Speaking at THE WEEK Health Summit 2024, Siddharth Bagga, Head (Retail, CPG and Healthcare), Google Cloud, said that while AI was set to make a 10x impact on health care, there are four major areas where it will be deeply felt:
1) Patient consumer engagement will see AI models that help with first-line triage and chronic condition management.
2) Clinical and care-team support, where AI models will minimise written tasks undertaken by clinicians (answering emails, chats, writing up case summaries and treatment plans.) It will also step in to assist with analytical tasks, like detecting scan abnormalities. The third bit would be assisting administrative duties, like a clinician handing off a patient to a nurse.
3) Longitudinal health, which needs an electronic medical record of patient health information generated over time by multiple hospitals. AI will analyse diverse records like X-ray images, scans, and clinical notes to generate a longitudinal view of health.
4) Digitisation of personal health records, instead of them being in bundles of paper stuffed into a file.
ALSO READ: 'AI helping diagnose diseases faster': AIIMS director Dr M. Srinivas
To illustrate functional examples of how AI is changing health care, Bagga mentioned a tuberculosis (TB) diagnosis project that Google and Apollo Hospitals are collaborating on. Briefly put, the project sends vans fitted with X-ray imaging devices to rural areas. X-rays taken of potential TB patients are beamed to a central facility, where AI screens the images and gives a diagnosis.
Instead of doctors physically travelling to villages to screen patients, a combination of AI and telemedicine could generate an early diagnosis. Patients who are positive for TB could then be provided appropriate care at a secondary- or tertiary-care facility.
Another example he provided was of an ePharmacy project with Manipal Hospitals, where any handwritten prescription issued by a doctor could be transcribed into a digital prescription. “And, it would place it in context,” Bagga said, emphasising that Google’s Gemini 1.5 AI model was not doing mere transcription.
He also said that Google was trying to map the human brain to understand mental health and related issues. While that will take an excessive amount of time, the first step was to map the brain of a mouse—which Google successfully did with “AI-based analysis and visualisation”.
Now, will AI lead to the loss of jobs in the health care industry? Bagga’s answer was clear: Will it replace doctors? No. Will it replace nurses? No. Will we need people who understand technology and medicine to make this work? Yes.
The hint was that while repetitive tasks might disappear, newer avenues will open. He also wanted patients to make the boundaries very clear as to what they want from AI. “Do not depend on AI for a diagnosis. Go to a doctor for that,” he said.
