No longer the realm of science fiction, AI and robotics are transforming healthcare.
Today, AI is used to diagnose diseases, recommend treatments, and even perform surgeries. And hopefully, in the future, AI will play an even more significant role in improving patient outcomes and reducing healthcare costs.
In Singapore, the Healthcare AI industry is still in its nascent stages. However, several startups are already using AI to develop innovative solutions for the healthcare sector.
Demystifying Healthcare AI: Historical Context and the Current State of AI in Healthcare
As healthcare professionals explore the use of AI in healthcare, several questions and concerns have been raised. This section aims to clarify the application of AI in healthcare and its potential implications.
The Origin of AI and its Application in Healthcare
AI has been around for centuries, with its origins tracing back to Greek mythology. In the tale of Pygmalion, a sculptor falls in love with a statue he has created. His wish is granted upon praying to the gods, and the sculpture comes to life.
The term “robot” was first used in the 1920 play RUR (Rossum’s Universal Robots) by Czech writer Karel Capek. The word “robot” comes from the Czech word “robota,” which means forced labour or drudgery.
But it wasn’t until the 1950s that AI started to be developed as a scientific discipline. In 1956, a group of researchers at Dartmouth College convened for a summer workshop to discuss the possibility of creating intelligent machines. This marked the beginning of AI as a formal field of study.
In the early days of AI, the focus was on developing machines that could replicate human intelligence. This was known as the “strong AI” approach.
However, it soon became apparent that creating machines that could think like humans was a much more difficult task than initially thought. As a result, the focus of AI shifted to developing machines that could carry out specific tasks, known as the “weak AI” approach.
That is the approach that is most commonly used today. And it’s why AI applications are often described as “narrow AI” or “applied AI.”
When Was Artificial Intelligence (AI) First Used in Healthcare?
AI found its way into the healthcare industry in the early 1970s, where it was applied to biomedical problems such as cancer detection and drug development. That saw a proliferation of AI research in the medical field.
In 1980, an international AI journal called “Artificial Intelligence in Medicine” was launched. AI would be incorporated into clinical settings in the decades that followed.
In the 1990s, a renewed interest in AI was driven by computer science and machine learning advances. That led to the development of new AI applications in healthcare, such as Fuzzy expert systems, artificial neural networks, Bayesian networks, and decision support systems.
In 2012, Geoffrey Hinton, a computer science professor at the University of Toronto, developed a machine-learning algorithm called “AlexNet.” That introduced the current era of “deep learning,” which has seen a significant increase in the accuracy of AI applications.
The Current State of AI in Healthcare
In recent years, AI has been making inroads into the healthcare industry. Deep neural networks are now performing at par with or exceeding human expertise in various medical tasks.
Some of the most promising applications of AI in healthcare include:
- Diagnosis: AI can be used to develop diagnostic tools that are more accurate than existing methods. For example, an AI system called “IDX-DR” can detect diabetic retinopathy, a leading cause of blindness, with an accuracy of 87%.
- Drug development: AI can be used to accelerate the drug development process. For example, Atomwise’s “AtomNet” platform has been used to design new drugs for Ebola and influenza.
- Treatment: AI can be used to personalise treatment plans for individual patients. For example, IBM’s “Watson for Oncology” system recommends treatment plans for cancer patients.
- Predictions: AI can be used to predict a patient’s health. For example, Google’s “DeepMind Health” unit is developing an AI system that can indicate if a patient is at risk of developing kidney failure.
The State of Health AI in Singapore
Singapore is a leading hub for AI in Asia. The city-state has been investing heavily in AI, with the government committing to invest S$180 million in national research and innovation to tap AI in healthcare and education.
Singapore has done a commendable job leveraging AI data to enhance its healthcare system. An example is the Command, Control, and Command Systems (C3) developed by the Tan Tock Seng Hospital (TTSH) and Integrated Health Information Systems and rolled out in 2019 to facilitate the day-to-day managerial operations of the hospital.
C3 uses AI to predict patient demand and optimise resources accordingly. Essentially it gives a bird’s eye view of thousands of patients visiting the hospital, allowing for better patient flow management.
What are the Main Types of AI and Its Applications in Healthcare?
AI isn’t one technology but rather an umbrella that incorporates various technologies, with new ones being developed all the time. Here are some of the main types of AI and its applications in healthcare:
Machine learning: Neural Networks and Deep Learning
Machine learning is a statistical technique that allows computers to learn from data without being explicitly programmed.
There are two main types of machine learning: supervised and unsupervised. Supervised learning is where the computer is given a set of training data, and its task is to learn to generalise from this data to make predictions about new data. Unsupervised learning is where the computer is given data but not told what to do. It has to learn from the data itself and try to find patterns.
One of the most popular types of machine learning is “deep learning,” a subset of machine learning algorithms known as “neural networks.” Neural networks are modelled after the brain and can learn by example.
Deep learning has been used to develop algorithms that can automatically detect tumours in medical images. It is also used to develop predictive models for various diseases.
Natural language processing: Chatbots and Virtual Nurses
Natural language processing (NLP) is a subfield of machine learning that deals with understanding and manipulating human language. NLP algorithms process and analyse large amounts of natural language data.
One popular application of NLP in healthcare is chatbots, computer programs that mimic human conversation. Chatbots can provide information on various topics, including health and wellness.
Virtual nurses are another NLP application that’s been gaining popularity in healthcare. Virtual nurses are computer programs that use NLP to provide personalised health advice and guidance to patients and health practitioners.
Rule-Based Expert Systems
Rule-based expert systems are AI applications that use a set of rules to make decisions. Typically, these rules are defined by humans, and the expert system is designed to mimic human decision-making.
Rule-based expert systems have been used in healthcare for various tasks, including diagnostic decision support, drug interaction checking, and disease management.
Robotics: Surgical Robots and Assistive Robots
Robotics is another area of AI that is increasingly making its way into healthcare. Surgical robots are perhaps the most well-known type of healthcare robot. These robots assist surgeons in performing surgery, and they have been shown to improve surgical accuracy and precision.
Assistive robots are another type of robot being developed for healthcare applications. Assistive robots can sense, process, and perform actions to support people with disabilities or chronic conditions. These robots help users with day-to-day tasks such as opening doors, pushing buttons, and picking up objects.
Robotic Process Automation
Robotic process automation (RPA) is a form of AI that uses software bots to automate repetitive tasks. RPA is used in healthcare for various tasks, including billing and claims processing, appointment scheduling, and prescription refill reminders.
Contrary to the name, RPA does not necessarily involve physical robots. Instead, RPA software bots are deployed on a computer to automate the tedious tasks often performed by humans.
It relies on combining workflows, user interfaces, and data to automate tasks.
How’s Singapore Building AI for Predictive Healthcare?
The development of AI in predictive healthcare is an area of active research in Singapore.
Several projects are underway to develop AI applications for various healthcare tasks, including early detection of disease, diagnosis and treatment recommendations, and predictive modelling of health outcomes.
At the National Hospital of Singapore (NHS), a team of doctors and researchers are developing a testing machine learning platform that automates the time-consuming process of reviewing and characterising a thyroid lump.
Ngiam Kee Yuan, the group CTO of the National University Health System (NUHS), is developing a system that automatically integrates AI models into Singapore’s healthcare system.
This system provides recommendations for diagnosis and treatment based on a patient’s medical history.
A Sandbox for AI in Healthcare
NUHS has also set up a “sandbox” environment called Discovery AI, which provides a safe space for researchers to experiment with AI in healthcare.
Discovery AI collates and aggregates data from various sources, including electronic medical records, clinical data repositories, biomedical literature, lifestyle habits, and genomic information.
The sandbox environment allows researchers to test AI applications in a real-world setting without disrupting the healthcare system.
The Future of AI in Healthcare
AI has the potential to transform healthcare, and Singapore is at the forefront of this transformation. With a strong focus on research and development, Singapore is well-positioned to become a global leader in AI in healthcare.
In the future, AI will play an increasingly important role in predictive healthcare, with applications ranging from early disease detection to predictive modelling of health outcomes.
AI will also be used to develop new drugs and therapies and provide clinicians with decision support.
Singapore Emerging as Global Hub for Healthcare AI
With its strong focus on research and development, Singapore is well-positioned to become a global leader in AI in healthcare.
It can all be traced back to 2019, when Singapore outlined plans to become a “Global Innovation hub” test-bedding, deploying, and scaling new technologies as part of its National Artificial Intelligence Strategy. Announced by Dr Vivian Balakrishnan, then-Minister in charge of the Smart Nation Initiative and then Minister for Foreign Affairs, the plan was to enable the country to grab the opportunities brought about by the 4th Industrial Revolution and create value for Singaporean citizens and the rest of the world.
Singapore has long aspired to be a launchpad for testing new ideas and innovative solutions. A test-bed for emerging technologies. A living lab where citizens can interact with technology in a safe and trusting environment. And a launchpad to scale up successful solutions for the rest of the world.
Since then, we have seen a flurry of announcements and initiatives by the Singapore Government to catalyse the development of Healthcare AI in the country.
The government invested S$500 million to further AI research innovation and adoption across different sectors. The country has since continued to raise its AI and R&D capabilities, strengthening the double helix partnership between the public and private sectors to make significant advances in AI. A good example is NExT++, a national program launched in 2019 that aims to catalyse the development of new technology and applications by supporting collaborations between companies, research institutes, and academia.
Five Covid-driven AI Trends That Will Reshape Healthcare
The healthcare industry has been under immense pressure this year as it scrambled to respond to the Covid-19 pandemic.
In the midst of all this, AI has emerged as a powerful tool that can be used to support and improve the delivery of healthcare services.
Here are five Covid-driven AI trends that are reshaping the healthcare sector:
#1. Virtual Care
The Covid-19 pandemic forced many healthcare organisations to rethink how they provide care. The need for social distancing forced virtual care to become an important part of the healthcare landscape.
Virtual care is about delivering healthcare services remotely, using technology such as telehealth and e-visits.
AI can be used to support virtual care in several ways. For example, it can give patients 24/7 access to their medical records, schedule appointments, and conduct video consultations with doctors.
#2. Expanding Lab Test Accessibility for Patients
The pandemic normalised at-home Covid-19 rapid testing kits, which made it possible for patients to get tested without having to visit a lab.
This trend is likely to continue even after the pandemic ends, as patients become more comfortable with the idea of at-home testing.
Remote clinical lab testing company Healthy.io uses AI, computer vision, and colourimetric analysis to turn a smartphone into a clinical-grade urinalysis device. The company’s app guides users through the testing process and provides them with results within minutes.
Varadharajan expects AI-enabled lab test accessibility to grow in popularity and eventually edge out traditional in-person lab testing for certain types of tests.
#3. Supporting the Adoption of Telemedicine
The pandemic has accelerated the adoption of telemedicine, as patients and doctors alike have embraced the use of technology to deliver and receive healthcare services remotely.
Telemedicine is the use of telecommunications and information technologies to provide medical care from a distance. It can be used for everything from consultation and diagnosis to treatment and follow-up care.
AI supports telemedicine in several key ways, such as by automating the scheduling of appointments, conducting video consultations, and providing 24/7 access to medical records.
#4. Reduced Radiology Costs
AI isn’t just changing the way healthcare is delivered. It’s also driving down the costs associated with pricey scans and tests.
Thanks to AI-assisted computerised tomography scans, radiology costs have dropped significantly in recent years.
And looking at the next wave of AI applications in healthcare, it’s safe to say AI will go beyond diagnosis to improve the patient’s overall healthcare experience. That translates to a quicker, more seamless, and affordable magnetic resonance imaging scan for the patient and fewer expenses for the hospital.
It’s worth mentioning that Facebook is collaborating with the New York School of Medicine to improve MRI scans and expedite the results using AI. The research team is working on a deep learning algorithm that can predict what an MRI scan will look like before it’s even taken.
If successful, the algorithm could be used to improve the speed and accuracy of MRI scans and even reduce the cost of these scans. That way, patients won’t have to make hours-long visits for the scans, as is often the case with MRIs. Fifteen minutes should be all it takes to get the results of an AI-assisted MRI scan.
#5. Picturing Computer Vision in Healthcare
The Unintended benefit of the global pandemic is that it spurred the development and adoption of new AI applications in healthcare, particularly in computer vision.
For those unfamiliar with the term, computer vision is a branch of AI that extracts information from digital images or videos.
Computer vision has many uses in healthcare. For example, it’s being used to develop contactless temperature screening solutions, which can detect fever in patients without physical contact.
It’s also being used to develop automated hand hygiene solutions to help ensure doctors and nurses correctly wash their hands before and after treating patients.
And last but not least, computer vision is used to build 3D models of organs, which can be used for pre-operative planning, surgical training, and patient education.
While computer vision is still in its early stages of development, it sure holds a lot of promise for the future of healthcare.
The BRAIN Behind it All
In 2017, Singapore’s health ministry planned to transform the healthcare sector in three major ways. The first was to make Singapore a leading healthcare innovation and research hub. The second was to use technology to empower patients and care providers. And the third was to develop a strong digital infrastructure to support care delivery.
One such solution is the Business Research Analytics Insights Network (BRAIN) platform.
BRAIN is a data analytics platform that uses AI to support the decision-making process of healthcare professionals. The platform consolidates data from different sources, including clinical records, patient surveys, and medical claims data.
BRAIN then analyses this data to identify trends and patterns that can be used to improve care delivery.
For example, in the case of diabetes, BRAIN allows the health ministry to determine who’s at risk of developing the disease and the most effective interventions for preventing or managing it.
To support patient safety and the ethical use of AI in healthcare, MOH to develop the MOH AI in Healthcare Guidelines (AIHGle — read as “agile”).
Click here to read the AIHGle
Share Good Practices with AI Developers: The guidelines share good practices with AI developers on how to develop, test, and deploy AI-powered healthcare solutions.
Ensure that AI is Used Responsibly: The guidelines ensure that AI is used responsibly by specifying the conditions under which it can be deployed. For example, AI can only be used if it is shown to improve patient outcomes.
Protect Patient Privacy: The guidelines protect patient privacy by specifying the conditions for collecting and using data. For example, data can only be collected with the patient’s explicit consent.
Complement HSA’s Guidelines of AI-MDs (AI-Medical Devices): The AIHGle complement HSA’s Guidelines of AI-MDs, which specify the conditions under which AI-powered medical devices can be used.
The guidelines are a living document that will be updated as the landscape of AI in healthcare evolves, such as the Smart Nation and Digital Government Office (SNDGO), to develop a set of ethical principles for using AI in the public sector.
Singapore Wants AI Startups to Think about Healthcare
The government is looking to steer AI technology development towards solving healthcare needs in the country.
The three-year collaboration between SGInnovate and SingHealth offers resources and opportunities for AI startups to develop prototypes and pilot their solutions with clinicians
The partnership aims to drive the adoption of AI and other frontier technologies to improve diagnostics, treatment, healthcare delivery, and clinical outcomes for patients in Singapore.