AI use cases in healthcare for Covid-19 and beyond

During the Covid-19 crisis, hospitals and healthcare companies have been rushed off their feet in trying to take care of affected patients. Alongside this has been the goal to find effective and safe treatments for the virus, which is still ongoing. However, digital technologies have continued to disrupt the healthcare sector, increasing efficiency and visibility, and AI is a key example.

“Healthcare is a discipline perfectly suited to reap the rewards that even the most basic task-based AI can provide,” said James Norman, chief information officer of healthcare at Dell Technologies. “Globally, the demand for healthcare is increasing at an unprecedented rate – far outstripping the supply of healthcare professionals trained globally.

“While obviously true in the developing world, across Europe an ageing population and a rise in chronic disease is causing unprecedented strain on resources.”


Norman went on to explain how AI has aided pathologists in executing round-the-clock medical results, proving to be useful for treating cancer cases.

“In Europe, the number of cancer cases continues to rise while the number of trained pathologists – those tasked with spotting cancerous cells – declines,” he continued. “Traditional pathology requires that a GP take a tissue sample from a patient, send it to a lab for analysis in a lab, where it’s manually placed on a glass slide to be examined, by a human pathologist, under a microscope. A pathologist, for all the training in the world, gets hungry, gets thirsty, gets tired, requires comfort breaks, and sometimes makes the wrong call.

“Fortunately, this most basic and critical task, that of spotting the cancerous cell, is that which task-based AI is almost perfectly suited to carrying out. Unlike a human, AI never tires and, if the algorithms are correctly coded, acts with incredibly precise results. Artificial intelligence can interrogate multiple libraries of images so that when a clinician detects a tumour, the database can be searched to find all similar tumours – thereby allowing the human pathologist to evaluate the treatment and subsequent outcomes before designing an effective personalised treatment for the patient.

“AI promises to alleviate mind-numbing, tedious repetitive work – in this instance staring down a microscope – and free clinicians to focus on work suited for humans – bespoke, targeted medical treatment.

“The benefits of digital pathology are maximised when this integrated data architecture is combined with high-performance computing, fast-servers, flexible scale-out network storage, and direct, secure access to a multi-cloud environment with big data analytics capabilities. At a time when demand is outstripping supply for the identification and treatment of cancers, artificial intelligence in digital pathology is going to allow patients far more accurate and quicker results that they have ever been able to receive previously.”

Rebuilding your data analytics capabilities in a post-Covid world

Conor McGovern, vice president at Capgemini Invent, discusses how to rebuild your data analytics capabilities in a post-Covid world. Read here

Instant updates

AI has also proven useful in the deployment of mobile healthcare applications, which can deliver real-time data and analysis.

Dr Mahiben Maruthappu, CEO of Cera Care, explained: “Acknowledging the need to move on from dated practices, at Cera, we have developed the UK’s first app-based care provider that incorporates predictive AI technology to keep those being cared for at home, and importantly, out of hospital.

“As an app-based platform, our programming offers a level of accountability that previous practices could never assimilate to. It means that everything is instantly updated, family can check on their loved one and communicate with the carer to make sure everything is as it should be, so there’s no surprises, and all stakeholders are reading from the same page.

“But where the app gets really smart is in using AI-powered predictive analysis to anticipate if a person being cared for is at risk of deteriorating. In older people, the deterioration of health conditions often starts with subtle signs that aren’t easily picked up on with simple note taking or by the naked eye.

“Our centralised digital systems are able to analyse these subtle changes and convert them into a risk assessment, so we can escalate care earlier on. Our office staff have a digital dashboard, continuously updating with new information, and can immediately act on issues as they arise, be that contacting a relative, their GP or calling 111.”

Drug discovery

Another key role that AI plays in healthcare is within drug discovery, an area that has seen numerous collaborative and multi-national projects come to fruition.

Corona Accelerated R&D in Europe (CARE) consortium launched

A new initiative dedicated to accelerating Covid-19 therapy development, the Corona Accelerated R&D in Europe (CARE), has been launched. Read here

“Even before the coronavirus outbreak, TCS was working with AI-based methods to explore chemistry and medical manufacturing,” said Ananth Krishnan, CTO at TCS. “AI methods can learn representations based on existing drugs, allowing scientists to find new drug-like molecules with the potential to cure diseases including coronavirus. Additionally, an AI-based approach can reduce the initial phase of the drug discovery process from several years to a few days thanks, in part, to its ability to optimise several drug characteristics simultaneously very fast.

“The rate at which the coronavirus pandemic has spread has meant that time has been of the essence, making AI particularly useful, especially if you already have the extensive neural network-based generative and predictive models built up as TCS does. Using these models, we discovered 31 molecular compounds that could potentially act as a cure for Covid-19 by targeting one of the well-studied protein targets for coronavirus, ‘chymotrypsin-like (3CL) protease’. This protease is responsible for the virus’ survival and replication in humans; essentially if you can find a way to stop this, you can stop the spread.

“The AI model used to discover these molecules was initially trained on a dataset of 1.6 million drug-like molecules. The model was further trained to incorporate synthetic feasibility. Further tweaking of the model allowed the team to design molecules with optimised physiochemical properties.”

Graph database technology

AI has been effective in increasing data visibility for organisations, and this benefit is no different within the healthcare sector.

Dr Alexander Jarasch, head of data and knowledge management at the German Centre for Diabetes Research (DZD), explained how diabetes research in particular can benefit from graph database technology, combined with AI.

“In order to better understand diseases and combinations of diseases, we try to connect the data that are by definition related,” said Jarasch. “In research into diagnostics around and the therapy of diabetes, we’re always looking for the hidden insights behind the newly connected data.

“To get there, we’re now starting to rely on pattern recognition through a combination of graph technology and machine learning. Graph database technology helps DZD’s researchers connect highly heterogeneous data from various disciplines, species and locations in order to create a hugely valuable body of knowledge. We are doing this by connecting public knowledge with our internal data, enabling our scientists to find hidden connections between data.

“In parallel, applying advanced machine learning techniques to the resulting database has allowed us to get much closer to understanding the complexities of diabetes. As a result, we have moved a step forward in being able to help patients suffering from both diabetes and prediabetes.

Can we automate data quality to support artificial intelligence and machine learning?

Clint Hook, director of Data Governance at Experian, looks at how organisations can automate data quality to support artificial intelligence and machine learning. Read here

“We believe that this combination of graph technology and artificial intelligence means it is possible in the future to succeed in identifying risk groups more precisely. Advanced software or machine learning applications in healthcare will never replace doctors, but a combination of graph technology and machine learning can relieve and support them in both diagnosis and therapy so that they win back more time to look after their patients.”

Digital workers

Lastly, digital workers powered by AI have been found to be useful in maintaining patient records and appointments, freeing up time for healthcare professionals to attend to other tasks.

Blue Prism’s cloud-based intelligent automation platform is providing AI-powered digital workers into the NHS resource pool, to perform a wide range of activities that are being automated at unprecedented speed – across multiple operational functions,” said Peter Walker, CTO EMEA at Blue Prism. “This is helping the NHS overcome a huge range of recent challenges and is releasing more time to care for frontline NHS staff.

How digital workers are creating a work revolution

Digital workers are reworking how organisations are operating, helping them to overcome workload challenges. Read here

University Hospitals of Morecambe Bay are employing digital workers to help patients book, prepare for and follow up appointments – to ensure everyone receives a wealth of tailored communications, confirming each step of their treatment.

“With 600,000 hospital appointments booked a year, there is no way staff could proactively manage that level of personalised communication manually. For medical staff too, they see countless opportunities for removing the daily burden of updating patient record systems so that they can dedicate their time to providing frontline patient care.”

Avatar photo

Aaron Hurst

Aaron Hurst is Information Age's senior reporter, providing news and features around the hottest trends across the tech industry.