Current approaches to medical and welfare care and prevention largely rely on quantitative data—such as age, weight, and blood pressure—while often overlooking the qualitative aspects of individuals’ lives, such as their daily routines, emotional states, and subjective experiences. As a result, there are significant limitations in the ability to design care and prevention strategies that truly reflect individual needs.

Qualitative data—such as people’s feelings, experiences of hardship, or personal narratives—capture vital, non-numeric insights. With recent advancements in artificial intelligence, it has become possible to incorporate these qualitative data into predictive models, allowing for more tailored care and preventive strategies. AI can now analyze lifestyle and psychological patterns to estimate health risks and provide personalized health recommendations.

The Inochi Forum, in collaboration with the Social Solution Initiative (SSI) of The University of Osaka and its core project “Science and Humanity for Fostering a Super-aged Society that Respects Individual’s Views on Life and Death and Their Autonomy” (PI: Assoc. Prof. Miyae Yamakawa, Graduate School of Medicine), is building a data science platform that integrates healthcare and welfare data. The initiative aims to use qualitative data within AI-based predictive models to enhance health and well-being outcomes. This approach enables care and prevention to be customized for each individual and contributes to broader community health promotion.

While AI and predictive analytics have recently made significant inroads into the fields of medicine and welfare, existing AI models are still primarily based on quantitative data. As such, they are insufficient for capturing people’s lived realities and perspectives. To further improve these technologies, it is essential to incorporate qualitative data that reflect patients’ real-life conditions, enabling more empathetic and customized disease prevention and eldercare prediction.

Such integration can enhance patient satisfaction and reduce medical costs.

A conceptual model illustrating this integrated approach shows how combining healthcare and welfare data can lead to better care and support. Conventional AI models have focused on quantitative data such as clinical records and care logs.

However, these data alone cannot fully grasp the experiences and perceptions of patients. New models must include qualitative data such as patient and caregiver narratives. This richer data enables deeper understanding of patients’ environments and social contexts. This approach is called “CHR” (Contextual Health Records)— a framework that allows all stakeholders in healthcare and caregiving to share insights about patients’ lives and needs, enabling more personalized care and preventive interventions. CHR aims to elevate the quality of both care and support by offering optimal, individualized services.

Image of the new AI prediction model (CHR)

To advance this initiative, several concrete steps are being taken. To begin with, to gain a deeper understanding of the experiences of patients and caregivers, the project will collect qualitative data such as nursing and caregiving records, as well as patient narratives, and integrate them with electronic health records (EHRs) and personal health records (PHRs) into a unified, large-scale database. Next, the project will employ natural language processing (NLP) technologies to convert the collected data into formats that can be easily interpreted by AI systems. This will allow AI to learn from patients’ lifestyles and perspectives, enabling it to propose care plans tailored to individual needs.

Furthermore, the project will develop new AI predictive models that combine both quantitative and qualitative data, aiming to improve disease prevention and optimize caregiving. These models will be tested in real healthcare and welfare settings, where their effectiveness will be validated by comparing predicted outcomes with actual results, thereby enhancing the model’s accuracy. In addition, the project plans to host workshops in community spaces such as public libraries and civic centers. These workshops will not only help gather more qualitative data through people’s personal stories but also promote a better understanding of the role and value of AI-based models in healthcare.

Through its participation in these activities, the Inochi Forum seeks to improve the quality and sustainability of care by offering more personalized, empathetic support to patients and caregivers, and by strengthening collaboration between science and society.

[References]

・BMJ Health & Care Informatics: Roadmap for Developing and Evaluating AI Models

https://informatics.bmj.com/content/30/1/e100784

・Milliman: AI and Predictive Analytics in Healthcare
https://www.milliman.com/en/insight/ai-and-predictiveanalytics-are-transforming-healthcare-heres-how

・The University of Osaka SSI: Science and Humanity Core Project
https://www.ssi.osaka-u.ac.jp/activity/core/sciencehumanity/

[Action Platform]
Medical and Welfare

[SDGs]