For 39 years, the International Society for Quality in Healthcare (ISQua) has been a central meeting point for world leaders in hospitals including healthcare professionals to inspire and drive improvements in patient safety and health.

Despite these efforts, Eyal Zimlichman, Chief Transformation Officer and Chief Innovation Officer at Sheba Medical Center from Israel, lamented the slow progress regarding patient safety over the last 20 to 30 years, in an interview with Korea Biomedical Review.

“Together with ISQua, we have been working on bringing more technology and patient involvement into the world of patient quality and safety but more innovation also needs to come from inside the walls of hospitals to facilitate the desired changes,” Zimlichman said, explaining the need for organic innovation.

Organic Innovation: Physician-created solutions, more likely to succeed?

The Israeli innovation officer has been focused on redesigning healthcare to be more patient-centric, effective, efficient, and equitable. He founded Sheba Medical Center’s innovation arm called “ARC” which is short for “accelerate, redesign, and collaborate.”

ARC was launched in 2019 to bring together like-minded organizations to transform healthcare with a digital focus. It currently consists of 114 members and 35 healthcare systems in 20 countries around the world. 

“We're working to build ARCs around the world which already exist in the U.S., Canada, UK, Australia, and Bahrain, and will meet with three of the largest hospitals in Seoul including Asan Medical Center (AMC), Seoul National University Hospital (SNUH), and Samsung Medical Center, to find the right partner to join the ARC ecosystem,” revealed Zimlichman. 

“Although we have put in considerable effort to understand and tackle patient safety, patients are still getting harmed at hospitals at alarming rates,” explained Zimlichman, referencing a recent study.

He further pointed out that adverse drug events remain the foremost concern in hospitals, persisting as a decade-old issue that has shown minimal improvement. 

He proceeded to outline additional challenges, including unfavorable surgical outcomes, healthcare-associated infections, and instances of patient falls.

“The system is not built in a way to prevent these human errors because humans will always make errors but we need technology to prevent these errors from coming through,” said Zimlichman.

He gave examples of using artificial intelligence (AI)-based systems to identify prescription errors.  

“We found the AI system is 80 percent more accurate in pinpointing the actual errors compared with traditional systems trying to tackle this problem, as it can learn more about the patient and the drugs being prescribed,” he explained. 

On the other hand, he also revealed the application of AI to prevent diagnosis errors, especially in high-acuity cases in the emergency department.

“We have an AI that can read the CT scan, in milliseconds and diagnose life-threatening problems such as a stroke in the brain, to alert staff to prevent diagnosis delays, which can be fatal to the patient in emergencies,” said Zimlichman. “Our recently published paper demonstrated a 30 percent mortality reduction with this system.” 

Still, he noted that the AI system needs to be properly integrated into workflows to ensure its adoption into the health system.  

“We first wanted to use one of our AI systems as a clinical decision support but the doctors didn't accept it because they thought that the AI would not be able to improve their capabilities so we had to convince physicians that the AI would only prioritize cases and then we saw high adoption rates of the same technology,” recalled the Israeli.

Zimlichman emphasized the need for similar technologies to impact patient outcomes but also drive down the cost of care to ensure sustainable healthcare.

Four key elements to ARC success

In this context, he emphasized four fundamental components of ARC: digital health as the primary catalyst for healthcare transformation. fostering organic and open innovation to facilitate advancements within and beyond the hospital setting, establishing international collaborations, and ultimately adopting a methodical approach to innovation.

To implement this system, he mentioned two key points. 

Foremost, he highlighted the need to share data between different institutions to develop AI solutions applicable to broad audiences. On this note, the ARC at Sheba Medical Center recently announced building a new data platform together with the Mayo Clinic, Toronto General Hospital, and Albert Einstein Hospital in Brazil to validate AI solutions across multiple institutions. 

Secondly, he mentioned the importance of creating evidence for AI solutions similar to pharmaceutical drugs which are conducted in multi-site clinical trials. Additionally, he noted that regulators need to be flexible to allow sharing of data outside of institutions in a safe way so that AI’s capabilities would not be limited. 

Striking the right balance for AI regulations

“Currently, we are struggling with AI regulation because we haven't figured out the regulation in a way that we feel comfortable,” Zimlichman said. “Once we figure out the regulatory steps that are required to approve AI solutions, then it will become much easier to implement across multiple organizations.”

He mentioned the system at ARC is to conduct pilots as a clinical trial through an institutional review board (IRB) and a local ethics committee to create evidence to demonstrate safety. 

“Once there is sufficient evidence of safety and quality, then we go to the local regulatory body to approve the AI-based technology if we want to turn this into a product and introduce it across multiple organizations,” explained the expert.

On the other hand, he noted that if the AI solution developed in-house remains within the hospital, regulation is not required. However, Sheba Medical Center has internal processes to approve AI solutions before implementation. 

“We always approve new AI-based technologies first for either six months or a year and then we re-evaluate the impact before we approve it for the longer run,” said Zimlichman.

In this way, the Israeli re-emphasized the importance of organic innovation so that real-world effectiveness and evidence of safety can first be collected in the hospital where the device was developed, making it easier to go to the regulator later with sufficient evidence to gain approval for the AI-based healthcare innovation.