Artificial Intelligence

The healthcare sector’s imperative to transform has gained momentum during the pandemic, when delivering connected, integrated and seamless care to patients, especially those who are elderly or suffering with chronic conditions, led to an urgent need to disrupt and reimagine legacy systems and care delivery pathways. The pandemic accelerated the adoption of technology and has led to healthcare delivery models becoming more agile and resilient. However, several providers still maintain paper-based health records, with diagnostic results being transferred among providers through CDs and via fax machines, and appointment bookings conducted via telephone. In an industry re-shaped by advanced clinical technology and the use of robotics, there’s a growing need to develop, embed and deliver AI based innovations to support patient services and back office functions in care delivery systems. A KPMG survey based on 751 business decision-makers in 2021 found that 53% of respondents believe that the adoption of AI in healthcare surpasses other industries.^[1] Artificial intelligence (AI) in healthcare is the “study and design of intelligent agents” applied in healthcare settings.^[2] AI in healthcare has far reaching opportunities ranging from process automation to potentially saving lives.

AI supports both clinical and non-clinical decision making. For instance, Google’s DeepMind Health and US Department of Veteran Affairs developed a tool that could predict Acute Kidney Injury (AKI) up to 48 hours earlier.^[3] Google also developed an AI-based eye scanning tool to detect diabetic retinopathy and prevent blindness, implemented in a chain of eye hospitals in India.^[4] Google’s AI is helping radiologists make quicker and more comprehensive decisions. For example, Hardin Memorial Health (HMH) partnered with IBM Watson to implement ‘Patient Synopsis’ to deliver structured patient data for radiologists.^[5] Surgical Robotics enhanced with AI (Mazor Robotics’ spinal surgery robot arm) are assisting surgeons to precisely position the implant.^[6]

AI-based revenue cycle management (RCM) solutions are being increasingly adopted. For instance, Waystar’s ‘Hubble’^[7] and BUDDI AI’s ‘Practice.AI’^[8] (both used by providers in the US) assist denial predictions and root cause analysis, denials appeal and other functions. UAE based Klaim.ai^[9] is another AI-based tech innovator that supports operational efficiencies in claims management, settlement and offering short-term cash flow funding to health providers in UAE. 

AI investment in drug design and discovery is growing exponentially (4.5x higher in 2020 compared to 2019).^[10] AI helps in determining what potential drugs are worth assessing and the most effective way of synthesizing them.

For instance, ‘COVID Moonshot’, a crowdsourced initiative supported by PostEra (an AI startup brought together 500 scientists globally) was used to accelerate the development of Covid-19 vaccines.^[11] Moreover, Moderna utilized its AI algorithm to develop an mRNA Covid-19 vaccine with reduced time-to-market and was one of the first FDA approved Covid-19 vaccines for emergency use in 2020.^[12] Similarly, Atomwise partnered with IBM to screen drugs for treating the Ebola outbreak in 2015 using AI^[13].

In the recent past, the role of AI was instrumental in supporting public health initiatives. Contact tracing apps (US, UK, China, UAE etc.) as well as a machine learning-based tool called Capacity Planning and Analysis System (CPAS) (installed to support the scheduling of ICU beds, equipment, and workforce in UK) ^[14] were developed during the pandemic to support contact tracing and track the spread of the virus within communities and cities in real time.

AI is utilized through platforms like Qventus^[15] (used by several university-affiliated US hospitals) which suggests the fastest route for ambulances, tracks hospital waiting time and supports triage. Johns Hopkins Hospital and GE’s AI platform led to 60% improvement inpatient admissions and 21% rise in patient discharges before 12 pm^[16]. Chatbots and virtual nursing assistants such as Sense.ly’s ‘Molly’^[17] are also used for workforce optimization. Further, AI-enabled applications (H20.ai^[18]) are also helping detect and forecast sepsis in critical care pathways and interventions.

AI-powered tools will assist medical professionals in early and accurate diagnosis of viruses, bacterial infections and diseases. For instance, IBM’s Medical Sieve is a long-term exploratory project to develop a ‘cognitive assistant’ to support cardiologists’ and radiologists’ decision-making^iv. Medtronic recently launched an AI pilot program for the identification of high-risk cardiac patients while addressing gender and race differences.^[19] Human Longevity is a venture aimed at developing an algorithm for early detection of cancer and vascular disease through complete genome sequencing.^[20] Abu Dhabi based G42 Healthcare^[21] has invested considerably in its genomic sequencing program to support improved and early diagnosis of markers for cancer and other chronic conditions, and in the future to support longevity programs.

AI will have a huge impact on precision medicine and personalized treatment. Companies like Deep Genomics^[22] (Canada), Dyno Therapeutics^[23] (US) and Microsoft’s Project Hanover^[24] (US) are using AI-based proprietary platforms to identify high potential treatment and therapies (gene therapies) for rare and malignant diseases. This is also an area of focus of G42 Healthcare in the UAE—to support precision medicine for personalized treatment for cancer and other rare conditions and diseases.

Several robot-assisted surgery tools have been developed to support microsurgical procedures. AI-supported robots have been developed by Maastricht University Medical Center in Netherlands and Johns Hopkins (Smart Tissue Autonomous Robot) to suture small blood vessels/ tissues. AI-based surgical risk and quality management tools (e.g. Caresyntax’s qvident) have been developed to analyze the surgeon’s technical capabilities, and the success of these innovations currently in trail could have an impact in improving clinical outcomes particularly in emerging economies that lack the quality and availability of clinical workforce that is needed to serve their patient populations. AI-based clinical therapies and innovations are also expected to support patients with long-term neurological conditions and other disorders that require speech, physical and / or occupational therapy by personalizing, predicting patient response and performance and engaging the patient through gamification in some therapies where the patient can earn rewards with improved outcomes.

AI plays a significant role in predicting future disease outbreaks and pandemics. BlueDot Inc.^[25], Metabiota and the ‘HealthMap’ software, developed by the Boston’s Children Hospital, use a range of Natural Language Processing (NLP) to predict endemic disease across the globe.[26]

AI’s impact on the healthcare sector is wide ranging, and can make diagnosis quicker and more precise, which translates to better patient outcomes. AI adopted in clinical research and drug discovery will lead to faster drug discoveries, revenue optimization and cost reduction for the R&D pipeline.

Although the future of AI in healthcare is promising, we must acknowledge some of the practical and ethical challenges accompanying it. These span from data privacy concerns, a global shortage of AI scientists, limited skills and concerns about clinical workforce training, the need for significant capital investment in AI innovations and change management, and leadership support to address cultural resistance, as well as potential biases in the technology that may lead to inaccurate results.

Despite the concerns, few industries have more opportunity to use AI for the common good than healthcare. While there is growing evidence of its positive impact, from machine learning to support back office functions in hospitals, to disease surveillance mapping, there’s a much greater need to use AI to address patient safety risks, improve the patient experience and access to care (and therefore health equity), and support clinicians in using evidence-based, scientifically proven precision medicine to diagnose and treat critical illnesses.