Restoring care in Singapore's healthcare – with AI

Just last week, researchers in Singapore achieved a milestone in healthcare: they created the world’s largest genetic databank of Asian populations. Thanks to sequencing data of around 5,000 Singaporeans, scientists and doctors will now be able to better diagnose rare diseases and find the causes of chronic ailments.

This four-year project was started partially in response to mounting pressure on Singapore’s healthcare system. Whilst being one of the world’s most effective and efficient – raking second only to Hong Kong on Bloomberg’s Health Efficiency Index – it is faced with rising demands of an ageing population and chronic diseases.

The Ministry of Health (MOH) described this rate of growth in healthcare spending as ‘unsustainable’. A potential solution to this problem could be tackling a major challenge doctors struggle with nearly every day: making the correct diagnosis. Patients trust their doctors to diagnose them accurately and quickly. However, doing so can be difficult when a patient has a rare condition or non-specific symptoms. To make matters worse, doctors usually don’t have a lot of time to spend with the patients, especially in countries that are faced with an ageing population and shortage of healthcare professionals.

In Singapore, it is not uncommon for the wait time to see a doctor to be longer than the duration of the actual consultation. With a doctor to population ratio of 1:410, most consultations don’t last more than 10 minutes. As a result, doctors might not have enough time to obtain enough information on the patient’s current symptoms, lifestyle, and medical and family history. This may increase the likelihood of misdiagnoses or delayed diagnoses.

Humanising healthcare
Some countries are turning to artificial intelligence (AI) to deliver better patient care. Singapore, for example, is ranked third in the world in terms of use of AI in healthcare, with automating administrative tasks and improving diagnostic accuracy as the most common use cases.

Whilst there are concerns about AI replacing human healthcare professionals and eliminating the human touch in medicine, those worries are unfounded. The reality is that AI will augment healthcare professionals’ abilities and re-humanise healthcare in the following ways:

• Empowering doctors to make more efficient diagnoses

Since AI is less likely to make errors, it can be used to detect signs that may be hard for the human eye to spot. For example, AI can reduce diagnostic errors in breast cancer detection by 85%. It can be “trained” to see the very earliest changes in cell structure that typically lead to the development of cancerous cells. Oncologists can therefore be alerted and guide patient care protocols with greater care and effectiveness.

• Recommending personalized treatments

Unlike humans, AI can analyse massive sets of data from both medical literature and the patient’s history. For the latter, it can not only know the patient’s other medical conditions but also track outcome patterns following a treatment to identify the right treatments or medication based on the patient’s profile. Individualised interventions and treatments can help eliminate post-treatment complications and accelerate recovery. In the case of heart patients, using AI has helped 61% of them to avoid invasive angiograms, which in turn reduced treatment costs by 26%.

• Supporting patient’s mental health and daily tasks

To ease some healthcare professionals’ workload (especially in nursing home or community hospital settings), AI-powered service robots can engage patients in conversations and remind them to take their medicine so that they will feel less lonely. These robots can also perform basic routine check-ups like temperature, blood pressure, and sugar levels.

• Accelerating the development of life-saving drugs

Isolating molecules is a time-consuming task and it can take researches months before they find a promising result. AI can speed up this process drastically by automating the analysis of potential molecule possibilities – giving researchers time back, which they can use to focus studying the most promising molecules. As such, AI can increase the productivity of research into drugs for Alzheimer’s, cancer, and multiple sclerosis by as much as 10 times.

Unlocking AI’s potential in healthcare
Recognizing the transformational value of AI, the Singapore government has provided the fundamental ingredients for widespread AI adoption. It has pushed out relevant policies, is offering grants and supporting various public-private initiatives to catalyse the development and deployment of AI to tackle real-world issues including those faced by the healthcare sector.

However, it takes more than that to harness the full value of AI. For AI projects to be successful, healthcare institutions need to start with choosing an appropriate initial use case, before putting a skilled team together and building a solid IT infrastructure.

Since AI is fueled by data, it is crucial for healthcare institutions to have an effective and efficient data pipeline that allows them to capture, prepare, access, move, and protect large volumes of data from multiple sources (be it the edge, core compute, or the cloud). This will enable them to train, validate, and operationalise an AI algorithm to identify patterns, develop predictive insights, and enable increasingly accurate autonomous systems. By using AI to automate manual tasks and act as a complementary decision support, healthcare professionals will gain more time to re-establish the human connection in the doctor-patient relationship and deliver enhanced care.

In sum, delivering effective patient care is not an easy task, as shown in Angel’s case. Challenges such as a limited number of healthcare professionals and emerging diseases have led to deteriorating doctor-patient relationships, as doctors do not have sufficient time and insights to fully understand what is going on with their patients. Doctors are therefore forced to be reactive and focus on treating symptoms and illnesses as they come up, instead of being proactive and helping their patients stay healthy.

From medical imaging to genomic analysis to drug discovery, AI is getting better and more sophisticated at doing what humans do — doing it more accurately, more quickly, and at lower cost. By freeing healthcare professionals from manual tasks and providing greater insights to patients, AI can help re-humanise patient care as well as empower medical professionals to deliver more effective and efficient care to patients. However, only healthcare institutions with an AI-ready infrastructure – one that breaks down data silos, connecting disparate datasets to generate deeper insights, and be able to seamlessly move data from devices at the edge to the core compute and to the cloud – will be able to unleash the full potential of AI and deliver enhanced patient care.