The shortage of doctors, especially in rural areas, is one of the biggest challenges before India’s healthcare sector. There is one doctor for every 1,511 Indians, but this is only an average. Most doctors are based in cities, so the doctor-patient ratio in India’s villages is far below the WHO norm of 1:1,000. Could artificial intelligence (AI) – which is being hailed as the next big frontier in healthcare globally – help solve this crisis? Dr Prateek Sharma, a renowned gastroenterologist who is the first chair of the American Society for Gastrointestinal Endoscopy (ASGE) AI Task Force, explains in a talk with Durgesh Nandan Jha
AI seems like a technology for the future, is it really here today?
Yes, it is here and all of us are already using it in our daily life. Several apps and websites, such as Instagram, Gmail, Amazon and others, are AI-powered. For example, if you share an image on an app, AI will help predict the location. If you start typing an email, the machine will prompt you to complete the sentence based on the phrases you commonly use. If you have searched for a product on Amazon, the next time you log in, it will take you to the same product. All this is being done today, and by AI.
How does AI work in medicine? Are there real-world examples?
Traditionally, the diagnosis and treatment of diseases has been based on patient-physician interaction. When a patient visits a physician with a particular set of symptoms, the physician works at a differential diagnosis, orders specific tests, and, based on the test results, treatment is designed.
Artificial intelligence predicts the outcome of a diagnostic or prognostic process by using its ‘experience’ that’s based on the data of previous patients used to train the AI system. It may be considered as a super calculator that reduces our error rate.
AI has shown benefits in a whole range of specialties. In primary care, AI studies have been conducted to help monitor patients’ blood pressure and glucose levels to suggest early changes in medications, including blood pressure medications and insulin doses that allow better control of patients’ hypertension and diabetes. In ophthalmology, automated image analysis has helped in the detection of eye conditions without the need for a specialist in remote areas. In the field of gastroenterology, AI has helped in the detection of polyps during colonoscopy and the detection of cancer in the GI tract.
Which existing and future specialties might benefit the most from the introduction of AI?
The field of radiology has seen the growth and adoption of artificial intelligence. AI can analyse CT and MRI images, and help detect subtle changes. Fields like cardiology and ophthalmology have also seen the growth of AI in helping detect and diagnose diseases. In GI (gastroenterology, or the medical specialty dealing with the stomach and the intestines), AI devices have been approved that allow the detection and characterisation of precancerous polyps that can be removed to prevent colon cancer.
AI can extract information from natural language processing, helping detect patterns and information that would otherwise be hidden in electronic data. The future of AI would involve all healthcare fields, including quality clinical care, early diagnosis and treatment, precision care medicine at a patient level, and mass-level care at a population level.
One of the potential benefits of AI is in improving healthcare access in remote and resource-challenged areas. What needs to be done to realise this potential?
AI has the considerable benefit of improving healthcare across the globe, and not only for resource-rich countries. Theoretically, AI may be useful when there are no human resources. Examples are pathology and ophthalmology. For this to be successful, there should be equitable distribution of AI technology. Open access algorithms, subsidies for AI technology, education and awareness among physicians and patients, and integration of the technology with local healthcare delivery models will be critical in AI’s successful adoption and utilisation.
How have other countries deployed AI in healthcare?
AI has shown a great deal of promise in preclinical as well as clinical areas of medicine. In the preclinical areas, AI has helped researchers identify multiple pathways by which diseases are caused and thus identify potential targets to help discover drugs.
In gastroenterology, for example, AI devices approved in the US, Europe, and Asia have improved the detection and differentiation of colon polyps and cancer. Numerous pathology-based services are currently run using artificial intelligence technology that has helped pathologists identify subtle differences on slides that help diagnose diseases better.
Are you saying that in the future AI can replace doctors? What are the potential pitfalls of dependence on such a system?
Not at all. At this time, the majority of AI medical devices and techniques are ‘assist devices’ – they help the doctor make a diagnosis or decide on a treatment. They are not replacing doctors. They can help doctors become more efficient, reduce costs, reduce unnecessary testing and make processes more automated. The ultimate decision is still the doctor’s, not the machine’s.
AI seems like a technology for the future, is it really here today?
Yes, it is here and all of us are already using it in our daily life. Several apps and websites, such as Instagram, Gmail, Amazon and others, are AI-powered. For example, if you share an image on an app, AI will help predict the location. If you start typing an email, the machine will prompt you to complete the sentence based on the phrases you commonly use. If you have searched for a product on Amazon, the next time you log in, it will take you to the same product. All this is being done today, and by AI.
How does AI work in medicine? Are there real-world examples?
Traditionally, the diagnosis and treatment of diseases has been based on patient-physician interaction. When a patient visits a physician with a particular set of symptoms, the physician works at a differential diagnosis, orders specific tests, and, based on the test results, treatment is designed.
Artificial intelligence predicts the outcome of a diagnostic or prognostic process by using its ‘experience’ that’s based on the data of previous patients used to train the AI system. It may be considered as a super calculator that reduces our error rate.
AI has shown benefits in a whole range of specialties. In primary care, AI studies have been conducted to help monitor patients’ blood pressure and glucose levels to suggest early changes in medications, including blood pressure medications and insulin doses that allow better control of patients’ hypertension and diabetes. In ophthalmology, automated image analysis has helped in the detection of eye conditions without the need for a specialist in remote areas. In the field of gastroenterology, AI has helped in the detection of polyps during colonoscopy and the detection of cancer in the GI tract.
Which existing and future specialties might benefit the most from the introduction of AI?
The field of radiology has seen the growth and adoption of artificial intelligence. AI can analyse CT and MRI images, and help detect subtle changes. Fields like cardiology and ophthalmology have also seen the growth of AI in helping detect and diagnose diseases. In GI (gastroenterology, or the medical specialty dealing with the stomach and the intestines), AI devices have been approved that allow the detection and characterisation of precancerous polyps that can be removed to prevent colon cancer.
AI can extract information from natural language processing, helping detect patterns and information that would otherwise be hidden in electronic data. The future of AI would involve all healthcare fields, including quality clinical care, early diagnosis and treatment, precision care medicine at a patient level, and mass-level care at a population level.
One of the potential benefits of AI is in improving healthcare access in remote and resource-challenged areas. What needs to be done to realise this potential?
AI has the considerable benefit of improving healthcare across the globe, and not only for resource-rich countries. Theoretically, AI may be useful when there are no human resources. Examples are pathology and ophthalmology. For this to be successful, there should be equitable distribution of AI technology. Open access algorithms, subsidies for AI technology, education and awareness among physicians and patients, and integration of the technology with local healthcare delivery models will be critical in AI’s successful adoption and utilisation.
How have other countries deployed AI in healthcare?
AI has shown a great deal of promise in preclinical as well as clinical areas of medicine. In the preclinical areas, AI has helped researchers identify multiple pathways by which diseases are caused and thus identify potential targets to help discover drugs.
In gastroenterology, for example, AI devices approved in the US, Europe, and Asia have improved the detection and differentiation of colon polyps and cancer. Numerous pathology-based services are currently run using artificial intelligence technology that has helped pathologists identify subtle differences on slides that help diagnose diseases better.
Are you saying that in the future AI can replace doctors? What are the potential pitfalls of dependence on such a system?
Not at all. At this time, the majority of AI medical devices and techniques are ‘assist devices’ – they help the doctor make a diagnosis or decide on a treatment. They are not replacing doctors. They can help doctors become more efficient, reduce costs, reduce unnecessary testing and make processes more automated. The ultimate decision is still the doctor’s, not the machine’s.