From dialysis to diagnosis: AI enters nephrology, questions follow

Bengaluru: As Artificial Intelligence begins assisting treatments such as dialysis in government hospitals, questions emerge about how far these systems can align with clinical judgement and patient care.

Doctors broadly agree that AI is a useful tool in treatment and diagnosis. Yet, it also raises ethical dilemmas.

AI can play a major role in detecting chronic kidney disease (CKD) faster than traditional tests.

Speaking to DH, Dr Topoti Mukherjee, Lead Consultant, Nephrology and Kidney Transplant at a private hospital, said, "Some of the recent studies indicate that deep learning algorithms can identify CKD risk factors as much as 6 to 12 months prior compared with clinical diagnosis."

"In addition, Artificial Intelligence imaging using convolutional neural networks for renal scan analysis has been able to detect abnormalities in kidney structure that are not visible via traditional imaging methods," she said.

This shift from reactive diagnosis to predictive risk stratification is particularly valuable in CKD, which is often asymptomatic in its early stages.

Addressing the conflict between AI assessment and clinical assessment, Dr Arun Kumar J, Consultant Nephrologist, said, "The conflict does happen and there are times when AI may not quite fit the clinical situation. Sometimes AI is picking up a trend that is not obvious at first glance."

"At the end of the day, the responsibility sits with the clinician. So, the final call has to come from us. AI can inform the decision, but it does not make it," he added.

Role in dialysis

AI is transforming dialysis by enabling more precise, personalised and safer treatment. It continuously monitors patient data such as blood pressure, fluid levels and toxin removal, predicts complications, and supports doctors in adjusting care.

This improves outcomes and reduces emergencies. It also lowers manual workload and enhances efficiency in dialysis centres, said Dr Kishan A, Consultant, Nephrology.

On the limitations of AI, he said, "However, its effectiveness depends on data quality, and it may miss sudden or rare complications. High costs, data privacy concerns, and the need for staff training remain challenges. AI cannot replace clinical judgement and requires constant human oversight to ensure safe, reliable and equitable dialysis care."

Ethical dilemma

Dr Anoop M Gowda MD, DNB (Nephrology), Senior Consultant Nephrologist and Transplant Physician, said, "Using AI in healthcare raises ethical and practical concerns. Patient care largely relies on human interaction, which technology cannot replace. Questions of accountability remain unclear if outcomes are poor — whether responsibility lies with the doctor, developer, or system."

"AI also depends on costly infrastructure and data often drawn from Western populations, limiting relevance in India. While promising, AI must be used cautiously alongside clinical judgement, with doctors ultimately guiding decisions and intervening, when necessary," he said.