AI, the doctor’s new assistant
Predicting heart attacks and diagnosing diabetes to performing precision surgeries, bots are changing the face of healthcare
Pushpa.Narayan@timesofindia.com 31.12.2024
Maheshwaran Rajendran doesn’t have the support system many elderly have – adult children – and couldn’t attend post-knee surgery rehab sessions at the hospital daily. So, doctors gave the 72-year-old an AI-powered app to help him exercise right in his bedroom. AI hasn’t replaced doctors, nurses, or paramedical staff, but it’s being increasingly adopted to improve treatment outcomes.
In some corporate hospitals, it’s a scribe for doctors and keeps a preliminary diagnosis and therapy options ready for approval. “The rehabilitation app is a perfect example,” says Dr G Balamurali of Kauvery Hospital. Many of the senior spine surgeon’s elderly patients skip rehab because they need help to travel to the hospital. “We now ask them to use an AI-powered app. They do the exercises we teach them in front of the phone camera and the app tells them if they are doing it correctly. Our rehab centre also keeps track of patients’ improvement.” People like Rajendran say hospital visits for rehab have reduced to once in two weeks compared to almost daily visits earlier.
“We have come a long way from telemedicine, when we could just see patients and offer consultations,” says Dr Balamurali. Over the past few years, hospitals such as Sankara Nethralaya have been using AI-powered equipment to screen eyes from remote and difficult-to-reach areas. High-risk patients are identified using AI tools and referred to ophthalmologists for treatment. “More accurate diagnosis, personalised care, and precision surgeries – AI and healthcare are proving they are made for each other, not just in lab experiments but also in hospital wards,” says Prof Krishnakumar R, department of medical sciences and technology, IIT Madras.
Over the past few weeks, the cath lab at Voluntary Health Services has been using AI algorithms to sharpen diagnosis. It has a full-time cardiologist to handle emergencies and staff trained by Dr Ajit Mullasari, director of cardiology for Madras Medical Mission, and his team. AI algorithms can quickly and accurately identify blockages and other abnormalities in angiograms, reducing the potential for human error and leading to more reliable diagnoses, says VHS secretary Dr S Suresh. Dr Mullasari says such tools can analyse vast amounts of data from electrocardiograms (ECGs) and other diagnostic tests to quickly identify heart attack patients.
“AI algorithms can also predict the risk of heart attacks by analysing subtle changes in medical images and patient data. This helps in early intervention and personalised treatment plans, potentially saving lives and reducing healthcare costs,” he says. While Dr Mullasari and his team are working with IIT-M on developing more AI tools for cardiology, scientists at IIT-M have been working on models that can diagnose diabetes before blood tests can show high sugar levels. “With AI, we have the chance to finally reverse the trend lines on chronic diseases,” says senior diabetologist Dr V Mohan. He admits it will take time to overcome hurdles such as patient privacy and convince policymakers to create a regulatory environment, “but it can certainly revolutionise treatment”. Labs in the city are already showing promise. Prof Krishnakumar of IIT-M, who with his team is researching the subject, says they can create a digital twin of a patient with all medical records in 45 minutes at their lab. “One hour after creating the synthetic patient on the computer, we can test at least five different types of treatment options based on these models. And nearly eight out of ten times, the outcomes we predict have worked.” Senior cardiac surgeons including Dr K R Balakrishnan have been working with the team to improve treatment outcomes in cardiac surgeries.
While organisations such as WHO are asking public health officials to be ready for more frequent outbreaks of infectious diseases, public health officials are hoping they can turn to AI for help.
.jpg)
.jpg)
.jpg)
