A 33-year-old man recently visited a doctor after multiple hospital and specialist consultations, during which he had spent several days admitted, and had incurred expenses running into lakhs of rupees. He had been experiencing high-grade, intermittent fever for three months, accompanied by rashes and body pain. Extensive investigations had been conducted to rule out infections, cancers, and autoimmune diseases. The only notable findings in his tests were elevated ESR and CRP—markers of inflammation—and a high serum ferritin level, a test typically used to assess iron stores.
When the tech-savvy doctor fed the patient’s symptoms and lab reports into an AI enabled EMR (Electronic Medical Record) the first diagnosis that the system suggested was Adult Onset Stills Disease which is an auto- inflammatory condition. The patient was then treated with steroids and immune-modulating drugs, which provided a prompt resolution of symptoms. What several specialists missed over almost three months, the AI EMR picked up in seconds, at virtually no cost.
Welcome to the exciting world of Artificial Intelligence-based Electronic Medical Records.
What is an AI-based EMR and what can it do that conventional medical record systems and even many doctors cannot?
Optimising diagnoses
An AI EMR does the thinking for the doctor/nurse to ensure that an optimum diagnosis is made, and the ideal treatment options are provided. It ensures that every lab result is factored in, every datapoint considered, when arriving at its recommendations. (It can do the thinking for you, the patient too, but that needs accurate and comprehensive information about your condition to be made available to the system.)
Hundreds of studies now show AI models outperforming physicians in real-life settings, medical MCQ examinations and even in the most complex of case scenarios. Yet, adoption faces huge roadblocks put up by a healthcare industry protecting its turf and livelihood.
Opportunity for India
AI today, offers a massive opportunity for countries like India to transform the quality of healthcare that the poor and underserved can receive. It is inexpensive, phenomenally accurate already (and will keep getting even better) and most importantly, easily accessible. In short, it can largely solve the three major problems healthcare faces: accessibility, affordability and quality.
AI to many means by default, Generative AI and the Large Language Models (LLMs) that have been created by the likes of OpenAI, Google and Anthropic. So the erstwhile, ‘Have you Googled your symptoms?’, is now, ‘Have you checked your symptoms on ChatGPT?’ But AI has a lot more to offer for healthcare than just Gen AI. And an AI based EMR is at the forefront of this revolution.
Combating existing challenges with AI EMRs
An AI EMR can ensure creation of a unified health record for a patient easily. Currently, this is a massive challenge due to patients seeing different doctors over time, going to different labs etc., and all this data remaining in disparate compartments, affecting the quality of care received. Data and unstructured data at that, can now be instantly tagged to a patient’s medical record and be extracted and made available for analysis. All that the patient needs to do is upload a PDF document of a lab report or take a photo using his mobile phone and upload the jpg images to the medical record. So, patients themselves can create their unified patient records and have them analysed using Gen AI within AI EMR systems. Though some of these technologies have existed before, their integration with AI has transformed them.
AI EMR systems can answer in simple, patient, understandable language, every doubt that the patient may have and the doctor may not have time for. For example, how essential is the surgery my doctor ordered, what are the alternative options, what are the risks if I do not go in for it etc. are standard doubts. Earlier, you could get generic information. Now, you get accurate answers based on your medical data, your symptoms, your lab reports and condition, even your economic constraints. In short, an AI EMR or PHR (Personal Health Record) can be your regular health assistant, one that is far more knowledgeable than an individual doctor could ever hope to be. The big difference from earlier is in the specificity of the answers tailored to the individual’s condition.
AI EMR systems can take language out of the equation. Today, AI EMR’s can support voice transcription of conversations between the doctor and the patient in any major language (and multiple languages too), extract the medical component of the data and even provide patient summaries in any major Indian language. Some EMR systems have patients talk to an AI assistant on their phone in their native language and it asks relevant questions based on symptoms and suggests the next steps or summarises the details for the doctor with great accuracy. Many of the technologies used in these processes are not AI (speech-to-text for instance, has been around for decades) but the integration of AI has given these technologies phenomenal power.
AI interpretation of medical images within EMR systems has greatly improved. AI medical record systems today can read X-rays, CT scan images and the like, with a great deal of accuracy, in many documented studies, outperforming radiologists. As this technology gets even better, the reliance on the local doctor to interpret a report will decrease.
AI EMRs also drastically reduce medical errors in hospitals and clinics. Doctors may be overworked and may forget critical data points that end up costing lives. AI-based models do not forget, do not get tired and are always available. As a patient safety tool, every hospital can benefit from using AI.
These are just a few examples of how AI EMRs can transform healthcare delivery.
Also Read: Benchmarks in medicine: the promise and pitfalls of evaluating AI tools with mismatched yardsticks
Roadblocks to implementation
AI is not 100% accurate and can make errors is the obvious caveat, but this should not be the deterrent to using it. The only question that matters is whether AI is better than existing healthcare providers at what it does. There is now overwhelming global evidence that it is significantly better.
So, what is preventing the healthcare community from embracing AI EMRs? Could it be a perceived threat to livelihoods and the ‘what’s in it for me’ attitude? Governments have a once in a millennium opportunity to transform healthcare delivery in a country like India through the use of AI EMRs but it is unlikely that they will antagonise the medical community and push for it in a big way.
Eventually the use of AI EMR/PHR systems in healthcare will likely be patient driven. As more and more patients find out that there is an accurate diagnosis and treatment option available for a large number of medical conditions and that too at almost negligible cost, they will start to embrace it. Maybe that is where the opportunity to transform healthcare lies.
(Dr. Sumanth C. Raman is a consultant in internal medicine and founder of Algorithm Health, an AI company in healthcare. He writes on healthcare issues. [email protected])
Published – July 03, 2025 03:36 pm IST