Abstract
Background: The comorbidity of mental disorders and diabetes is on the rise, presenting a significant global public health challenge that gravely impacts the physical and psychological health of patients and presents obstacles to their effective management. The COVID-19 pandemic, in particular, aggravated these challenges owing to restrictions on in-person care. Artificial Intelligence (AI) interventions have emerged as a promising solution to alleviate this burden. Thus, we conducted a scoping review to map the current evidence in the literature and provide a clear understanding of AI for mental disorders and comorbid diabetes.
Methods: This scoping review utilized the Arksey & O'Malley framework. Five electronic databases were systematically searched for studies published in the post-COVID-19 era, focusing on AI-assisted approaches for individuals with mental disorders and comorbid diabetes.
Results: Twenty-four studies were reviewed. Supervised learning algorithms were most commonly employed, with studies reporting positive performance metrics, while generative AI was essentially absent. The findings demonstrated promising outcomes in four functional domains: Risk assessment, Prediction, Diagnosis, and Personalised care. AI applications are predominantly at the capability stage of translation maturity, focusing mainly on model development and validation, with limited adoption in existing clinical workflows. Other limitations include little demographic reporting, limited external validation, scarce intervention-focused research, and ethical concerns.
Conclusion: AI shows promise in enhancing support for individuals with mental disorders and diabetes through targeted, data-driven strategies. Future studies should prioritize clinically integrated, patient-centred AI interventions and evaluate their effectiveness in improving functional outcomes, while addressing ethical considerations.