A health care framework for patient-centric use of AI

Hindustan Times

07-06-2025

Artificial Intelligence (AI) technologies are developing at a much faster pace than regulation or industry standards can keep up. With every new breakthrough and innovation, the potential of AI to transform the health care sector increases–whether through screening and diagnostics, public health and complex data analysis, or clinical decision-making. Generative AI (GAI) in particular has the potential to positively impact improve India's healthcare system by supporting clinical care, that is, providing clinical decision support and deepening clinical expertise at the point of care by processing vast amounts of data and knowledge. Ultimately, this leads to better informed decision making and improved health outcomes.
Use of AI in sensitive fields like health care is understandably mired in misconceptions and practical challenges. At the outset, it's essential to recognise AI should be seen as a tool to augment, not replace, human expertise the role of AI as a tool that supports the skills of doctor's clinicians, it's equally important to recognise that adoption of AI technologies raises valid concerns about operational efficiency, patient data privacy and algorithmic bias, in addition to broader ethical and regulatory challenges.
As AI develops through newer, relevant use-cases for health care, the need of the hour for India is to proactively build a self-reliant framework and roadmap for ethical and responsible use of AI. This can not only ease adoption of emerging AI technologies across India's private and public health care, but can also help address systemic challenges in health care systems. A clear roadmap of AI integration with guardrails can help both India's clinicians and patients benefit from the immense potential of modern AI.
In a field as dynamic and ever-evolving as health care, GAI enables critical support to clinicians by creating real-time learning opportunities. It facilitates differential diagnosis especially in complex cases with co-morbidities; by driving sharp insights from large databases to help clinicians and nurses take better-informed calls. GAI platforms trained on Large Language Models (LLMs, clinicians can be understood as a tailored assistant that can answer targeted queries at the point of care and pull up relevant answers efficiently and quickly, allowing clinicians to perform their tasks better by enabling easier access to pertinent data and information. When employing differential diagnosis for patients, GAI solutions based on latest evidence can function as a concentrated pool of knowledge for clinicians to rely on.
Modern health care systems are built on technological foundations, from data management, diagnostics to surgical support. AI is the natural next step to increase productivity and accuracy in healthcare by augmenting the skills of medical professionals. GAI can rapidly analyse and synthesise large volumes of medical literature to provide clinicians with the most relevant, reliable evidence to make clinical decisions at the point of care. One of AI's key offerings is personalisation–it can generate patient-specific treatment recommendations and tailor its research based on the unique needs of individual and population profiles. Health care outcomes, for the same conditions and diseases, can vary across regional and population groups. AI can help track and consolidate data for particular population and regional groups. This allows for improved prognostication: clinicians will be able to better track trends from their own practice or areas, instead of relying on data that may not be best suitable for their patient groups. This is particularly significant, as it can help India understand and analyse its population health metrics better. Such an approach can also empower patients to assume a more active role in collaborative decision-making with their clinicians by allowing them greater access to improved quality of information.
While GAI promises to be a gamechanger for improving individual and public health, the use of any AI, as an emerging technology, poses questions of data privacy and biases in algorithms which may impact outputs for diverse population segments.
As GAI needs voluminous datasets to train on for increased accuracy, confidential patient data must be carefully handled to avoid misuse and breaches, especially to third parties. To build trust in AI and its relevance for health care, there is an urgent need to increase public and medical trust in data handling by AI systems.
First, there is a need for clean, evidence-based data. Secondly, a clearly outlined framework for patient privacy is needed. To tap AI's potential to improve health care outcomes, it must be leveraged through a comprehensive framework that governs its usage and applicability. The creation of such a framework necessitates multi-disciplinary collaboration between clinicians, data scientists, ethicists, and policymakers, as it has broad ramifications across medical, ethical, legal, and social lines.
At the outset, data privacy vulnerabilities must be thoroughly accounted for, ensuring that any data used is with consent, and that patient data is accurate, properly anonymised and only shared with authorised parties. Algorithms require continuous human oversight to detect and mitigate gender or historical biases. Additionally, operational issues such as transparency and accountability must be prioritised by organisations adopting and integrating AI into their existing systems. As regulation around AI is still evolving, a constant eye on regulatory compliance is necessary to stay up to date with legal requirements. Continuous monitoring and evaluation are key to ensure use of GAI is both ethical and efficient.
This article is authored by Dr Arun Khemariya, senior clinical specialist, Elsevier India.