Laparoscopic and General Surgeon, formerly the Director, AIIMS from 2013 -2017 and ex-AIIMS, Head of the Department, General Surgery foresaw the need for specialized trauma care early on and established the Apex Trauma Centre at AIIMS.
Q. How can we see the role of digitalization in the medical and laboratory business, and how can hospitals leverage this digital revolution’s potential?
My journey in healthcare dates back to an era where the concept of digitalization was unthinkable. Back then, everything was manual, from documenting patient profiles to scribbling emergency notes by hand. It was evident that many essential details were often missed or undocumented. Then, in 1981, we were introduced to ultrasound technology, followed by the advent of CT scans in 1988. These were significant steps towards digitalization, but progress was slow. Today, no hospital can function without digital technology. It benefits both healthcare professionals and patients. Telemedicine, which was once a concept we talked about, was judiciously extensive in use during the COVID-19 pandemic.
The digital age has also introduced us to Artificial Intelligence (AI), although we’ve just started on the surface of its great potential. We currently use AI for tasks like appointment scheduling, automated patient responses, and sharing lab reports digitally.
Q. This technology, which was once a mere dream of the IT industry, has now become widespread. I’d like to hear your thoughts on how artificial intelligence offers advantages over traditional analytics and clinical decision-making techniques.
Artificial intelligence fundamentally relies on algorithms designed to process and categorize data effectively. For instance, consider a scenario where we have data on 1000 patients with fever. AI can help us categorize the most common and less common causes of fever, assisting in differential diagnosis. In the healthcare context, AI is finding utility in various areas, particularly in medical imaging and reporting. By using algorithms, AI can recognize abnormalities in X-rays or MRIs, differentiating between conditions such as cancer and inflammation. While AI shows great promise, its current scope is somewhat limited.
However, areas like robotics, which also fall under the AI umbrella, are gaining traction, with applications in remote surgery. In the healthcare sector, AI will inevitably play a substantial role across disciplines, including imaging, pathology, clinical medicine, and research.
Q. Given the various potential applications of AI in healthcare that you’ve mentioned, what, in your view, is the most promising application?
The potential applications of AI in healthcare are vast and hold immense promise. While some applications are already in use, such as AI assisting in imaging and pathological reporting, the true potential of AI is yet to be fully realised. AI is expected to shine in areas like reporting CT scans and MRIs, as well as aiding in pathological diagnoses. Clinical medicine also stands to benefit from AI, although this field is still in early stages in terms of AI integration. The key challenge I would say lies in accumulating vast datasets, and the geographic prevalence of certain diseases means that algorithms will need to be customised to specific regions. In India, for example, tuberculosis is a common concern, while it is rare in the USA. As a result, AI’s role in clinical medicine is just beginning to take shape, and time will reveal where it will be most impactful and where its limitations lie.
Q. How do you view the concept of preventive and personalized care for patients?
Indeed, personalised medicine is an exciting avenue, but we’re still at the early stages of understanding how individual variations can impact treatment outcomes. For instance, what works for one person may not be suitable for another due to differences in genetics and response to medication. Even the same medication can yield different results in different individuals. So, while the idea of personalized medicine is promising, we have a long way to go. We need extensive data to predict how an individual’s genetic makeup may influence their response to specific medications accurately.
Q. Could you provide some examples of AI-managed activities within the healthcare sector?
In the realm of healthcare, one significant area where AI is making a substantial impact is medical imaging. This includes tasks like interpreting X-rays, CT scans, and even ultrasound images. AI algorithms are being employed to assist in diagnosing medical conditions.
For instance, let me share an intriguing case. We had a patient, a woman, who came in with a brief history of abdominal pain. Surprisingly, her vital parameters appeared normal, and she didn’t seem severely ill based on our clinical judgement. She complained of pain and difficulty passing stool and gas, which raised concerns about a potential bowel obstruction. To investigate further, we conducted a CT scan. However, the CT scan failed to detect the multiple perforations she had in her intestines, which was quite an unusual situation. This incident highlights the limitations of traditional imaging methods. To overcome such limitations, we need to accumulate extensive patient data, which can then be used to develop AI algorithms.
As I mentioned earlier, AI’s role in clinical medicine is significant, although it does have its limitations. Not all aspects of healthcare can be solely managed by AI or robotics. Take robotic surgery, for example. While robots can aid in precision, they cannot replace the expertise and decision-making abilities of human surgeons. Nevertheless, they are invaluable in procedures like arthroplasty and knee replacements, offering precise incision planning and execution. Robotics, guided by AI, provide enhanced magnification, precision, and instrument maneuverability, granting surgeons greater flexibility. Additionally, let me illustrate with the example of the gamma knife used in brain surgery. It enables precise targeting of specific areas in the brain, aided by meticulous planning. AI plays a pivotal role in these advancements.
Q. Presently, we are actively engaged in 3D image reconstruction and bioprinting. You mentioned the significance of identifying features from images earlier. We are currently generating 3D models from DICOM data, which include CT, MRI, and PET scans. For instance, these models are highly beneficial in brain tumor segmentation and cranial implant generation. Could you elaborate on other potential areas where this technology can be advantageous?
Ans: Certainly, another area where this technology holds immense promise is in liver resection surgery. Much like the brain, the liver is a vital organ, and when dealing with liver tumors, precision in planning is crucial. In the case of brain surgery, preserving healthy brain tissue is paramount, given the importance of every neuron. The use of 3D printing and planning can significantly aid in ensuring that only the diseased part is removed, leaving the healthy tissue untouched.
Robotic surgery also benefits from 3D modelling. It provides surgeons with three-dimensional depth perception, a crucial aspect when performing intricate procedures. This depth perception enhances precision and can be especially advantageous during minimally invasive surgeries. Traditional laparoscopy, which offers only a two-dimensional view, can’t match the level of precision achievable with 3D modelling and robotic assistance. Therefore, this technology has had a transformative impact across various surgical disciplines, ranging from general surgery to orthopedics.
Q. How do you see the three decade journey of technological evolution?
The evolution of technology in the medical field has been remarkable over the last three decades. Video technology has played a pivotal role, progressing from the first generation of endo cameras to triple-chip cameras with enhanced colour precision. Nowadays, we utilise high-definition cameras and even 4K technology, offering unparalleled clarity during surgical procedures. It’s essential to recognize that AI and digital technology have been driving these developments.
Moreover, technological advancements are not limited to surgery alone. They have permeated all aspects of medicine, including diagnostic tools. For example, consider the use of ultrasound technology. It has evolved significantly, with newer energy sources that allow ultrasound waves to coagulate and ligate vessels, eliminating the need for sutures. Vessel sealing systems have also revolutionised surgical procedures, enhancing safety and efficiency. Even in seemingly routine procedures like dental treatments, sensors are now employed to prevent nerve damage during root canals, ensuring a painless experience for patients. In essence, AI-related components are being integrated into various medical fields, enhancing the quality of healthcare across the board.
Q. Another application of AI in healthcare is disease screening and classification. It appears that AI models are being developed for conditions like diabetic retinopathy classification, determining whether a patient falls into the high-risk category requiring immediate intervention. We also discussed the use of 3D models for retinal detachment. Could you shed more light on the potential of AI technology in the healthcare sector?
AI’s ability to create three-dimensional models, as we discussed earlier, has transformative potential. These models can be applied not only to the brain and liver but also to other parts of the body. Imagine being able to visualize a tumor or abnormality in three dimensions, allowing for precise planning and targeted interventions. This level of precision and depth perception is invaluable in healthcare, as it can lead to better patient outcomes.
In the case of trauma patients, such as those involved in road accidents, AI can play a crucial role in assessing injuries. For instance, if someone has experienced a fall or an accident, a whole-body CT scan can be performed to determine the extent of internal injuries. AI algorithms can assist in swiftly identifying and assessing these injuries, enabling healthcare providers to make informed decisions and provide timely care.
Furthermore, the potential applications of AI in healthcare extend beyond diagnostics and treatment planning. Consider the vast amounts of data generated in healthcare settings, from patient records to medical imaging. AI can help manage and analyze this data efficiently, leading to better patient care and more effective resource allocation.
To fully harness the power of AI in healthcare, we must continue to build robust datasets and develop sophisticated algorithms. This requires collaboration between healthcare professionals, researchers, and technology experts. By doing so, we can unlock AI’s full potential and revolutionize healthcare delivery, offering patients more precise, accessible, and effective care.
Q. How is integrating AI beneficial for academics and research in the healthcare sector?
Interesting question. In academia, AI can enhance medical education. Tools like task-sensitive tabletop displays with full-scale 3D human body graphics can be used to teach anatomy effectively. These tools provide a detailed understanding of anatomy without the need for cadaveric dissection by the students because many times they don’t get medical colleges and do not get enough cadaverics to work on. The scope is far and wide and it is up to us.
Q. When we talk about this huge amount of health data that is then required. So how does AI ensure privacy and security of my health data? And what are your thoughts on it?
The issue of privacy is indeed crucial, particularly when dealing with health data. When data is anonymized, with identities concealed, there’s a layer of confidentiality maintained. Only authorized personnel with specific access rights can retrieve this live data. For example, if I reside in Delhi but fall ill in Mumbai, only those granted access rights, such as healthcare providers involved in my treatment, can view my data. It’s akin to a Hospital Information System (HIS) where not everyone can log in for every patient; specific patient-related login credentials are necessary. Preserving confidentiality is not that much of a problem.
However, it’s important to note that there are continuous checks and balances in place. Breaches of confidentiality, like data breaches or hacking incidents, do occur, as exemplified by a recent hacking incident involving the AIIMS website and we were off the internet for four months. The same cybersecurity risks apply to health data. So, preserving confidentiality remains a challenge that necessitates ongoing attention.
Q. What are the limitations of AI technologies in the healthcare sector?
These AI algorithms, essentially data-driven, have the potential to aid in research, analytics, medical diagnosis, and treatment planning. It’s important to note that AI won’t replace human interaction, especially in clinical medicine. Ultimately, AI is a product of human design and will be a valuable tool to complement human expertise. The next five to ten years will likely reveal how AI continues to evolve and integrate into healthcare.
Despite the advanced imaging technologies available today, such as ultrasound and CT scans, there remain cases where the exact diagnosis remains elusive. In these situations, pathological testing or direct visualization through procedures like endoscopy remains necessary. Thus, technology has its limitations, and it can only assist us up to a certain point. Notably, in cases like Dr. Deepak Agarwal’s work with the Gamma Knife, AI can help pinpoint the precise location of brain tumors for radiation treatment planning. However, the ultimate decision-making and planning still rest in human hands.
Q. Can you shed light on other challenges that healthcare institutions face when integrating AI into their practices? How can these challenges be addressed or overcome?
Good question. Presently, our healthcare system operates in isolation, with individual hospitals and healthcare providers working independently, even within the public and private sectors. This lack of integration poses significant challenges. We lack comprehensive, consolidated data sets, making it impossible to track procedures across India. For example, how many gallbladder operations are done all across India, every year, or every month or every week. Unlike the USA, where such data is readily available, we struggle to access this information because of the lack of integration. It is mandatory to report the death. And that’s why we have data on that. But we do not have data on various medical or surgical illnesses, cancer or cardiovascular diseases.
The National Digital Health Mission announced by the Prime Minister is one step towards establishing a common platform for comprehensive data collection. This initiative could help replicate the successes we’ve seen in managing diseases like COVID. To do so, we must make it mandatory to report all patient data from every hospital in the country. Whether it’s Ambani Hospital in Mumbai or Fortis Hospital in Delhi, we need a standardized approach for collecting essential patient data, including age, gender, final diagnosis, and outcome.
Q. It seems interoperability is a key issue to address. Do you believe there is a gap in access to AI- powered healthcare, particularly in underserved communities? If so, how can we bridge this gap?
Indeed, there exists a gap in access to healthcare and affordability, which is not unique to healthcare but extends to other professions as well, like legal services. Addressing these issues, especially in underserved communities, is a critical concern. Initiatives like Ayushman Bharat aim to bridge this gap by offering healthcare benefits to those with limited incomes.
One significant aspect to address is the lack of information among underserved communities. Some individuals may lack the resources to travel to distant healthcare facilities, even within their districts. However, this can be mitigated by establishing connectivity between primary health centres, community health centres, and district hospitals.
ASHA workers and other healthcare personnel at the grassroots level can play a pivotal role by identifying healthcare needs and facilitating telemedicine consultations when necessary. Creating such a network can significantly improve accessibility to healthcare in rural areas.
Furthermore, timely data sharing and telemedicine can aid in trauma cases, ensuring patients reach healthcare facilities within the “Golden Hour,” critical for trauma victims. While challenges remain, with dedicated efforts, we can work towards an integrated healthcare system, improving healthcare access and affordability in underserved communities. This is essential for India’s development goals, aiming for a developed country status by 2047 or beyond.
Q. So I would also then, like to understand from you how technology innovation hubs, like ours, DRISHTI CPS Foundation here at IIT Indore can put innovative digital health initiatives into action?
Technology Innovation Hubs like the IITI DRISHTI CPS Foundation at IIT Indore have a crucial role to play. They can identify local challenges and collaborate with nearby healthcare institutions, particularly in underserved areas. These hubs can empower healthcare workers, such as ASHA workers, to collect essential health data, conduct basic health assessments, and enable telemedicine consultations with higher-level healthcare facilities. Additionally, these hubs can facilitate data transmission, even for educational purposes, promoting awareness and healthcare knowledge dissemination. The integration of healthcare systems is a complex task, but with dedicated efforts, we can work towards this goal, ensuring better patient care, accessibility, and affordability.