The Venetian hotel in Las Vegas feels like walking though Venice, Italy but in a luxury hotel. There’s even a quarter-mile long canal that meanders through shops and restaurants under a blue sky with fluffy white clouds painted on the ceilings. It gives the illusion of being outdoors and even turns dark as evening arrives.
The hotel is built for events, and we recently visited it for an experience covering the latest and greatest in digital health: HIMSS 2025. At the rear is a convention centre – it was the venue for our week of healthcare immersion in an exhibition hall done the Vegas way at a scale I've not seen before. Rows and rows of immense stands. Others tower high in corporate colours. A show favourite is the puppy pen where you get to relax with a pack of rescue puppies.
Then hundreds of speaking sessions to learn about all types of digital health-related content: how hospitals are traversing HIMSS' digital maturity models, how data is being used to identify and support at-risk groups, and how interoperability is being achieved across hospitals to name a few.
The top trend on everyone's lips, the crux of many demonstrations, and theme of a large proportion of talks is artificial intelligence (AI). I am eager to learn more and have picked a two-hour panel discussion with a view to be enlightened about how hospitals are achieving digital maturity by adopting the right tech, at the right scale for the right people.
The Icebreaker question to all those assembled: "In which areas is AI being used within your organisation?" We all scramble for our phones and scan a QR code on screen that grants us access to an app that collates our views. The room is filled with hospital professionals, and a few others from the healthcare supply chain.
The poll results align with what I would expect. Lots leverage AI with business functions: transcribing notes, producing content, translations etc. And then a solid chunk dedicates AI to healthcare functions - radiology, pathology and clinical trials. Healthcare, just like the rest of the world, is leveraging AI at varying levels to make life easier and increase efficiency.
Question two to the group: "Is your organisation committed to strong AI governance, and how are you progressing in this area?" I think of a major blunder by workers who unwittingly leaked top secret source code. "Yes" I transcribe. The responses from the wider group range from "what governance?", to "not at the moment" and "slow" trigger a little nervousness within me. Is there a risk of sensitive data leaking to the outside world, or are results being taken as reliable and accurate?
Missing governance, too much faith in results and uneven usage are highlighted as challenges by the panel chair. The answer to his next question, "what does everyone want from AI?" is captured succinctly by a speaker: "Less admin, less burnout, more joyful staff."
The session moves to a live demo of AI in action. On a large screen, we see a verbal consultation being transcribed directly to an electronic medical record with a useful structure applied.
As the panel discussion begins in earnest, we hear how AI is great at speeding up manual processes, like summarising consultation notes, but it also requires clinicians to think differently. Charting isn't mindless administration; it's a mechanism to organise thoughts. And if things become overly automated, does that reduce the power of observation and experience to spot anomalies?
Digital transformation is already letting us do more with the same which is critical as we seek to maintain standards as the population ages. Many countries are showing strong maturity in digitisation, with electronic medical records achieving major strides, for example. AI is still at the early adoption stage, but there are areas of significant uptake. Radiology is showing strong advances as x-rays, ultrasounds, CT scans, and MRIs can be more effectively analysed by computers as they learn about anomalies greater than the human eye unaided by AI can achieve.
Institutional change is not linear. A speaker mentioned six healthcare trusts they had worked with and their attitudes to AI. The largest took the lead and shared best practice. Two never even thought about it. The others were somewhere in the middle.
Within trusts, those curious enough to try new things often become the champion which adds pressure to the day job, inevitably putting the brakes on innovation. AI remains fragmented with lots of products remaining siloed or pilot based. Organisations struggle with a lack of governance and clear guidelines, plus clinical and administrator buy-in remains lacking.
But not everyone is ready to embrace AI. I had heard nurses talk about unions blocking AI use for fear of losing jobs despite the pressure from staff shortages. There was also a fear that if AI made nurses more efficient, leaders would just expect more and more, with no time given back to patients. The situation is delicate.
Efficiency is always at the forefront of healthcare leaders’ minds. In the UK, integrated care boards (ICBs) are collaborating and merging to reduce costs, and a large healthcare organisation in the region recently reported huge staff reductions.
AI has the power to drive efficiency without the pain of changing business structures and cutting jobs. Administrative tasks could be handed over to AI, chatbots and recorded messages. Here are some examples shared during this panel discussion:
· In one use case, we hear how results are traditionally shared within 28 days by a doctor over the phone. If it is bad news, doctor delivery is important, as there will be many questions and emotions to work through. But if it’s an all-clear, an AI-generated response is now being used to contact patients. Less clinician time, and a faster end to anxiety.
· In other examples shared, AI is being used to predict who won't show up for appointments, based on historical behaviour. Staff can proactively intervene, assess the risks of readmission, and more accurately predict length of stays.
In part two of this series, I’ll share more examples of how clinicians are effectively using AI today as well as how they can overcome data and connectivity challenges. Visit our blog next week for these and other details.
