Artificial intelligence in community medicine in India: The present and future!

Abstract

Community medicine emphasises the health of the population as a whole, rather than concentrating solely on individual patients.[1] As community medicine is a data-driven subject, AI holds an important potential to facilitate public health actions and policy. At the population level, artificial intelligence (AI) domains such as machine learning, deep learning, natural language processing and also the latest revolutionary large language models such as ChatGPT can be employed to analyse and train on large amounts of data that are available from electronic health records (EHRs). This analytical and predictive capability of AI algorithms makes it an effective tool in meaningfully interpreting data of public health importance, such as predicting any disease outbreak, improving the nutritional status of the community, or making health accessible to all.[1] The use of official and unofficial public health data in disease surveillance is one of the cornerstones of community medicine. One of the most important strengths of AI lies in making meaningful interpretations of unstructured data sets from a variety of sources such as health programme portals, media and EHRs. The introduction of the Integrated Health Information Platform (IHIP) – a web-based, real-time information platform further enhanced disease surveillance by integrating AI for event-based surveillance and media scanning in India. The IHIP platform provides comprehensive demographic, temporospatial and epidemiological characteristics of diseases reported as outbreaks from formal and informal sources. In addition, IHIP allows effortless exchange of data at all levels of the ministry and across various sectors, facilitating informed, evidence-driven decisions during major health events.[2] During the COVID-19 pandemic, AI-driven prediction models became highly essential to predict the progression of the disease all over the world. Various portals were created to collect authentic data at the national, state and district levels in India. This collected data were analysed in real-time enabling policymakers, researchers and doctors to track the peaks and the troughs of the epidemic curve along with the geographical distribution of COVID-19 cases, allowing to control and prevent further outbreaks.[3] While these AI models are proven to be successful in disease surveillance, it is crucial to note that these models might fail to give a clear picture if the training data are not consistent, the data collected are not accurate or a mass gathering occurs despite lockdowns. India launched its free-of-cost National Telemedicine Service, eSanjeevani, which is available in two distinct variants. The first is eSanjeevani AB-HWC (Ayushman Bharat-Health and Wellness Centres) (doctor to doctor) which links patients and community health officers at AB-HWC with specialists in tertiary hospitals/medical colleges. The second variant, eSanjeevani outpatient Department (patient to doctor), provides citizens access to specialist doctors directly through their smartphones. The main motto of this initiative is to provide Universal Health Coverage, strengthening the aim of the Ayushman Bharat scheme.[4] To date, eSanjeevani has served more than 35 crore patients, making it one of the world’s largest telemedicine programs that bridges the rural–urban divide.[5] As India is a multilingual country, eSanjeevani effectively leverages AI to generate multilingual prescriptions for patients to download. It also employs AI to provide a Clinical Decision Support System to the doctors through collecting patient symptoms, summarising them and suggesting differential diagnoses. AI-models also help in analysing doctor–patient interactions.[6,7] Despite its swift adoption, some technical challenges need to be addressed. Key issues such as lack of digital infrastructure in rural settings, lack of standard operating protocols, inadequate training in triage, unstandardised recording of patient symptoms and target-based referral approaches lead to over-referral or referral to wrong specialties, driving up the opportunity cost for already preoccupied specialist government doctors.[8] Recent studies indicate that the application of AI in the field of nutrition is still in its infancy and is primarily used in dietary assessment, with some applications in understanding diet-related diseases and lifestyle intervention on an individual level.[9] AI is efficient in analysing food composition and nutrients using food pictures and daily dietary assessment through various apps available for smartphone users. It aims to provide feedback about their fulfillment of nutritional requirements through what people eat and suggest better alternatives, ultimately focussed on improving the relationships between diet and health.[10] Although the guidelines for nutritional intake of the Indian population are released by the Indian Council of Medical Research – National Institute of Nutrition, the majority of these nutrition-related apps are developed by private businesses in the health and fitness industry.[11] To tackle malnutrition among school-going students in India, several measures were taken. Notably, Indian Administrative Service officer Shubham Gupta collaborated with organisations like Feeding India and Udyog Yantra together and came up with the effective implementation of an AI machine called NutreGro to monitor and improve the quality of meals served in public schools under the aegis of the mid-day meal program. NutreGro is an AI-enabled machine that takes a snapshot of a student’s plate, analyses the meal’s quality and quantity, provides nutritional analysis and correlates it with the BMI of the child.[12,13] In addition, AI-powered applications such as the child growth monitor are being piloted by the government to instantly detect malnutrition by directly scanning the children through a smartphone at Anganwadi Centres (AWCs).[14] To address the challenges of nutritional surveillance, the Indian government introduced a data-driven tool called ‘Poshan Tracker’ in 2021. This app tracks anthropometric outcomes through growth monitoring in children between 0–6 years of age, the distribution of supplementary food to children, women and adolescent girls at the AWCs. It has led to improved frequency in the collection of anthropometric data and the distribution of food packets for supplementary nutrition. In addition, it has also improved the accountability of AWCs.[15] From an overall nutritional perspective, AI has a significant unrealised potential to address a wide range of nutritional concerns, including identifying causes and interventions related to malnutrition, obesity, cardiovascular diseases, diabetes and hypertension on individual as well as community levels.[16] AI has the future potential to leave a remarkable impact on community medicine. However, for AI to function at its full capacity, it has to be available, accessible and sustainable. Improving digital infrastructure and digital literacy in rural India will make AI available and accessible beyond the urban landscape.[17] As most AI applications are being developed and trained on data sets in high and middle-income countries, they are not representative of the population residing in low- and middle-income countries like India, leading to selection bias.[18] Extending ‘Make in India’ effort in AI towards community medicine will not only yield data that is more accurate and free of bias, but training it on multilingual datasets and integrating cultural contexts will make it more acceptable. To ensure the safe and ethical use of AI-driven healthcare solutions, it is necessary to implement rigorous regulatory monitoring and strong data governance systems to ensure privacy and security. Many private firms, for example, those who sell digital fitness watches, nutrition tracking apps, have important data that can be utilised for public health purposes if community physicians build public–private partnerships to leverage them.[17] AI is undoubtedly reshaping community medicine, providing robust solutions for early disease detection, nutritional security and accessible health care. The future of AI in public health lies in innovatively thinking about public health problems with new perspectives and cross-collaboration between policymakers, community physicians and AI experts to harness its full potential.

Ähnliche Arbeiten