CNN-based facial expression recognition and TRIZ-inspired service redesign in combined medical and elder care

In current combining medical and elder care (CMEC) services, there remains a lack of systematic and technical approaches capable of leveraging real-time emotional feedback from older adults to effectively drive the redesign of micro-level service processes. This gap significantly constrains the capacity for continuous optimisation and personalised responsiveness. To address this challenge, this study integrates service blueprint, facial expression recognition, and the theory of inventive problem solving (TRIZ) to develop the ‘4S’ model, which comprises four key phases: service blueprint construction, service failure diagnosis, service redesign solution generation, and service redesign solution evaluation. The model was field-tested in three large-scale CMEC institutions in Fujian Province, China. Through the analysis of older adults’ immediate emotional responses, three categories of critical service failures were identified, leading to the development of six targeted service redesign solutions. The innovativeness and feasibility of these solutions were positively evaluated. This study proposes the integrated ‘4S’ model to enable real-time emotion-driven service redesign; constructs a structured CMEC service blueprint, through which three critical service failure types are identified, and six TRIZ-inspired redesign solutions are generated; and extends service science by fully leveraging the temporal, spatial, and cognitive dimensions in service design, thereby maximising their combined impact.