In an era where technology continually pushes the boundaries of innovation, artificial intelligence emerges as a transformative force, particularly in healthcare. One of the most promising applications of AI lies in early disease detection, revolutionizing how we approach healthcare and paving the way for more proactive and personalized interventions. The ability of AI algorithms to analyze vast amounts of data with speed and precision opens new avenues for identifying subtle patterns and anomalies that may signal the onset of various diseases.
We asked Naithy Cyriac, a Partner from the YCP Solidiance Singapore office with extensive experience advising healthcare providers, pharmaceutical companies and medical device firms across Asia Pacific, for her insights on the role that AI plays in early disease detection, exploring its potential to redefine diagnostics, improve patient outcomes, and ultimately reshape the landscape of healthcare.
What are the most promising areas of application in AI for early detection?
AI has proven to be valuable for the early detection of diseases across various medical fields. It can help enhance diagnostic accuracy, detect early signs of diseases, and has the potential to enhance patient outcomes by enabling timely interventions and personalized treatment plans.
In oncology, when AI is used to analyze medical imaging data, such as mammograms, CT scans, and MRIs, it can help detect cancerous lesions. It can also help identify patterns and anomalies that may be indicative of cancer at its early stages. In cardiology, AI can be employed for analyzing electrocardiograms (ECGs) and other cardiac imaging data to detect early signs of heart diseases, such as arrhythmias or abnormalities in heart structure. AI applications in genomics are especially fascinating as it can analyze genetic data to identify patterns associated with hereditary diseases and conditions, which can help with early detection and the development of personalized treatment approaches.
In Thailand, the IBM Watson supercomputer analytics has been integrated into the oncology department at Bumrungrad International Hospital to advise doctors on the best treatment plans for cancer patients. In Indonesia, AI is being used by a startup named CekMata for the early detection of cataracts.
In another interesting example of proven AI applications in Southeast Asia, the National Environment Agency in Singapore has been utilizing data analysis and predictive modeling to anticipate and monitor the outbreak of dengue fever. AI algorithms analyze data such as weather patterns, population movement, and historical dengue cases to predict potential hotspots and enable proactive measures. Hence, the possibilities are diverse and impactful for patients, healthcare facilities and governments alike, and the field continues to evolve rapidly.
What are the limitations of using AI for early detection, and what support do healthcare professionals need to use AI effectively?
AI models in healthcare rely heavily on the quality of the data they are trained on. If the training data is biased or lacks diversity, the AI model may produce biased results. AI models developed for a specific disease or population may not be relevant to other diseases or populations. The lack of standardized protocols for developing, testing, and deploying AI models in healthcare can hinder interoperability and collaboration among different healthcare systems and institutions. This will also impact integration with existing clinical workflows or health management systems, requiring a whole systems approach within an organization.
Ensuring the reliability and safety of AI models for healthcare applications may also require extensive validation and regulatory approval. As a result, significant upfront costs may be required in terms of infrastructure, process transformations, and training. Without seamless integration, it may be challenging for healthcare professionals to adopt new technologies that may disrupt established routines.
Targeted training for healthcare professionals is a prerequisite for AI deployment. Shortages of skilled professionals, limited digital literacy among healthcare providers, and the need for specialized training are common challenges that arise. This is where public-private partnerships can help bridge the gap. As part of a Memorandum of Understanding signed between AI Singapore, the national program established to boost AI capability, and SingHealth, healthcare professionals in Singapore will receive AI training and qualifications catered for their industry, in the form of e-learning videos, physical classes, and workshops starting in 2024.
How will this innovation impact healthcare disparities and access to care?
Access and affordability of care remain critical challenges in the healthcare sector. Telemedicine gained prominence during COVID-19 as a valuable tool to access healthcare from the confines of homes and is also very effective in reaching remote populations, or in the case of countries like Indonesia or the Philippines with populations spread across multiple islands.
AI-based telemedicine platforms can help leverage machine learning algorithms for preliminary diagnosis. In Indonesia, for example, the government used an AI-powered app called Telemedicine Indonesia to link patients with hospitals and doctors during the height of the COVID-19 crisis. Although in its early stages, we are seeing similar deployment of AI in the Philippines in telemedicine platforms, personalized health monitoring systems, and diagnostic tools driven by various startups and other private companies. AI-powered telemedicine platforms that allow patients to seek medical advice remotely also help to reduce the burden on healthcare facilities, minimize travel costs for patients, and facilitate early discussions about health concerns.
The deployment of a deep learning AI software called Selena+ in Singapore for analyzing eye images to flag abnormalities as signs of possible diabetes retinopathy in the form of a semi-automated model is to result in 20% cost savings. By 2050, Singapore is projected to have one million people with diabetes. At this rate, the estimated annual cost savings would be 15 million Singapore dollars.
In Malaysia, AI is helping to reduce healthcare costs by automating routine tasks, such as scheduling appointments and processing insurance claims. AI can also help reduce the length of hospital stays and associated costs for patients by predicting which patients are at risk of complications and intervening early to prevent them. A pilot study conducted in smaller healthcare facilities in Da Nang, Vietnam has also demonstrated how AI can especially support hospitals with limited resources and reduce the burden on healthcare professionals, who are often overworked and understaffed. Hence, these are some of the many examples where AI is helping to improve patient outcomes and reduce healthcare costs while easing the burden on healthcare professionals.
Are there ethical considerations that must be considered when developing and deploying AI for early detection? Also, will there be an assurance that these AI-powered systems are fair, transparent, and accountable?
Southeast Asian governments have already taken initial steps in advancing AI, with Singapore leading the charge when it launched its National AI Strategy in 2019. Indonesia, Malaysia, the Philippines, Thailand, and Vietnam have also drafted their national AI strategies and roadmaps between 2020 and 2022. The need to balance the perks and challenges of AI has pushed Singapore, as rotating chair of the ASEAN Digital Ministers’ Meeting and Related Meetings in 2024, to collaborate with partners in the grouping to develop an ASEAN Guide on AI Governance and Ethics. The guide is expected to serve as a “practical and implementable step” towards supporting the safe deployment of “responsible and innovative AI” in the region.
In healthcare, it is vital to ensure that patient data is kept secure and that AI is used to respect patient privacy. Various personal data privacy laws are still applicable. Thailand’s Personal Data Protection Act requires explicit consent to collect sensitive personal data, with certain exemptions. It is allowed when mitigating danger to the data subject's life or health or when it is necessary for legal claims or compliance. This can include preventive medicine, employee health assessment, medical diagnosis, care system management, public health interests, employment protection, social security, research purposes, and substantial public interests.
Indonesia’s Personal Data Protection Law defines personal data handling principles, including consent, data updating, breaches, transfers, and sanctions. The country’s MOH Regulation No. 24 of 2022 specifically governs personal health data, outlining obligations for storage, deletion, and confidentiality of medical records, encompassing a patient's identity, examinations, medications, and related services. Both define health data as records or information about an individual's physical and mental health or health services. Thailand also does not allow medicines to be prescribed through a teleconsultation, while Indonesia mandates teleconsultations to be only conducted at registered health facilities. With the proper integration and training, institutionalizing governance processes, and strengthening data privacy regulations, AI has the potential to revolutionize healthcare in Southeast Asia and globally.
How do you see AI changing the way diseases are diagnosed and treated in the next 5-10 years?
Southeast Asia is still facing limited private sector investment and slow response from governments to keep up with rapidly evolving AI technology and the expanding pool of potential AI users. The region is also subject to relatively inequitable and sparse provision of ICT infrastructure. Apart from Singapore, most countries in the region struggle to keep pace with the needed internet speed and widespread infrastructure for advancing digital health. As a result, AI adoption is still in the nascent stages. While many companies are piloting AI initiatives, only a few are in the advanced stages of AI implementation. Addressing these limitations over the next 5-10 years requires ongoing research, collaboration between healthcare professionals and AI developers, regulatory frameworks, and a commitment to ethical AI practices in healthcare.
For healthcare providers, enhanced integration of AI tools with imaging devices and hospital management systems will allow clinicians to increasingly rely on AI-driven decision support systems for diagnostic assistance, treatment recommendations, and the interpretation of complex medical data. Increased natural language processing capabilities will streamline clinical documentation, making it easier for healthcare providers to input and retrieve patient information, improving overall efficiency. Voice-activated AI interfaces may also pave the way to facilitate natural communication between healthcare professionals and AI systems, enhancing the accessibility of information.
For patients, AI can facilitate access to increasing levels of individualized care based on genetic predispositions, lifestyle factors, and environmental influences, enabling proactive interventions for prevention. AI can also further strengthen current telemedicine platforms to offer personalized health insights, allowing patients to receive timely advice and interventions without the need for in-person visits. AI-powered wearables are another interesting segment that can enable continuous monitoring of vital signs, providing real-time data for the early detection of health issues and more effective management of chronic conditions.
Naithy is our Partner based in Singapore, with over a decade of consulting experience. She has worked extensively with MNCs exploring investment opportunities in Asia across strategy and operations consulting projects such as market entry strategy, channel and customer engagement, M&A advisory, business and strategy planning with other project stints in Australia, Iran, Malaysia, Singapore, Thailand, and UAE.