A NASSS-based framework for AI-driven prevention of non-communicable diseases in Iran: a qualitative study

Non-Communicable Diseases (NCDs) are recognized as the leading cause of mortality worldwide, accounting for approximately 74% of all deaths. In Iran, this figure exceeds 80%, underscoring the critical need for innovative preventive strategies. Artificial Intelligence (AI), with its capabilities in early detection, risk prediction, and the design of targeted interventions, holds significant potential for mitigating the burden of NCDs. However, the successful integration of AI into a health system necessitates a thorough understanding of local cultural, organizational, and structural contexts. Grounded in the NASSS (Non-adoption, Abandonment, Scale-up, Spread, and Sustainability) theoretical model, this study aimed to design and validate a localized framework for the application of AI in NCD prevention within the Iranian health system.

This study employed an exploratory-sequential qualitative design to develop and validate a localized framework for the application of AI in preventing Non-Communicable Diseases (NCDs) within the Iranian health system. In the initial phase, data from 34 semi-structured interviews with experts in health, technology, ethics, and policy were analyzed using a hybrid inductive-deductive thematic analysis. In the subsequent phase, the preliminary framework was refined and finalized using a modified Delphi method and two focus group discussions with 13 experts, yielding a high overall content validity index of 0.91.

Analysis guided by the NASSS model identified five principal themes and 19 initial subthemes. Following Delphi validation and framework refinement, the final framework comprised five major dimensions and 18 operational components tailored to the Iranian health system. The framework emphasizes algorithmic transparency, data standardization, digital health literacy, and workforce empowerment as key requirements for successful AI implementation in NCD prevention.

The finalized framework provides a locally validated model for the ethical, sustainable, and context-sensitive integration of AI into NCD prevention. It offers practical guidance for policymakers, health system managers, and researchers in Iran and other countries with similar healthcare contexts. The framework may assist decision-makers in prioritizing digital infrastructure investments, strengthening data governance, and implementing context-sensitive AI strategies for NCD prevention.