Machine learning speeds up digital transformation at leading Saudi Arabia hospital

Machine learning and artificial intelligence (ML and AI) have been at the heart of the King Faisal Specialist Hospital & Research Centre’s (KFSH&RC) response to the COVID-19 pandemic in Saudi Arabia, accelerating a digital transformation journey that has been underway since the start of the 21st century.

The rapid development of a highly integrated COVID-19 digital support machine learning platform, using predictive analytics to optimise the hospital’s operational response and patient care delivery, has been a game-changing experience for the organisation – and in particular, for the Healthcare Information Technology (HIT) team led by CIO Dr Osama Alswailem.

“The hospital as an organisation is transitioning from ‘smart’ to ‘intelligent’ systems,” says Alswailem. “Before the pandemic, we were already moving from interoperability, data warehousing and simple analytics into more machine learning projects from genomics to 3D printing. When COVID happened, we shifted our focus from the defined use cases that we had into a platform that could use real-time, multi-dimensional data to enable focused organisational decisions.”

Adapting to the uncertainties of a rapidly developing pandemic demanded a platform that could be integrated with every internal and external operational and clinical function, including the hospital’s Integrated Clinical Information System (ICIS), bringing real-time data to care providers and administrators so that decisions could be made and resources such as beds and devices allocated based on the latest knowledge.

“We took all of the supply chain and integrated it into the COVID platform, from medication, Personal Protective Equipment (PPE) to ventilators – we even brought our research ventilators into the pipeline in case they were needed – to rapid tests and 3D printing for face shields,” says Alswailem. “This allowed the supply chain to make the most insightful decisions based on demand, quantity and price across both facilities in Jeddah and Riyadh.”

Tackling pain areas 

The platform has delivered a predictive model accuracy of 92% – a significant contribution to some important outcomes: zero sentinel, stock medication, PPE, bed, ventilator and staff shortage incidents.“We incorporated staff workflows so that we could see who was trained and could be transferred from their respective wards to ICU, who was under quarantine or isolating, and created a self-learning contact tracing solution. Since then, we have added vaccination data so we know how many staff are not vaccinated and we have targeted those groups via social media campaigns to the extent that we have now achieved a 99% vaccination rate.”

Other initiatives include a scenario-based capacity tool that measures the Reff rate of COVID-19; which predicts how many people will be infected due to each positive case tied with government restrictions, curfews, etc. “We are very proud of all the work that went into building this platform, which has led to innovation in many other areas of the organisation” says Alswailem.

“We have been able to do this due to KFSH&RC leadership support and our unique approach to HIT, which is led by the business rather than being a back office function. Our HIT team is made-up of physicians, nurses, pharmacists, lab supervisors – we all speak the same language as the healthcare business. Since 2019, we’ve had a clear plan about how to advance our use of AI and ML, focusing mainly on patient experience, medical imaging, applied personalised medicine, clinical diagnostics and prognostics.”

Alswailem says that by tackling ‘pain areas’ rather than simply using the out-of-the-box technology, KFSH&RC has been able to attain significant achievements: building Arabised sentiment analysis around the public use of social media into its prediction models; for appointment no-shows and targeted reminders; genetic disorder and cancer prediction; and micro-genomics feeding into patient information to help with diagnosis and treatment.

This also connects with Saudi Vision 2030, in which healthcare is a major focus, and the hospital is leading the way by sharing the knowledge and experience it is gaining throughout the digital transformation process, nationally and regionally.

Virtual upgrade of the ICIS

KFSH&RC’s 30-year relationship with Cerner – the latest stage of which was a virtual upgrade of the ICIS during the summer of 2020 – has been an essential part of the hospital’s digital transformation. “Much more than a piece of software,” says Alswailem, “the partnership has created a great wealth of knowledge, expertise and solutions and has been strengthened by the ongoing cycle of change management and technology adoption.”

“It’s the interoperability that puts the workflow together, the constant data generation has become one of the biggest assets of the organisation,” he says. “I think we are reaching the stage where people can just ask for research related data, and it takes a few hours compared with few months as it used to be.”

Alswailem says there are three crucial areas in which the system has enhanced healthcare delivery across the hospital: interoperability, patient engagement and the advent of a new model of patient care.

“EMR interoperability has disrupted the way healthcare is conducted,” he says. “Having the patient information without the layers of delays, planning the length of stay and improving care co-ordination among providers. Too many people used to be involved in taking care of a patient and through organisational interoperability we have seen a major difference.

“Having the patient as a positive and interactive member of the healthcare delivery process rather than just a passive receiver would never have been possible without the patient hub that sits on top of the EMR – much more than a portal. It was very evident during COVID that the patient felt involved, knew that care was available and they could reach out to their provider, and we could see the results and impact of their treatment on their quality of life.”

Patient Experience AI Model goes live 

The Patient Experience AI Model went live in March 2021 and has achieved a predictive model accuracy of 87%, enabling 693 total distinct patients to be reviewed and 3,840 actions to be taken to enhance patient experience.

“A new model of patient care would never have been possible without this technology – whether it’s in the genome research and having access to petabytes of data or just providing a focused message for the physician at the right time. But it is not just a piece of information to help the physician talk to the patient, it’s a solution that continues from the previous session and looks to the next one, knowing what goes to the lab, what goes to the pharmacy and everything throughout the patient journey.”

Alswailem says we are in a unique era for healthcare, in which technology of every kind enables care models that bring value to the patient. As a physician, he suggests, nothing has nudged behaviour greater than his smartphone, which constantly monitors his blood sugar. From the IOMT (Internet of Medical Things), virtual/augmented reality to 3D printing, the horizon is constantly widening.

The hospital’s digital transformation journey has been a flagship project, leading the way in Saudi Arabia and across the Middle East, and earning KFSH&RC numerous awards for the innovation taking place under Alswailem’s influence. But he is quick to point out that this has not been a one-man show.

“This recognition is the result of a journey that the hospital has been on for the last three decades,” he says. “During that time the focus has shifted from day-to-day operation to the transformation of healthcare delivery. It meant that our reaction to the pandemic would be strategic and calculated. And now we have a duty to share our information and technology in ways that will improve healthcare for the whole population.”

Source: Read Full Article