How AI is Transforming the Healthcare Landscape
Praveen had extensive experience in Supply Chain Management and ERP solutions, having worked at several global companies including i2 Technologies, where he helped customers streamline operations and adopt best practices.
In a conversation with Keerthana H K, Correspondent, CIO Insider Magazine, Praveen Singh Bist, CIO at Amrita Hospital, shared his views and thoughts on the key benefits that AI can bring to healthcare as well as how we can implement AI tools to enhance early disease detection.
According to you, what are the main challenges to implementing AI in healthcare, and how do you address them?
Main challenges and obstacles to implementing AI in healthcare are as follows:
(a) Having complete digitized data (EMR) for the patient, both from the requirement of input data to train / validate the AI model in Indian setting as well as for implementation of AI based Precision medicine for a specific patient. Now that Indian Government has laid down Digital Health policy (ABDM), we need to collectively start adhering to its requirements.
(b) Transparency issue around the decision making – so need some kind of explainable model/interface which can help answer the common questions like, why did AI gave the answer it came out with?
(c) Ethical questions around the responsibility – who is responsible if something goes wrong? So should be clear policies and guidelines around this.
(d) Since AI is an evolving field, the lack of trained resources is a huge challenge and we have to invest in capacity building at all levels including training medical professionals starting with their college as well as IT resources.
(e) Unless the AI solution / module get properly integrated with the current healthcare systems, it will not be easily adopted by clinicians. So AI should be a part of the clinician workflow, ‘Day-in-the-life-of’, instead of being a stand-alone solution.
(f) Last but not the least, the investment required is always a barrier which needs to be taken into consideration. But with time, AI solutions will become more affordable so that the investment required, will justify with the benefits it will bring.
In your opinion, what are the key benefits AI can bring to healthcare, particularly in patient care and
what will be the outcomes?
AI can be adopted in healthcare in almost every aspect, starting with mundane and routine administrative tasks all the way to complicated Surgeries. So RPA (Robotic Process Automation) tools can help with some common administrative tasks, while Robot assisted surgeries allows you precise control and enable surgeons to perform complex surgeries, which were always thought to be very risky. AI can also be used for Precision Medicine, where starting with the early detection of Diagnosis, to treatment planning options and then all the way to Prognosis, it can be a great assistant to clinician. Use of AI in clinical research is already bearing fruit, and the use of AR / VR (Augmented and Virtual Reality) for Training and Education is already proven. With the advancement of Image and Speech recognition, AI has become a great tool for Physician to make their job easier and efficient.
In your view, how should we implement AI tools to enhance early disease detection, and how can we measure their effectiveness?
AI tools to enhance early disease detection should be a well thought exercise and one should not jump into making any quick decision. The reason for this is that most of the AI tools have been trained using patient data which does not represent Indian population (and it has been proven that South Asian are more prone for having Cardio related issues compared to their counterparts in the West). So it’s extremely important that we understand the Cohorts used and ensure that there is no bias in the AI tool, before implementing the same. On the same hand, there are some proven AI tools, especially when it relates to the Image recognition which can be a good asset to rely on, in the area of Radiology and Pathology for early disease detection.
Since AI is an evolving field, the lack of trained resources is a huge challenge and we have to invest in capacity building at all levels to bridge the gap, this include training the medical professionals starting with their college as well as IT resources.
How should we educate healthcare professionals about AI technologies, particularly in improving patient care? What strategies could foster a culture of AI adoption within the organization?
This is an important point which gets easily ignored. We need to have a good strategy to engage and educate healthcare professionals and staff, since this will reap continuing benefits. To foster the right understanding of AI technologies, we need to expose our healthcare professionals at early stages of their education, which will allow them to develop better appreciation of both the positive and not-so-positive perspective of AI application. And then when they grow into their career as a healthcare professional, will be able to take better decision – AI is here to stay so we have think from a long-term perspective.
For any organization to foster a culture of AI adoption, the best strategy is to identify cross-functional team members, which will have Clinicians, IT, as well as, Administrative department represented. Clinicians will be able to identify the problem area or issues and IT will have good understanding of which type of AI tool can possibly help, while administrators can provide inputs relating to the implementation and proposed workflow.
What is the current status and future role of AI in healthcare, and what strategies should be implemented to leverage AI for the organization's benefit?
Currently AI can easily be adopted for automation of administrative tasks and RPA (Robotic Process Automation) is proven tool, so should start with such low-hanging fruits. With such implementation the organization will start reaping immediate benefits and also provides them better understanding of AI in healthcare, both its potential as well as limitations. And to best leverage the potential, should have a very well thought out adoption of digital workflow, so that AI has all the required data for the patient being captured digitally.