AmarDoctor: An AI-Driven, Multilingual, Voice-Interactive Digital Health Application for Primary Care Triage and Patient Management to Bridge the Digital Health Divide for Bengali Speakers
Nazmun Nahar, Ritesh Harshad Ruparel, Shariar Kabir, Sumaiya Tasnia Khan, Shyamasree Saha, Mamunur Rashid · 2025
Read AmarDoctor: An AI-Driven, Multilingual, Voice-Interactive Digital Health Application for Primary Care Triage and Patient Management to Bridge the Digital Health Divide for Bengali Speakers by Nazmun Nahar, Ritesh Harshad Ruparel, Shariar Kabir, Sumaiya Tasnia Khan, Shyamasree Saha, Mamunur Rashid (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-24724.txt.
Introduction
Nazmun Nahar, Ritesh Harshad Ruparel, Shariar Kabir, Sumaiya Tasnia Khan, Shyamasree Saha, and Mamunur Rashid made AmarDoctor, a health app for Bengali speakers. Big health apps tend to serve users in rich lands and main tongues, while Bengali speakers face gaps in care and in tech ease. Their app speaks and listens, and serves both sick users and care staff.
For users, a step wise ask flow weighs signs of ill health and points to the right kind of doctor. Voice help guides users who find screens hard to use. For staff, the app adds AI tips for care calls. It works in three main Bengali speech forms.
Catalog record
| CREATOR | Nazmun Nahar, Ritesh Harshad Ruparel, Shariar Kabir, Sumaiya Tasnia Khan, Shyamasree Saha, Mamunur Rashid |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-24724
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
AmarDoctor: An AI-Driven, Multilingual, Voice-Interactive Digital Health Application for Primary Care Triage and Patient Management to Bridge the Digital Health Divide for Bengali Speakers Nazmun Nahar, Ritesh Harshad Ruparel, Shariar Kabir, Sumaiya Tasnia Khan, Shyamasree Saha, Mamunur Rashid Abstract This study presents AmarDoctor, a multilingual voice-interactive digital health app designed to provide comprehensive patient triage and AI-driven clinical decision support for Bengali speakers, a population largely underserved in access to digital healthcare. AmarDoctor adopts a data-driven approach to strengthen primary care delivery and enable personalized health management. While platforms such as AdaHealth, WebMD, Symptomate, and K-Health have become popular in recent years, they mainly serve European demographics and languages. AmarDoctor addresses this gap with a dual-interface system for both patients and healthcare providers, supporting three major Bengali dialects. At its core, the patient module uses an adaptive questioning algorithm to assess symptoms and guide users toward the appropriate specialist. To overcome digital literacy barriers, it integrates a voice-interactive AI assistant that navigates users through the app services. Complementing this, the clinician-facing interface incorporates AI-powered decision support that enhances workflow efficiency by generating structured provisional diagnoses and treatment recommendations. These outputs inform key services such as e-prescriptions, video consultations, and medical record management. To validate clinical accuracy, the system was evaluated against a gold-standard set of 185 clinical vignettes developed by experienced physicians. Effectiveness was further assessed by comparing AmarDoctor performance with five independent physicians using the same vignette set. Results showed AmarDoctor achieved a top-1 diagnostic precision of 81.08 percent (versus physicians average of 50.27 percent) and a top specialty recommendation precision of 91.35 percent (versus physicians average of 62.6 percent). K eywords Digital Health \cdot Clinical Decision Support System(CDSS) \cdot Accessible and Inclusive Healthcare \cdot Healthcare Management 1 INTRODUCTION In an era where technology is integrated into nearly every aspect of daily life, healthcare has undergone a significant shift toward digitization. Digital health solutions, including AI-powered symptom assessment tools, virtual consultations (e.g., telehealth, telemedicine), and lifestyle and wellness guidance, are now at the forefront of this transformation, becoming increasingly prevalent among patients who turn to the internet as a primary resource for health-related concerns. AI-powered self-triaging solutions such as WebMD, Symptomate, Ada, and Klinik Health have played a key role in this digital health revolution, empowering individuals to take greater control over their healthcare decisions [ 1 ] , [ 2 ] , [ 3 ] . Amidst the growing demand for virtual primary care, healthcare systems worldwide are embracing these AI-powered digital triaging applications to achieve more efficient patient management [ 4 ] . These cutting-edge tools are redefining primary care delivery by alleviating workload pressures on medical staff in many developed countries ( Klinik Healthcare, May 2021 ). However, despite these advancements and their rapid adoption in developed countries, digital health solutions face significant challenges in scaling equitably across diverse contexts. While these technologies hold the potential to address gaps in healthcare access – it’s been estimated that half of the world’s population lacks access to essential healthcare services [ 5 ] , [ 6 ] – there’s growing evidence indicating that they are, in some cases, widening existing inequalities [ 7 ] , [ 8 ] , [ 9 ] . This is particularly evident in low- and middle-income countries, where various factors hinder the effective implementation and adoption of digital health solutions. In this context, the case of Bengali speakers is particularly salient. Bengali, the world’s seventh-largest ethnolinguistic group, with 270 million speakers across Bangladesh, India, and the Gulf region, remains one of the most underserved populations in digital healthcare. Especially, Bangladesh, with its 180 million population, has one of the worst primary care systems among its contemporaries, with only 9.9 doctors for every 10,000 patients [ 10 ] . The situation is even more severe in rural areas, where there are only 1.1 doctors for the same number of patients [ 11 ] . Due to the absence of an established primary care system, trivial illnesses often escalate into health emergencies, pushing 6 million people into poverty due to out-of-pocket (OOP) emergency health expenditure. Despite high smartphone penetration (60 million) and good internet coverage, digital health solutions are not scaling well in Bangladesh. Several primary barriers seriously hinder the scaling of digital health solutions in LMIC countries like Bangladesh. Poor Digital Literacy These include poor digital literacy among both patients and doctors [ 12 ] , and the fact that the majority of available digital health solutions primarily support English or European languages, failing to encompass the linguistic nuances around the medical vocabulary for many non-European languages. Furthermore, digital health services available via mobile or web applications require patients to type in a digital medium (e.g., smartphone or keyboard). There are several versions of Bengali digital keyboards (e.g., Unicode-based Bengali keyboards) and a mixture of phonetic-based approaches, each with different layouts. This lack of standardization makes typing Bengali in digital media unpopular among low socioeconomic groups and those with poor digital literacy [ 13 ] . Bias in Current Clinical Decision Support Systems Furthermore, AI-driven clinical decision support systems offer the potential to help physicians reduce consultation time and increase operational efficiency, which is a critical priority in many developing countries. However, the effectiveness of these platforms is often limited by their reliance on patients’ ability to accurately report their medical concerns through a text-based interface. This presents a significant challenge for many non-Latin language speakers with low digital literacy, as their languages frequently do not translate seamlessly onto standard digital keyboards. This issue is further compounded by the fact that current recommendation systems are often trained on data predominantly from white European patients, which can significantly reduce their accuracy and effectiveness when applied to patients from non-European ethnolinguistic backgrounds [ 14 ] , [ 15 ] , [ 16 ] . Consequently, these oversights can disadvantage native language speakers who lack adequate English skills and prefer to use their native tongues when accessing digital health solutions, thereby increasing the risk of exacerbating existing health inequalities [ 17 ] , [ 9 ] . The convergence of these technical challenges with prevalent poor literacy has significantly impeded the progress of healthcare digitization and the widespread adoption of digital technologies. Lack of Domain Specific Foundation Models in Bengali Recent advancements in large language models (LLMs) are beginning to reshape the digital healthcare landscape, offering unprecedented capabilities in clinical support and patient engagement. However, the efficacy of these powerful models is not universal; their performance is highly contingent on the linguistic and clinical context for which they are trained [ 18 ] . This creates a significant disparity for the world’s 270 million Bengali speakers. Despite its global prevalence, Bengali remains a low-resource language in the digital domain, with a pronounced lack of medical-specific foundational models, which has hindered the ability to adapt cutting-edge AI for this population [ 19 ] . Consequently, while multi-modal LLMs are empowering millions worldwide with access to sophisticated medical advice and user-friendly conversational bots, Bengali speakers have largely been excluded from these advancements, deepening the digital divide in healthcare access Figure 1: AmarDoctor’s patient-centered medical triage and primary healthcare delivery. The workflow begins with a multilingual, voice-interactive chatbot that classifies