AI can potentially play a crucial role in global public health, combating epidemics and pandemics. Despite all the hype, AI adoption within the healthcare sector has been dismal.
AI (Artificial Intelligence) has emerged rapidly as one of the most disruptive generic technologies in the last decade, impacting several fields such as banking, finance, insurance, and healthcare. AI experts claim the wide-ranging prowess and ensuing possibilities, which may match or even exceed human intelligence and capabilities on tasks such as complex decision-making, reasoning and learning, sophisticated analytics and pattern recognition, visual acuity, speech recognition and language translation. It is in this background that there has been a growing rhetoric of AI’s potential to change the face of healthcare delivery around the world.
AI is expected to solve many of the challenges in the healthcare ecosystem. It can help address the lack of adequate numbers of trained medical specialists, improve access to healthcare, enhance preventive care and quality of life, produce more accurate diagnoses and treatment plans, and overall, it can lead to better patient outcomes. As a result, AI can potentially play a crucial role in global public health, combating epidemics and pandemics. Despite all the hype, AI adoption within the healthcare sector has been dismal. According to a McKinsey report of March 2020, the lack of multidisciplinary development, lack of early involvement of healthcare staff in the developmental process, and limited collaboration by AI and healthcare teams, were cited as major barriers to addressing quality issues early in the development cycle of AI software in healthcare and adopting these solutions at scale. As a testimony of the challenges relating to adoption of AI in healthcare, a 2020 survey by the US-based research group Brookings found that less than 1 percent of job postings in healthcare required AI-related skills.
Our interactions with experts in both AI and Healthcare in India reveal that there are at least five challenges which slow down the adoption of AI in healthcare.
1. Learning by making mistakes
Experts say that AI is a natural learning platform, and it improves itself from errors that are made through its learning during its development phase. Errors in healthcare AI systems can be a cause for serious concern and can even mean life or death of the patients. There is considerable reluctance from both doctors and patients to be early adopters of unproven AI systems. Neither of them wants to be guinea pigs during the time when the AI system is learning! The recent IBM Watson Health’s experience is a standing example of potential pitfalls. One of Watson Health’s biggest setbacks was the revelation that its cancer diagnostics tool was not trained with real patient data, but instead with hypothetical cases provided by a small group of doctors in a single hospital. This data reflected the doctors’ own biases and blind spots and wasn’t necessarily generalizable to all patient cases. As a result, the tool was accused of making inaccurate and unsafe recommendations, leading high-profile hospital partners to cancel their collaborations with Watson Health. It is important for AI systems to learn to fail gracefully when they are uncertain (which often means abstaining from making predictions), rather than give inaccurate predictions.
2. Access to data (patient privacy & authenticity of data)
A major concern regarding healthcare data is of privacy and ethicality. One of the major concerns the patients have is that once their healthcare data is online, it could be used by regulatory agencies/banks to track them. Their second concern is, ‘How will medical insurance companies use/misuse this data for processing claims?’ The fact that AI systems depend on the availability of enormous amounts of data poses a major impediment for building indigenous AI interventions in India. Datasets for healthcare in India are fragmented, dispersed and incomplete to say the least. The data that are readily available for AI companies are thus likely to be unrepresentative of a significant segment of the population. Hence the solutions being generated by the AI software may not be accurate and may not be appropriate for the entire population.
3. Ethical/Moral/Security considerations
Healthcare decisions have been made almost exclusively by humans in the past, and the use of smart machines to make or assist them raises issues of accountability, transparency, consent and privacy, not to mention the fact that a patient is more amenable to a doctor whom he/she trusts giving out treatment options, than a software doing the same function. Accountability is difficult to assign in AI-driven treatment options. A case in point is the recent cyberattack on the prestigious AIIMS (All India Institute of Medical Science), Delhi, which brought under the scanner the safety of electronic medical data, which can be held ransom by cyber attackers or can be encrypted and made inaccessible and unusable.
The dataset in India pertaining to healthcare is scattered, unverified, and largely representative of the urban population. Thus the data that is readily available for the AI companies would most likely be unrepresentative of the entire population.
4. Regulatory Restrictions
In India, the current legal framework pertaining to e-health protection is governed by the provisions of the Information Technology Act, 2000, and the Information Technology Rules, 2011, which together offer some degree of protection to the collection, disclosure, and transfer of sensitive personal data, which covers within its ambit medical records and history. However, legislation has not been updated to take into consideration the rapid development in technology and leaves many aspects unaddressed. Considering this, the Government has introduced DISHA (Digital Information Security in Healthcare Act of the Government of India) and the Personal Data Protection Bill, 2019 (“PDP Bill“). These bills when implemented, will change the legislation about data protection (both personal and health data) in India, making it more in tune with global standards. Until that time, the issue of unregulated AI solutions is a major concern for different stakeholders in the healthcare ecosystem.
5. Reluctance to adoption of AI-based healthcare systems by the healthcare community
AI-based diagnosis and treatment recommendations are challenging to embed in clinical workflows and EHR (Electronic Health Record) systems. Most of the healthcare workforce in India is either unaware or apprehensive about AI. A major reason for this apprehension is the fear of losing their jobs to AI based solutions. Another reason is the lack of knowledge about the positive impact AI-based solutions can have in day-to-day medical practice.
A way forward to break this imbroglio in the present scenario is by creating an AI ecosystem through partnerships to co-develop the right solutions for the local population and to make a compelling narrative on the value generated through AI-based solutions, both with patients and practitioners. This can be done by defining and developing the right use cases jointly with the end users. Another crucial step would be to define and address skill gaps in digital literacy for the healthcare staff in an organisation, which may provide a strong value proposition for AI talent to join and stay with the organisation. Lastly, it is essential to address data-quality, access, governance, and interoperability issues, while nurturing a culture of entrepreneurship in the organisation.
Composed by: “Shweta Jaiswal is a senior Healthcare Professional with over 18 years of experience as an Anaesthetist & Intensivist. She has worked across India in various tertiary care hospitals. Her field of expertise is Cardiac & Neuro Anaesthesia and Cardiac Critical Care & Transplant Critical Care.”
“DVR Seshadri is a Professor of Practice in the Marketing Area at the Indian School of Business. His areas of interest are healthcare, business-to-business marketing, and climate change.”
“Raj Krishnan Shankar is an Associate Professor of Strategy and Entrepreneurship at Great Lakes Institute of Management, Chennai.”
