AI: Rx For Medical Professionals
In January 2020, the World Economic Forum said artificial intelligence (AI) will change healthcare by 2030. It provided three ways this would happen: predictive care, connected care and better patient and staff experience. That prediction requires a drastic revision.
AI is already changing medicine, and we have barely gotten started with 2023! While AI in the form of ChatGPT (Generative Pretrained Transformer) is shaking up industry after industry, it is the play of AI in medicine we will welcome the most. Because saving lives is more important than owning autonomous cars or using AI to cheat on your high school essay.
Incredible as this may sound, listen to what the World Health Organisation is saying: there is a one in a million chance of a person travelling by air being harmed, but a one in 300 chance of being harmed during healthcare. 42.7 million adverse events occur when patients are hospitalised, making it the 14th leading cause of the global disease burden.
Anything can cause this harm, ranging from inadequate patient data to a lack of experience on the doctor’s part, administrative errors, staff shortages, and poor working conditions. But now, AI can step in and fill in the gaps. For example, AI can look up data from millions of patients for a condition, quickly flag missing data, and eliminate administrative errors. Best of all, it can help doctors reduce medical errors by using the experience of thousands of other doctors whose decisions are meticulously recorded and preserved in medical databases.
This reminds me of an episode of New Amsterdam on Netflix (season 2, episode 14) which is about doctors at a hospital in New York addressing the business, medical, social, ethical, regulatory, administrative, and personal challenges their patients pose. Dr Vijay Kapoor, played by Anupam Kher, is given a device called DAWN (diagnostic assessment wellness network) that listens to him, parses, and then analyzes vast amounts of medical data to reach an accurate diagnosis.
But DAWN fails to diagnose a critical patient’s condition, while Dr Kapoor—deeply suspicious of the AI—throws up his hands and said, “This is not artificial intelligence; it is my intelligence!” Finally, Dr Kapoor figures out what is wrong with the patient when listening to the patient’s daughter narrate an innocuous, seemingly unrelated event. An AI could never have done this. New Amsterdam is worth watching – because it tells us why AI will never replace doctors. But it also makes an excellent case for AI that can become an adjunct to a system that is overworked, stressed, and fighting against odds to save precious lives.
AI can run right through the lifecycle of a patient, cutting down the time to diagnosis, reducing errors, and pointing to possible treatment regimes. Let’s consider the patient’s lifecycle and where AI can create interventions:
Friendly, voice-based chatbots that analyze symptoms and check on cures to treat the illness. Many, like Buoy, developed by a team at the Harvard Medical School, use data, ML and algorithms with differing capabilities to diagnose health conditions and recommend what to do next. The advantage is that this type of AI can run in any language, 24X7, and scale without adding to costs.
Faster hospital admissions can significantly impact the patient experience. The Johns Hopkins Hospital uses AI to manage operational activities. Patients are being assigned beds 38 per cent faster, ensuring they are not kept waiting. The efficiency has also contributed an estimated USD 16 million in annual revenue by opening 16 beds daily without adding to staff strength or requiring a new wing.
Helping medical professionals with an accurate diagnosis by analysing CT scans, MRIs, X-rays, ultrasounds, retinal images, etc., with Arterys being one example. This capability has a dual advantage. On the one hand, it improves diagnostic accuracy and speed; on the other, it can reduce cost because AI analysis can be done in low-cost facilities.
Managing insurance payments is always a painful process for patients. AI can help automate the processes by assembling the required documentation—from patient data to treatment details, medical forms, and other insurance requirements to make payments painless and fast.
Post-care tracking is an essential part of modern medicine. After being discharged from the hospital, proper care ensures that patients need not be readmitted or visit doctors. Patients, for example, may be prone to falling after knee surgery or a hip transplant, or a cardiac procedure. Can their movements be tracked remotely? One example is Virtusense, a company that senses when a patient is about to get up, analyzes balance, and predicts the risk of falling for the next 12 months.
AI can improve care but equally enhance the experience and lower costs. Technology may never match the skills of a trained professional, but it will help doctors in making better decisions and save millions of lives.
