Machine Learning-Enhanced Clinical Decision Making in Dementia Care – A Systematic Review

Background: A strategic focus on dementia care has become critically urgent in the current world settings. Dementia has become a rapidly escalating mass-scale global health and societal crisis. According to various credible estimates, it impacts between fifty-five million and sixty-five million people worldwide today. This imposes unsustainable economic costs on healthcare systems and demands highly skilled caregivers. It further necessitates population-level interventions, integrated health systems, universal policies, and human-centered care. In this context, advanced tools such as clinical decision support systems are becoming increasingly vital for advancing the quality and efficiency of dementia care. Objective: This study hypothesizes that integrating artificial intelligence (AI), particularly machine learning, into dementia care can substantially improve decision-making accuracy in dementia care. It also hypothesizes that this can broaden the support for critical areas in dementia care such as equity and individualized, patient-centered care, as well as facilitate policy-level interventions. The paper aims to assess the extent to which current research on machine learning-based clinical decision support systems addresses these multidimensional needs beyond the customary focus on diagnosis. Methods: A thorough review of empirical literature was conducted. The review was guided by the PRISMA methodology. The paper examined peer-reviewed studies focusing on machine learning-powered clinical decision support systems in dementia care. Particular attention was paid to the distribution of research focus across diagnostic and non-diagnostic domains. Results: The findings reveal that contemporary research on machine learning-based clinical decision support systems in dementia care is overwhelmingly concentrated on diagnostic tasks, not excluding early detection, multimodal data integration, and clinical workflow optimization. Key areas such as policymaking support, equity-driven decision-making, caregiver empowerment, and holistic patient-centered care remain underexplored. This paper recommends that while diagnosis is a direct corollary of medical decision-making, significant decisions are also required at other levels and in other areas of dementia care. Therefore, it is imperative that future research on smart decision support in dementia care extends its focus beyond diagnostics to address these other crucial dimensions.