MindCare: An Integrated On-Device Affective Computing, Hybrid Adaptive Screening, AndClinician-in-the-Loop Tele-Consultation Framework For Mental Health Triage

19 Nov

Authors: Dr. R Ashok Kumar, Satavisa Deka, Shuchi Sharma, Varun R

Abstract: Digital mental health screening, affective text analysis and access to professional support are usu-ally delivered through separate systems. Fixed questionnaires adapt poorly to what an individual writes, while cloud-based conversational tools raise privacy and governance concerns. This paper presents MindCare, a soft-ware framework that combines on-device lexical text analysis, adaptive screening, clinician-facing summaries, specialist recommendation and tele-consultation in one application. The text analysis extracts sentiment, emo-tion and risk indicators from free-text responses and combines them with a 15-question bank covering anxiety, depression, stress and interpersonal difficulties. Question selection follows a hybrid scheme with general, spe-cific and confirmation phases, a safety override and an evidence-based early-stopping rule. The prototype is built with Flutter and Dart, using local storage, Cloud Firestore and Agora video calls. It was checked with 64 unit and widget tests across 13 modules, all of which passed. These results show consistent software be-haviour only; they do not show clinical validity, accuracy or effectiveness. Clinical testing, privacy and security hardening, and validation against recognised instruments are required before clinical use.

DOI: https://doi.org/10.5281/zenodo.24361794