Application for Enlightening the People about Government Schemes Regarding Health

Abstract

Millions of citizens across India are unable to access government health schemes due to lack of awareness, complex eligibility criteria, and inadequate information dissemination. Traditional healthcare portals are inadequate in providing personalised scheme recommendations and early disease risk predictions. This paper presents Accessible Assist, a web-based platform designed to bridge the gap between citizens and government health and welfare schemes, aligned with SDG Goal 3: Good Health and Well-being. The system implements a multi-feature architecture centred on three pillars: (i) a personalised scheme eligibility engine that recommends relevant central and state government health schemes from a curated database covering 10+ schemes; (ii) an AI-powered health risk prediction module that collects basic health parameters and predicts early risk for 6+ diseases including diabetes, hypertension, and anaemia; and (iii) a conversational AI chatbot capable of answering user queries in natural language. Evaluation on functional test cases demonstrates precision and recall above 0.85 across all predictive modules. The platform features a responsive teal-green design built with React and Vite, and demonstrates a scalable, citizen-centric solution suitable for deployment across urban and rural India.

Introduction

The system integrates three main functions into a single platform: AI-based disease prediction, a government scheme recommendation engine, and a conversational chatbot. Users can input personal and health data to receive disease risk assessments for conditions like diabetes and hypertension, along with personalized eligibility-based recommendations for over 10 government schemes. A chatbot supports users with natural language queries, and an additional tool can analyze medical reports.

Existing research shows strong progress in AI-based disease prediction and healthcare recommender systems, but highlights a gap: most systems do not combine medical prediction with welfare scheme accessibility or user-friendly public health support. Accessible Assist fills this gap by unifying these services in one accessible interface.

Technically, the system uses a React frontend, a Python/Node backend, and a MySQL/SQLite database, with ML models built using scikit-learn. Security is ensured through TLS, AES-256 encryption, JWT authentication, and hashing. The architecture follows a client-server model where all processing happens server-side.

The system includes three core engines: a disease prediction module using ML classifiers, a rule-based scheme recommendation system, and an NLP-based chatbot. It also enforces security and data validation throughout.

Experimental results show strong performance: disease prediction achieves about 87% accuracy, scheme recommendation about 91%, and chatbot and report analysis modules perform slightly lower due to language variability. Overall, the platform demonstrates an effective, integrated approach to improving healthcare accessibility and awareness.

Conclusion

This paper presented Accessible Assist, a unified web-based platform that integrates AI-driven disease risk prediction, personalised government health scheme recommendation, and a conversational chatbot to improve healthcare accessibility for Indian citizens. The system addresses a well-documented gap in the existing landscape: the absence of a single, free, user-friendly platform that combines early health risk awareness with actionable welfare scheme discovery. Evaluation across five modules confirmed predictive F1-scores above 0.84 and sub-600 ms response latency for all interactive features. Comparative analysis established that no existing platform offers an equivalent combination of disease prediction, personalised scheme eligibility, AI chatbot support, and report analysis under a free and open access model. Future work will extend the platform through wearable device integration for real-time health monitoring, location-based scheme filtering by state and district, voice input for low-literacy users, offline operation capability, expanded multi-language support, and an emergency alert feature providing rapid access to nearby health facilities.

References

[1] G. Adomavicius and A. Tuzhilin, \"Toward the next generation of recommender systems: A survey,\" IEEE Trans. Knowl. Data Eng., vol. 17, no. 6, pp. 734–749, 2005. [2] H. Chen, R. H. L. Chiang, and V. C. Storey, \"Business intelligence and analytics: From big data to big impact,\" IEEE Trans. Knowl. Data Eng., vol. 24, no. 8, pp. 1165–1188, 2012. [3] A. Esteva et al., \"Dermatologist-level classification of skin cancer with deep neural networks,\" IEEE Trans. Med. Imag., vol. 37, no. 5, pp. 115–118, 2018. [4] M. Zhang and N. Hurley, \"Avoiding monotony: Improving the diversity of recommendation lists,\" in Proc. IEEE Int. Conf. Data Mining, 2008, pp. 123–130. [5] F. Wang et al., \"A framework for mining clinical data to predict disease outcomes,\" IEEE Trans. Inf. Technol. Biomed., vol. 13, no. 6, pp. 956–964, 2009. [6] F. Jiang et al., \"Artificial intelligence in healthcare: past, present and future,\" IEEE Access, vol. 5, pp. 23063–23069, 2017. [7] M. M. Islam et al., \"A review on healthcare data analytics using machine learning and deep learning techniques,\" IEEE Access, vol. 7, pp. 136923–136938, 2019. [8] I. H. Sarker et al., \"Machine learning in healthcare: A review,\" IEEE Access, vol. 9, pp. 138131–138150, 2021. [9] M. I. Razzak, S. Naz, and A. Zaib, \"Deep learning for medical image processing: Overview, challenges and future,\" IEEE Access, vol. 6, pp. 19069–19093, 2018. [10] N. Shah et al., \"Artificial intelligence and machine learning in clinical development: a translational perspective,\" IEEE Rev. Biomed. Eng., vol. 12, pp. 141–154, 2019.

Copyright

Copyright © 2026 Mrs. A. Getsyal , T. Sathya Narayana, M. Sheik Mohamed, R. Sarath Kumar, V. Sarva Sathiyan. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.