Artificial intelligence (AI) and its use as a means of managing complex systems of information has found a home in most aspects of healthcare. AI and Machine Learning are implemented in devices and systems in almost all sectors of a hospital, ranging from AI-assisted surgical and diagnostic tools to AI-assisted methods of managing staff.
Some healthcare facilities have followed this path to its logical conclusion and made a concerted effort to link all of these single-point-of-use systems into a cohesive singular AI system, linked through a central hub, with the goal of creating an overarching AI system designed to run every facet of a hospital from patient and staff scheduling, right through to how laundry is handled.
Go deeper with GlobalData
GlobalData’s Medical Device Intelligence Centre forecasts that one of the major drivers for the adoption of AI in hospital settings will be the increasingly common use of AI and machine learning tools embedded into medical devices such as diagnostic devices. These results may then be fed into other AI-driven tools such as clinical decision-making tools. Staff operating the diagnostic devices could increasingly have their work rota’s managed and assigned by an AI rostering program. According to GlobalData analysis, the overall AI market saw sales of $93bn in 2023, up 12% from 2022.
It is no wonder then that a market would emerge designed to identify places in healthcare systems where AI tools could be used, then linking the output of all the AI-powered systems into one much more easily understood and managed singular hub.
One such example of a hospital making concerted efforts to create an AI hub out of its many disparate AI tools is Israel’s Sheba Medical Centre, a self-described “hospital city”, which has brought in AI advisory firm ARC Innovation in a bid to connect up the myriad AI systems working throughout the complex. Hospital Management sat down with Dr. Gal Goshen, the new Head of ARC Innovation’s Sagol AI & Big Data Hub, to find out more. This interview has been edited for length and clarity
Joshua Silverwood: Tell me about your role at Sheba Medical Centre
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Thank you!
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form
By GlobalDataGal Goshen: I think the Sheba Medical Centre brought me onboard because they wanted to transition from an organisation that has a few single-point-of-use AI systems here and there and make it so that there is an AI hub that works as a sort of production centre for AI solutions in-house. We also oversee and consult on AI solutions coming from outside of the hospital as well.
Our other mission statement is that by 2030 the hospital will be the leading AI hospital in the world, and my job is to explain to staff what that will mean and to make it happen. Right now we do a lot of work looking at more systematic approaches to the hospital organisation wide. Employing processes to decide things such as what kind of AI we will look to develop in-house and what we will look to source from outside. What kind of knowledge or manpower to build this system from something that is a series of different singular points, into an organisation-wide system?
JS: How do you go about linking together all these systems?
GG: We target a few specific areas that we basically map. To give you an example through the journey of your standard oncology patient – we map all the different interventions and operations and interactions with the hospital, and we locate the most impactful areas where AI can be implemented.
So, it can be something like using AI to schedule this person’s appointments or to schedule follow-up. We can also do things like produce personalized medicine by matching his genome sequence to the correct treatment. So, we map patients throughout their journey and plan out where we can use AI and that mapping, we are doing across all areas of the hospital.
Sheba Medical Centre is very big, it’s more like a city. It is six hospitals linked together and for each of these hospitals, covering things such as women’s health or cardiovascular conditions, we map the whole organisation and all its processes and find areas where we think AI will be the most impactful for the patient journey. We effectively have an internal scoring system for each AI we are considering integrating.
As part of my homework for this, I interviewed the CEOs of many medical AI companies and asked them what we needed to do at the Sheba Medical Centre to make the integration of AI into the hospital easier. I also asked them what some of the challenges they faced. Surprisingly it’s not about the technology. They found that most hospitals already have the infrastructure in place. What we don’t know how to deal with is the decision-making process and appealing to the correct decision makers that have the relevant position within the organisation to authorize these things. So, a big part of our mission is to educate decision and policymakers and department heads.
JS: Have you encountered much pushback from staff at the hospital when trying to integrate AI tech?
GG: You will always find that there are some staff who are more conservative or are just dealing with things in a day-to-day way, which can be very exhausting. If you talk to an ER doctor AI is really the last thing on their mind. Even though they might resist at first, this can be helped by having a discussion with them about their unmet needs and what efficiencies you could add to what department to make people’s lives just a bit easier. Then you aim to develop or source AI specifically for that need instead of coming up with something that they don’t think they will need. So having those kinds of discussions are really the solution to solving pushback on these issues.
JS: Can you give me an idea of how AI systems can directly help hospital staff?
GG: To give another example, the most sophisticated, automated part of the hospital is not the department you think it would be – it is the laundry department. The Medical Centre, in fact, has the biggest laundromat in all of Israel. It is all automated, entirely hands-free and run by robots using an AI model. So, we need to make the hospital work a lot more like the laundry.
It’s funny, but it is because they have fewer challenges integrating these systems into laundry than they do into anoperating room. .
But we have many projects targeting many areas of the hospital, one system we have integrated just recently is a system that works to manage the shifts of healthcare staff in the case of an emergency, but at the same time in those situations, you also need to provide solutions to the children of these workers. Sheba is large enough to have a kindergarten here, but you also need a management system for that. So, we have a whole HR system designed to allocate specific manpower to support the kids and teachers and understand the ages of the children currently in kindergarten and what they would need. This was something we established through Covid-19, but now when there is an emergency a lot of staff-side planning is done through automated systems that we built.
JS: Lastly, you said that you spoke with many CEOs of AI companies, what insights have you gathered?
GG: What I found is that most would agree that technological capacity within hospitals at the moment is not the main barrier as some might believe, most hospitals these days are digitised and even if not, that is not the biggest hurdle to overcome.
The real problem is getting through to the decision-makers and ensuring that they allocate the correct amount of manpower to get AI systems integrated into the hospital. It’s about AI literacy, they need to understand what AI means, how to use it and what it can do. Usually, the friction comes from either decision-makers or decision-making processes that are just not good enough.
