The transformative power of Artificial Intelligence (AI) is reverberating across many industries, but in healthcare, its impact promises to be truly life-changing. From administrative healthcare functions to clinical research including medical imaging, risk analysis, diagnostics and drug development, AI is revolutionizing the workings of the healthcare sector with the goals to improve patient outcomes and make healthcare more affordable and accessible.
According to Frost and Sullivan analysis, the growth in the AI health market is expected to reach $6.6 billion by 2021 from 600 million in 2014, that is the compound annual growth rate of 40 percent [1]. AI can be considered as a self-sustaining engine in the context of healthcare. As per an analysis conducted by Accenture, key clinical health AI applications can create $150 billion in annual savings for the United States (US) healthcare economy by 2026 (figure 1) [2].
Figure 1: Estimated worth top ten AI applications in the US by 2025 [2].
The number of startups entering the healthcare AI space has augmented in recent years and will continue in the future. The market map shows different companies transforming different areas of healthcare with AI in the US (figure 2) [3].
AI in Healthcare: Global initiatives
While the United States leads in developing and using AI, many other countries, including China and the European Commission, are not far behind. The European Commission (EC), EU member states, Norway and Switzerland have come together for the development and use of AI in Europe, with the prime focus being healthcare and transport. They plan to emphasise on four significant areas – fostering and promoting research and talent, democratizing data, ensuring trust and increasing investment to reach at least €20 billion of private and public partnership (PPP) investments by the end of 2020 as told by the Vice President for the Digital Single Market, Andrus Ansip [4]. At the beginning of 2019, hospitals in the Guangdong province of China deployed 14 AI-enabled cameras to detect blindness causing diseases, which were co-developed by Baidu and Sun Yat-sen University.
To secure and maintain its position, the US government has recently issued an American AI Initiative, which appeals for the federal government to prioritize research and development for machine learning (ML) in healthcare and other relevant areas [5]. The order calls for the government to promote technical education and training, boost STEM and computer science in schools and universities, increase federal investment in AI research and development and provide access to federal data and other computational resources for AI researchers.
AI in healthcare: India
India is a big country with a vast population spread across different geographical locations and economic strata in society. Thus, making the accessibility of primary healthcare facilities difficult to everyone. The disparity in the availability of healthcare can also be seen across states and gender; for example, the life expectancy for women is around 66.8 years in Uttar Pradesh and 78.7 years in Kerala whereas for men, the numbers are 63.6 years in Assam and 73.8 years in Kerala [6]. With lagging infrastructure and less number of doctors per 1000 citizens, it is challenging for the government to promise good quality healthcare to its citizens.
The diversity and potential scale of the Indian healthcare system provide tremendous opportunities and incentives for AI to present creative and innovative solutions. According to a report by Accenture, AI could help add USD 957 billion to the Indian economy by 2035 (figure 3) [7].
AI in India is developing steadily, and the government is taking multiple steps to promote and implement the applications of AI in numerous areas including the Indian healthcare system. The national strategy for artificial intelligence document by NITI Aayog, a policy think-tank and task force on artificial intelligence report by Ministry of Commerce and Industry, Government of India provide roadmaps to the government’s AI ambitions and the sectors where AI can be leveraged in India, the challenges specific to India and the ethical concerns [8, 9].
This brief focusses on the key factors and approaches taken by the Indian government, private organizations and universities, their immediate progress and projected future advancements in the realm of AI in healthcare.
a) Research, collaboration and funding
Global tech giants independently or in collaboration with the government, non-profit organizations, Indian universities and other private entities have ventured big-time into AI healthcare in India and adopted the public-private partnership (PPP) models.
Microsoft India’s AI Network for Healthcare” initiative and Apollo Hospitals launched the first-ever AI-powered heart disease risk score API (application program interface) to predict the risk of cardiovascular disease (CVD) among the Indian population, and further help doctors drive preventive cardiac care. It has also announced a partnership with SRL Diagnostics to expand the AI to pathology to detect cancer. The state government of Telangana has implemented Microsoft Intelligent Network for Eyecare (MINE) that was developed in partnership with LV Prasad Eye Institute from Hyderabad. MINE uses ML and advanced data analytics to predict regression rates for eye operations. This aids the doctors to identify and decide the steps to follow to prevent and treat visual impairments [8].
To help bolster AI research, one of the government’s initiatives includes funding Indian researchers, scholars and university faculty for conducting AI-based research. Public source of funding includes from organizations like – Department of Biotechnology (DBT), Defence Research and Development Organization (DRDO), Ministry of Electronics and Information Technology (MeitY). As per the Global AI talent report 2018 ranks India 10th globally in terms of number of PhDs in AI research and 13th in terms of presentations in AI conferences [10]. Recently, a group of scientists funded by the DBT-India Alliance/ Wellcome Trust have developed a non-invasive, non-contact modality constructed using a combination of thermal-imaging and machine-learning to detect and predict shock that affects almost 30 percent of ICU patients up 12 hours in advance in patients in intensive care units (ICUs) by monitoring their hemodynamic status. A delay in detection leading to ineffective management of shock often leads to rapid deterioration of tissue function, failing organs and eventually death [11].
Further, Niti Aayog has partnered with Google to provide online training courses on AI to students, graduates and engineers in numerous cities across India. Educational and research institutes such as Indian Institute of Science (IISc) and Indian Institute of Technology (IITs) have decided to introduce AI as a full-fledged course at the undergraduate and graduate levels. Additionally, short term courses are also being offered by companies in collaboration with the educational institutions to provide training in AI. Such courses are primarily targeted to the existing working-class who wish to update themselves with the latest technology. IBM has joined hands with IIT-Delhi (IIT-D) to participate in a multi-year research collaboration on AI in India, with a focus on sectors such as healthcare and medicine.
Besides, the Central Board of Secondary Education (CBSE) has introduced AI as an optional skill subject in class IX at the school level from the academic session 2019-2020. The board has assured that it will provide the necessary support and guidance towards the training and capacity building of teachers and other aspects for the successful implementation.
b) IT infrastructure, data storage and access to information
Cheaper computing power, data storage via cloud and digitization of data are considered as important factors driving the AI/ ML boom. Thus, attempts towards shifting towards electronic counterparts of paper-based reports are encouraged, and continuous upgrade of the IT infrastructure and development of AI data storage is being looked into.
For AI to work, having a vast amount of structured clean data is critical. National AI Marketplace (NAIM) is proposed by Niti Aayog that will collect and annotate data and evolve deployable models. Also, the access of data and open-source AI technologies has been initiated and proposed by the government under the Digital India Initiative. The Indian government has already provided access to a range of data collected by various ministries through the Open Government Data Platform (www.data.gov.in/sector/health).
The Ministry of Health and Family Welfare (MoHFW), Government of India has set up the national health portal (NHP) (www.nhp.gov.in/), which aims to establish single point access for authenticated health information for citizens, students, healthcare professionals and researchers. The users will have access to detailed information related to diseases, health services and programs, insurance schemes, health apps, helpline numbers and blood bank details etc.
Augmenting of computing infrastructure for AI/ML research and applications by setting up a national high-performance computing infrastructure that is rich in Graphics processing units (GPUs) and specialized hardware for AI research is taken up by a number of universities. Additionally, the design and manufacture of such systems in India are also encouraged. The report by Niti Aayog proposes the setting up a national computing infrastructure known as AIRAWAT, 100-petaflop computing system for AI/ML applications. Several foreign and Indian private companies such as IBM, Microsoft and TCS, Wipro and Infosys respectively have ventured into providing the IT infrastructure for AI in India.
c) Conceptualization and implementation
The Indian Start-up ecosystem is demonstrating an increasing trend of applications based on deep learning and AI in the healthcare domain. Philips is engaging with entrepreneurs who are developing AI-enabled solutions for improving clinical and operational outcomes. In that effort, Philips has introduced a start-up collaboration programme based on the application of artificial intelligence (AI) in healthcare and selected the most promising 19 Indian start-ups from 750 applicants working on AI in healthcare [12].
Private investors are actively supporting and investing in AI start-ups in various domains, including healthcare and agriculture sectors. Recently, Niramai Health Analytix, an AI-based health tech startup from Bangalore, has raised $6 million in Series A funding led by Japanese VC firm Dream Incubator and other investors. Its product thermalytix is a portable, non-invasive and non-contact solution for early-stage breast cancer detection. The capital generated will be primarily used to increase the operations in India, hiring workforce, and for international expansion [13]. Microsoft’s corporate venture fund M12 has made its India entry with its first investment in healthcare data analytics startup Innovaccer. M12 has invested an additional $10 million in Innovaccer’s Series B round, leading the total investment to $35 million. Innovaccer leverages on machine learning and healthcare-related expertise to enable its users to consolidate financials, claims, patient, and operational data together to provide a comprehensive patient 360-view for better decision-making, care coordination, and reporting [14].
As part of the UK-India Tech Partnership, UK-India Healthcare Artificial Intelligence Catalyst (HAIC) program has been set up that oversee the delivery and execution of AI-based national health programs in India. The HAIC has invited and encouraged applications from UK based to participate in ‘AI based healthcare solutions’ competition to revolutionize the Indian healthcare system. UK has decided to invest £1 million in this collaboration with India [15].
Challenges of AI-healthcare in India
Artificial intelligence (AI) can and is being applied to almost any field in healthcare and medicine, and its potential contributions to biomedical research and delivery of health care seem limitless. It is going to be involved in every sector by replacing the existing systems. It will hamper the existing system in place and create a new and robust one. It will be like the acceptance of computer technologies (CBS) from the traditional typewriter in the banking system. In a large developing economy like India, AI will be needed to sustain its economy as in future much of the commercial services will be driven by AI-based mechanism. This powerful technology creates a novel set of challenges that must be identified and alleviated.
In India, there is already an absence or scarcity of the public health infrastructure in underserved areas. Here, the proposed adoption of the AI technologies by the healthcare sector comes into inquiry. No doubt AI will make the healthcare sector accessible to all, but then the question arises how to make it available and affordable at the same time. We need adequate infrastructure, a skilled workforce, increase R&D spending etc. Even in a metropolis like Delhi, we lack the proper medical facilities in the government hospitals, where the majority of the people visit.
AI technology has tremendous capability to threaten patient preference, safety, and privacy. At present, legal, policy and ethical guidelines for AI technology are lagging behind the progress AI has made in the realm of healthcare. There is a pressing need for the policymakers, healthcare AI developers, and medical practitioners in India to consider all the ethical and legal aspects while planning, implementing, and regulating AI in the healthcare field. Some of the most urgent concerns include addressing the added risk to patient privacy and confidentiality, liability among the physician, hospital, and the AI system developer, trainer and manager in case of medical negligence. AI is all about data collection and creates the complex issue of privacy and IPR. For example, data in AADHAR has been breached many a time, if we are going to implement the AI system, then the issue for protecting the data should be of utmost priority. There is the need for the creation of new skill set to supplement the AI set up.
The labour laws in India have largely remained unchanged after India’s independence. Some of these laws may have to be amended to keep them relevant in context of AI-driven workforce in healthcare. One of the shortcomings of the IT Act, 2000 that makes it inept in the AI technology era is the limited onus and liability. Section 79 of the Act states that no person providing any service as a network service provider would not be held liable for the misuse of information and data barring some exceptions. This rule may have to be re-examined and reframed keeping in mind the applications and use of AI in healthcare [16,17].
There are other challenges of AI in terms of the Indian context that are not just limited to healthcare but to other fields that use AI or could potentially benefit from it. Lack of good quality labelled data sets and lack of adequate computer infrastructure to perform state of the art AI research and develop subsequent applications, mundane AI research in universities, limited university-industry interaction and insufficient funding are primary sources of lack of enthusiasm in students and researchers to pursue AI research in India and they prefer to go abroad where there excel, thus leading to brain drain.
What can be done?
To address the challenges that are being or in future be encountered by AI, a comprehensive ecosystem needs to be developed to ensure compliance so that the benefits of AI outweigh the risks within the context of Indian healthcare. Firstly, there is a need for the creation of a new skill set to supplement the AI set up alongside the infrastructure requirements. For this the investments are massive. Just by launching a policy framework won’t help. The policy that would come up should be broad enough to cover the entire system by balancing both the old and new system.
It is time for the Indian lawmakers and the judiciary to step in and set up a comprehensive legal framework and pave the way for a successful transition of the Indian healthcare into the age of AI and solve India’s complex healthcare problems. In addition, relevant policies are required to be framed that address the concerns of the degree of openness of AI technology, privacy and security of data, appropriate calculation of risks associated with AI and political factors associated with the use of AI for the common good.
As the challenges and limitations in implementing AI have been raised, appropriate measures are being taken to overcome such bottlenecks. The adoption of AI in India is at a turning point and ready for sustained growth. All the aspects ranging from the availability of technology, resources and government involvement is driving AI in the right direction.
References
[1] Frost & Sullivan, http://ww2.frost.com/news/press-release/600-m-6-billion-artificial-intelligence-systems-poised-dramatic-market-expansion-healthcare
[2] Accenture, https://www.accenture.com/us-en/insight-artificial-intelligence-healthcare
[3] CB Insights; “From Virtual Nurses to Drug Discovery: 106 Artificial Intelligence Startups In Healthcare;” posted February 3, 2017big-time at https://www.cbinsights.com/blog/artificial-intelligence-startups-healthcare/
[4] https://ec.europa.eu/digital-single-market/en/news/coordinated-plan-artificial-intelligence
[5] https://www.whitehouse.gov/presidential-actions/executive-order-maintaining- american-leadership-artificial-intelligence/
[6] India: Health of the Nation’s States. https://www.healthdata.org/sites/default/files/files/policy_report/2017/India_Health_of_the_Nation%27s_States_Report_2017.pdf
[9] https://dipp.gov.in/whats-new/report-task-force-artificial-intelligence
[10] http://www.jfgagne.ai/blog/2018/2/7/the-global-ai-talent-pool-going-into-2018
[11] Nagori et al. Predicting Hemodynamic Shock from Thermal Images using Machine Learning. Scientific Reports. Volume 9, Article number: 91 (2019)
[12] https://www.analyticsindiamag.com/philips-launches-startup-collaboration-programme-ai-healthcare/
[13] https://www.dealstreetasia.com/stories/niramai-dream-incubator-123028/
[14] https://news.microsoft.com/en-in/m12-innovaccer-investment-india/
[15] https://digitalhealth.london/the-uk-india-healthcare-ai-catalyst-haic-competition/
[16] Indian Law is yet to Transition into the Age of Artificial Intelligence, The Wire (September 26, 2016)
[17] https://www.wipo.int/edocs/lexdocs/laws/en/in/in024en.pdf