Contributing writer

How AI and big data are changing healthcare in the Middle East

Dec 20, 2021
Artificial IntelligenceBig DataHealthcare Industry

AI and big data analytics are allowing healthcare providers in the Middle East to make faster, more cost-effective diagnostics, but security concerns are top of mind.

healthcare 3.1
Credit: Dell Technologies

AI and big data analytics are allowing healthcare providers in the Middle East to make faster, more cost-effective diagnostics, according to a broad cross-section of healthcare professionals. Along with the increasing use of AI and big data, though, security concerns about data privacy are also growing.

AI is one of the fastest growing segments of the global healthcare market today. According to Frost & Sullivan forecasts, it will reach US$6.6 billion by the end of this year. Such growth rates are possible thanks to the huge amounts of data generated by a wide variety of devices, which can be analysed and acted on.

In the Middle East, healthcare professionals attest to the effectiveness and increasing usage of AI.

 “We started the digital transformation journey last year with the focus on technology being the core to the foundation of the American Hospital. We’ve teamed up with some big names like Oracle and Microsoft to build a new level of intelligence when it comes to models,” said Ahmad Yahya, CIO of American Hospital Dubai.

The hospital’s IT team created a new COVID diagnostics application built on its clinical database and modelling from Cerner. It was customised and validated by the clinical staff at the hospital intensive care unit, and helped identify risk factors for patients, as well as in deciding who would go to the ICU.

Two other AI-based diagnostic applications are aimed at identifying asthma patients and predicting whether emergency patients will go to an in-patient ward.

“We are currently working on all these AI models with one of them, the COVID one, having been already validated, while the other two are close to being validated. We are also working on going live early next year with real-time monitoring of patients’ sentiments, satisfaction, and (hospital) capacity,” which can serve to help allocate resources, Yahyah disclosed earlier this year, at the Arab Health 2021 event in Dubai.

With a wealth of historical medical records available for analysis, AI can be helpful in making a diagnosis and choosing an appropriate treatment, providing the doctor with a “third opinion”, healthcare professionals say. AI applications are able to analyse all available medical information about a specific disease, and find out which treatments and drugs have been the most effective in the entire history of medical practice.

Big data fuels growth of AI applications

Medicine is a data-rich field in which accuracy is perhaps the most critical factor. The more data that algorithms process, the more accurately and correctly they will be able to formulate conclusions based on them. Meanwhile, different types of technology in use today are generating an increasing volume of health data, according to Massimo Cannizzo, CEO of Gellify, a venture capital company that in October launched a US$50 million fund with management group Azimut to invest in companies providing healthcare and emerging technology in the Middle East.

There are, for example, various wearable devices that are gaining popularity and generating health information, including portable heart rate and blood pressure monitors — devices that can continuously monitor your heart rate or blood sugar.

As their cost decreases and the functionality of the already popular fitness bracelets expands, AI-based diagnostic systems will receive even more data on the health of each individual patient, giving the doctor the opportunity to more accurately and efficiently prescribe a treatment plan.

The rise of so-called augmented healthcare is demonstrated by the growth figures of wearables market in the MENA (Middle East and North Africa) region, where 25% of the adult population is expected to be using a wearable device by 2022, according to Cannizzo.

AI for disease, workflow optimisation

Algorithms and AI models — programmes or sets of algorithms that use a set of data to recognise patterns and perform tasks — are constantly improving, and this progress is already finding expression in specific applications in the medical field.

AI will not only help make clinical assessments, but will also streamline operations and workflow, according to Kentaro Suzuki, general manager at medical equipment maker Canon Medical. For example, it can shorten the scan time in MRI, Suzuki said.

“We believe AI is a tool for the medical profession in order to improve the current capability of the imaging equipment,” Suzuki said.

AI applications can analyse medical images and find early signs of a disease that a doctor may not notice. “It will never replace humans or directly diagnose patients,” Suzuki said. The “third opinion” that AI can offer, though, is especially relevant for oncological diseases, in which an early start of treatment can significantly improve the prognosis for a speedy recovery.

The role of AI in genomics

AI can also change the way we think about genetics, and push the boundaries of genomics research. Recent research shows that an artificial neural network is capable of identifying and detecting patterns in large amounts of genetic data, thus revealing groups and sequences of genes associated with specific diseases.

Since diseases are “encoded” in the genetic sequence of a person, the ability to understand genetic information at the most detailed level is currently the key to determining how to treat them.

Advances in genomics have been difficult to make due to the complexity of genetic data. Thanks to the ability of AI to classify and analyse a wide range of data in a short period of time, though, practitioners now expect breakthroughs in genomic study.

 “I think we are all shifting towards more automated work flows and smart technologies, including in genomics by default because we generate large amount of data that is impossible to manually assess and to make sense of,” said Prof. Walid Mohammad Abuhammour, clinical molecular geneticist, director of the genomics centre at Al Jalila Children’s Specialty Hospital, and associate professor of genetics at Mohammed bin Rashid University.

AI can, in particular, make headway in helping to cure rare diseases. Some 80% of the so-called rare diseases are genomic, with 50% of them found in children, noted Joshua Symons, director of data strategy at Genomics England.

“We want to embed the combination of AI and genomics into routine medical care to approve outcomes and therapies for patients to discover new drugs and improve lives on a national scale,” Symons said.

These types of approaches, embedded into routine clinical medicine, will enable decision support systems to determine what the best therapies are for patients, who, for example, now can perform diabetes blood tests at home, Symons said. Technology is evolving that will allow cancer patients to take their own blood sample and have the ctDNA(circulating tumor DNA) analysed to determine whether they are responding to treatment, he said.

AI security concerns on the rise

As an increasing number of companies apply emerging technologies such as AI and robotics to healthcare, however, concerns about security and patient privacy are on the rise.

“We have a lot of companies in digital health and many of them use AI. We also have companies that use surgical robots or manufacture robotics. The number of those companies is increasing,” said Marwan Janahi, managing director of Dubai Science Park (DSP), created to position the UAE as a major destination for research and development.

There are more than 400 companies at DSP, with over 4,000 people working there and the healthcare sector is constantly evolving, Janahi said. The role of AI in healthcare is increasingly important, but at the same time healthcare professionals should be very careful about the information concerning patients and how their privacy is secured, he said.

There are a variety of ways to approach security for patient data. For instance, the UAE ICT law that was issued in January 2019 requires patient data to be stored in the country, where there are very strict data protection laws, Janahi said.

The data should be owned by the patient, but the same time there is a need for flexibility because sometimes there is a need to share information with other parts of the world to advance healthcare knowledge, and to get second opinions, Janahi pointed out.

Healthcare officials should take a variety of approaches toward data security, said American Hospital’s Yahyah. “We use a typical approach when it comes to cybersecurity, like having firewalls, etc. But the weakest link is your people and what is really important is [security] awareness, which is critical to us,” he said.

As a preventive measure the hospital does some mock demo attacks to raise awareness among its people within the enterprise, he said.

More AI training is needed

While healthcare has evolved to be one of humanity’s great success stories due to advances in medicine and technology, at the same time it is in a crisis because of challenges it is facing, said Khalid Ghaloua Adine, director of solutions marketing for digital healthcare at Etisalat Digital.

While life expectancy is higher today than ever before, there is a huge shortage of healthcare professionals to cover the demand. There is also a concern on how to manage and control costs. That’s the reason why healthcare professionals seek new technologies that can optimise costs and support human caregivers, he said.

To be prepared for the challenges ahead, there is a need for more data scientists, data governance engineers, and other types of roles. In addition to CIOs, “why not a chief AI officer?” Adine asked.

Healthcare organisations should look ahead at skill sets they need to have in the future, he suggested.

 “For the last 20 years we have been generating data that will need a huge number of people to process it. If we don’t have people with the right set of skills to interpret and model that information then we will not be able to overcome those challenges,” Adine said.