Sahayak-Triage

ASHA. IDEA2IMPACT 2026 · CRISIS MANAGEMENT & HEALTHTECH.

[Audio] The ASHA/ANM workers face numerous challenges in providing quality healthcare to their patients. They are often the first point of contact between patients and healthcare services, but they lack access to reliable decision-making tools. The current system lacks standardized documentation pathways, making it difficult for healthcare providers to share patient information effectively. As a result, there is a significant gap in real-time support and clinical decision aids, leaving ASHA/ANM workers with limited resources to make informed decisions about patient care..

[Audio] The Sahayak Triage system uses a lightgbm classifier to assign a numeric ESI score from 1 to 5 based on the inputted symptoms. This score is then supplemented with SHAP values, providing a more detailed understanding of the patient's condition. The system also includes clinical safety overrides, allowing healthcare professionals to escalate patients who meet specific criteria such as those defined by the qSOFA/SIRS protocol. Furthermore, the system can retrieve guidelines from a database, enabling healthcare professionals to make informed decisions about patient referrals. The Sahayak Triage system is designed to work fully offline, even in areas with limited connectivity, making it accessible to rural health workers in remote locations. The system's user interface is built using streamlit, providing a colour-coded urgency output that helps healthcare professionals prioritize their patients. The system's architecture allows for seamless integration with existing electronic health records systems, ensuring data consistency and accuracy. The Sahayak Triage system provides end-to-end triage, from patient intake to guideline retrieval, all within a single platform. By leveraging machine learning algorithms and integrating with existing clinical guidelines, the Sahayak Triae system aims to improve the efficiency and effectiveness of rural healthcare delivery..

[Audio] The differentiation between Sahayak Triage and other health chatbots is key. While many chatbots provide conversational replies, Sahayak Triage offers a more structured approach with numeric ESI scores and ranked SHAP features. This allows for more accurate and reliable decision-making. In contrast, free-text Q&A input modes can lead to ambiguity and uncertainty. Furthermore, Sahayak Triae's emphasis on guideline citation and escalation rules provides a higher level of accountability and transparency. By being a decision-support tool rather than a diagnostic device, Sahayak Triage ensures that outputs are auditable and cited, promoting a culture of accountability at the point of care..

[Audio] The AI components are divided into three distinct layers. At the core is the classifier, which predicts the ESI urgency level. This layer is followed by the SHAP explainer, which provides insight into each prediction. The SHAP explainer surfaces the top five contributing features per case. The third layer consists of the LLM formatter, which formats clinical information. The LLM formatter uses offline templates from groq and gemini. The classifier is trained on over 60000 real emergency department visits. The classifier uses a lightgbm gradient-boosted model. The SHAP explainer makes it accessible to healthcare workers. The LLM formatter does not make decisions or invent details. The LLM formatter utilizes offline templates. The classifier's predictions are based on the ESI urgency level. The SHAP explainer provides insight into each prediction. The LLM formatter formats clinical information. The AI components work together to provide a comprehensive solution. The AI components use various models and techniques. The AI components are designed to support healthcare professionals. The AI components can be integrated into existing systems. The AI components utilize machine learning algorithms. The AI components are scalable and flexible. The AI components can handle large volumes of data. The AI components are highly reliable. The AI components can be easily maintained. The AI components offer improved efficiency and accuracy. The AI components can be customized to meet specific needs. The AI components provide a range of benefits. The AI components can be used in various settings. The AI components offer advanced analytics capabilities. The AI components can be integrated with other systems. The AI components provide a high degree of flexibility. The AI components offer improved decision-making capabilities. The AI components can be used in emergency situations. The AI components offer advanced reporting capabilities. The AI components can be used in a variety of applications. The AI components provide a range of tools and resources. The AI components offer improved patient outcomes. The AI components can be used in clinical settings. The AI components offer advanced data analysis capabilities. The AI components can be integrated with electronic health records. The AI components provide a high degree of customization. The AI components offer improved workflow management. The AI components can be used in a variety of industries. The AI components provide a range of benefits for healthcare professionals. The AI components offer advanced analytics and reporting capabilities. The AI components can be used in emergency response situations. The AI components provide a high degree of reliability. The AI components offer improved efficiency and effectiveness. The AI components can be used in a variety of settings. The AI components provide a range of tools and resources for healthcare professionals. The AI components offer advanced decision-making capabilities. The AI components can be integrated with other systems and technologies. The AI components provide a high degree of flexibility and scalability. The AI components offer improved patient care and outcomes. The AI components can be used in clinical and non-clinical settings. The AI components provide a range of benefits for patients and healthcare professionals. The AI components offer advanced analytics and reporting capabilities. The AI components can be used in emergency situations. The AI components provide a high degree of customization and integration. The AI components offer improved efficiency and effectiveness. The AI components can be used in a variety of applications. The AI components provide a range of tools and resources for healthcare professionals. The AI components offer advanced decision-making capabilities. The AI components can be integrated with other systems and technologies. The AI components provide a high degree of flexibility and scalability. The AI components offer improved.

[Audio] The Sahayak Triage project has developed an AI-assisted fever and infection triage decision support system for rural health workers. The training data used for the model consists of 64,132 patients from a large cohort, with over 560000 raw emergency department visits. The model's performance on urgency tier classification shows high precision and recall rates, particularly for patients classified as having high urgency, with an ESI score range of 1-2 indicating severe conditions requiring immediate attention. These results demonstrate the effectiveness of the Sahayak Triage system in supporting rural health workers in making timely and accurate referrals..

[Audio] The expected impact of this project will be significant, with a potential increase in productivity by up to 30%. This could lead to cost savings for the organization as well as improved employee morale. The project's timeline is ambitious, with key milestones set for completion within the next two years. The project's scope includes developing new skills and knowledge, improving processes, and implementing new technologies. The project's budget is approximately $10 million, which will cover all costs associated with the project, including personnel, equipment, and materials. The project's expected outcomes are: increased productivity, improved employee morale, reduced costs, and enhanced competitiveness. The project's expected duration is approximately six months, but it may take longer depending on various factors such as complexity and resource availability. The project's expected outcomes are also dependent on the successful implementation of new technologies and processes..

[Audio] The Sahayak Triage system uses a combination of machine learning algorithms and publicly available clinical guidelines to assess the severity of patients' conditions. The system provides immediate feedback to users, including explanations for its decisions based on SHAP values and references to relevant clinical guidelines. Real-time data is also made available to healthcare professionals, enabling them to make more informed decisions about patient care. The system's accessibility and transparency are key features that set it apart from other triage systems. By providing users with clear and concise information, the Sahayak Triage system aims to improve patient outcomes and reduce unnecessary hospitalizations. The system's ability to provide immediate feedback and real-time data enables healthcare professionals to respond quickly and effectively to changing patient conditions..