cannot be fully quantified, and take responsibility for the consequences is not something that can or should be outsourced to an algorithm. Prioritisation under pressure, handling clinical uncertainty, and the quality of interaction between patient and clinician all require human presence and responsibility.
What AI can do is handle the routine complexity that currently consumes enormous amounts of time and cognitive capacity; structuring fragmented information, predicting how waiting lists will develop, and optimising schedules against multiple constraints simultaneously. These are tasks where computational approaches can significantly outperform manual processes, and where automation frees up the human attention that should be directed at judgement-intensive decisions.
The goal is not to remove people from the planning process, but to give them better information and more time to focus on the decisions that genuinely require their expertise.
Co-development with healthcare professionals
The experience at Lovisenberg Diaconal Hospital illustrates both the potential and the conditions required for this approach to work in practice.
When the hospital began exploring AI-assisted scheduling together with a Norwegian health tech company, the surgical planning team was managing an extraordinarily complex puzzle largely by hand. Information was fragmented across departments and systems, making it difficult to build a reliable shared picture of what any given week's surgical programme required.
Rather than deploying a ready-made solution, the development process involved schedulers, planners, staffing coordinators, and clinical staff. This resulted in a final system that was truly designed around their actual working day, and in the implementation of a planning process grounded in international research.
"Instead of forcing clinicians to adapt to the system, the tool has been built around the real work situation of healthcare personnel," said Henrik Hofgaard, former assistant clinic manager there.
This co-design approach reflects a broader principle: technology in healthcare designed without deep clinical input tends to add administrative burden, rather than reduce it. For AI tools specifically, adoption depends not just on technical capability, but on whether clinicians trust the outputs and find that the system genuinely supports their work.
Putting the data to work
An SaaS solution integrates directly with the hospital's existing electronic health record infrastructure, ensuring that information flows consistently without requiring double data entry. It structures free-text clinical data, predicts waiting list development using machine learning, and generates optimised surgical programmes that account for staff availability, room utilisation, overtime risk, and recovery bed capacity.
For planners, this means receiving automated patient suggestions based on clinical priority and available operating time, with the system handling the combinatorial complexity that would otherwise require hours of manual work.
A broader opportunity
The problem Lovisenberg Diaconal Hospital faced is not unique to Norway. Across Europe, hospitals face structurally similar challenges: growing demand, constrained resources, and planning processes that rely too heavily on individual expertise and institutional memory.
The case for AI-assisted surgical planning is ultimately not a technology argument; it is a patient-access argument. Every week a patient spends on a waiting list carries an unnecessarily human and societal cost. Better utilisation of existing capacity is, in most health systems, far more achievable in the short term than expanding it.
The question for healthcare leaders is less whether AI has a role in surgical planning, and more how to implement it in a way that genuinely supports the people doing the work, and the patients depending on them.
About the author
Line Langeland is a doctor and ENT specialist with over ten years of clinical experience from Norwegian hospitals. Today, she works as medical director at Deepinsight, applying artificial intelligence to improve capacity management and resource utilisation in hospitals.
