Shot in the arm: AI’s healthcare journey
Hospitals generate massive amounts of data from the hospital information system (HIS), picture archiving and communication system (PACS) and laboratory information system (LIS). Analysing this data can have a major impact on health outcomes. The technology to securely and efficiently analyse this massive trove of data to gain any meaningful insights has been wanting, until now. AI, which by nature requires the ingestion of enormous amounts of data, was built for these types of environments.
Aster DM Healthcare is a medical group with a presence spread across nine countries with 12 hospitals, 107 clinics and 231 pharmacies with heavy patient footfall. Using artificial intelligence, Aster is now leveraging its massive pool of data through the recently-launched Centre of Digital Health Excellence (Aster CDHE).
Aster CDHE has four platforms. Aster Databank is a platform to leverage health data to build AI algorithms. It is currently working with a leading digital data transformation partner towards collecting, synthesising, annotating and storing data from various streams such as HIS, PACS, LIS and Citizen.
The second platform is Asterstine, an inhouse AI and big data research centre which seeks to partner with tech firms to codevelop and conduct interdisciplinary research in Aster’s medical facilities. The third platform is Aster Clinical Simulation Lab, a pre-clinical product development approach for faster, and compliant evidence generation for digital health solutions. Aster’s approach helps in evidence-generation, iterative agile development, clinical data modelling and conducting early validation experiments inside the group’s infrastructure.
Aster has also launched the Digital Centralised Clinical Trial Centre, a facility that is using digital technologies to shift conventional clinical trial processes. The centre also offers a platform to SMEs and startups with limited resources to undertake clinical studies in digital technologies while enabling larger corporates to avoid operational inefficiencies.
“There’s a lot of innovation happening in healthcare using AI. We have been actively involved in this area – during the Covid-19 pandemic when there was a shortage of intensivists, we connected our ICU with specialists sitting in a single location to help remote monitoring,” says Dr Azad Moopen, founder, chairman and managing director, Aster DM Healthcare.
“Aster’s innovation centre aims to achieve key milestones such as the introduction of innovative solutions for home healthcare with a focus on digital primary care; establishing an ecosystem of digital health partners from startups to academia; and starting digital health, informatics and medicine as a stream for future healthcare workers,” he adds.
AI offers healthcare providers the opportunity to leverage data analytics to better manage workflows, patient flow, personnel flow and dynamically adapt capacity to demand in proactive ways. A good example of this is command centre technology. “Hospitals command centres run like airport control towers that manage – in real-time and proactively – the operational data of a department or a hospital, to better match demand and capacity to improve efficiency,” explains Catherine Estrampes, president and CEO of GE Healthcare Europe/
Russia, Middle East and Africa (EMEA).
GE Healthcare has installed about 200 command centres in hospitals around the world in the past year. Regionally, Estrampes says GE is working with clients in Saudi Arabia and the UAE on command centre technology which will be deployed soon.
Healthcare data is one of the most guarded the world over, for obvious reasons.
Dr Satish Rath, group chief officer for innovation and research, Aster DM Healthcare, says records handled by the group are anonymised so that no patient privacy information is released. “Secure federated learning is a shining light in this direction. In this new architecture, patient data needn’t be shared, only data models are shared. This will revolutionise AI research speed in coming years while protecting patient privacy and laws of the land,” he explains.
Estrampes says GE Healthcare follows a very strict charter around data protection and data privacy. “We are very careful in making sure we handle data in full compliance with each country’s laws.”
GE Healthcare is also one of the first healthcare companies to publish and address ethics around AI in healthcare, creating transparency around what the company is working on, the use of the data and guarding against AI bias, says Estrampes.
Early AI use cases are emerging, with promising results.
The GE Healthcare ‘critical care suite’ is an on-device AI for fixed and mobile digital X-ray systems which notifies radiologists of cases with potential critical findings. It uses AI to detect pneumothorax, or collapsed lung, which is a lethal condition if it doesn’t get treated quickly. It not only notifies staff, but it also sends the information to the PACS so that it gets prioritised reading.
Another example from GE Healthcare is the ‘thoracic care suite’ that uses AI to detect abnormal chest radiograph findings, which supports tuberculosis detection and helps find pneumonia or ground glass opacities, which are indicative of Covid-19.
GE Healthcare has also developed a deep learning image reconstruction engine on the Revolution Apex CT machine, which is an AI-powered reconstruction engine that reduces scan time by about 30 per cent while keeping the same image quality. “The care delivery improvement and gains in efficiency, in this case, are remarkable,” says Estrampes.
Asterstine is partnering with a global semiconductor giant to set up a federated learning-based simulation infrastructure for AI-based research programmes. In one of the projects, CDHE is developing and validating an AI-based X-ray diagnostics system to detect more than 30 types of radiological presentations. Aster’s other partnership with a global software solution provider is exploring an audio-based diagnostic system for respiratory, digestive, and cardiovascular diseases.
At the Aster simulation lab, an earlier project leverages the company’s biological research capability in pathological imaging for renal and metabolic diseases.
Dr Rath notes that earlier medical inventions such as stethoscopes and even drugs were invented inside hospitals by physician-scientists. With the introduction of complex machines in the post-industrial era, invention moved away from hospitals to pharmaceutical, medical devices or digital solutions manufacturers, where products are invented outside healthcare centres and introduced to medical workers. This has resulted in late adoption and conversion of products.
“We want to change this. We want to see the ‘concept-tocash’ cycle happening within a healthcare environment. The real healthcare problems doctors face every day should turn to motivation for solutions that are picked up by startups or technology partners,” Dr Rath says.
Covid-19 has led to a redefinition of care delivery strategies, says Estrampes. Healthcare operators are looking to alter the way they deliver care where hospitals focus on acute care while other less-vulnerable patients are treated at outpatient centres or at home. “In such cases, the use of IoT and telehealth will enable remote patient monitoring while keeping patients out of the hospital,” explains Estrampes.
Meanwhile, 15 surgeons from 13 different countries recently undertook 13 mixed reality orthopaedic operations as part of a global 24-hour showcase event. In the UAE, the surgery was carried out at Burjeel Medical City in Abu Dhabi. The surgeons virtually collaborated using Microsoft HoloLens 2, Microsoft’s mixed reality smartglasses.
Equipped with the HoloLens, the surgeons in each country were able to visualise and operate surgery via hologram, share their real-time view of the surgery while benefitting from remote peers’ expertise on different clinical cases, and train their peers remotely.
With growing success and the push towards innovation, the bond between AI and healthcare appears to have a healthy future.