JMIR Publications

ABSTRACT

Background:

The growing integration of Artificial Intelligence (AI) into mobile health (mHealth) applications offers new opportunities to improve healthcare access, particularly in low-resource settings. Yet, the opaque nature of AI-generated recommendations poses challenges of transparency, trust, and cultural relevance. These issues are often overlooked by existing Explainable AI (XAI) frameworks designed in high-resource contexts.

Objective:

This study aims to develop a human-centered design framework for explainability in mHealth applications adapted to the Bangladeshi context, addressing gaps in transparency, trust, and cultural relevance.

Methods:

A mixed-method approach was employed, combining quantitative surveys (n=137) with qualitative interviews and focus groups (n=20) involving end-users, developers, clinicians, and XAI experts. Survey responses were analyzed descriptively, and qualitative data were analyzed using thematic analysis to identify explainability expectations, trust perceptions, and design priorities.

Results:

Trust perceptions varied significantly across education and age groups among 137 surveyed users. Younger users (18-24 years) with lower formal education rated apps highest in perceived trust and clarity (mean Likert score: 5.0), while postgraduate degree holders (31-40 years) rated explainability lower (mean: 3.25), and users aged 40+ expressed strong skepticism (mean: 2.0). Despite recognizing AI's utility for preliminary guidance, users emphasized the necessity of human validation and expressed concerns about understanding AI's decision-making logic. Interviews with 6 XAI experts, 4 developers, and 10 medical professionals revealed critical gaps. Developers acknowledged minimal explainability implementation in current mHealth apps, while medical professionals unanimously prioritized clinical judgment over automated outputs and advocated for physician-mediated AI systems. Synthesizing findings across all stakeholder groups revealed five core requirements: (1) Human–AI collaboration and clinical validation, (2) Transparent logic paths, (3) Contextual personalization, (4) Cultural and linguistic relevance, and (5) Trust calibration with ethical safeguards.

Conclusions:

The framework bridges stakeholder misalignments and offers actionable guidance for design, deployment, and policy alignment in resource-constrained environments. By situating explainability within the sociocultural realities of the Global South, this research advances XAI beyond algorithmic transparency toward equity, inclusion, and user empowerment in digital health. Clinical Trial: N/A