Importance and perceived performance of smart public healthcare services among adults in Taiwan: an exploratory survey integrating modified SERVQUAL, importance–performance analysis, and SEM-based measurement diagnostics

Assessing smart public healthcare services requires understanding both which service attributes users consider important and how they perceive those attributes to perform in actual services. Such evidence can help identify attributes warranting further service design inquiry. This exploratory study combined a modified SERVQUAL framework and importance–performance analysis (IPA), using structural equation modeling (SEM) to diagnose the proposed measurement structure, among adults in Taiwan. An exploratory cross-sectional online survey using convenience and network-referral sampling was conducted in Taiwan in December 2024 ( n = 203 adults). Respondents rated the importance of 20 ideal-service attributes and the perceived performance of 20 corresponding attributes of actual services across five modified SERVQUAL dimensions (tangibles, reliability, responsiveness, assurance, and empathy), plus four overall evaluations. Performance-minus-importance differences were treated as descriptive discrepancies rather than conventional SERVQUAL expectation–perception gaps. SEM was used diagnostically to examine the measurement structure of separate importance-based and perceived-performance models; IPA used standardized item scores. Mean perceived-performance ratings were numerically lower than importance ratings across all five dimensions. Reliability showed the largest mean discrepancy (− 0.526), followed by tangibles (− 0.510), assurance (− 0.473), responsiveness (− 0.454), and empathy (− 0.379). The SEM diagnostics did not support the proposed five-dimensional structure: the dimensions were not empirically distinct (latent correlation up to 0.992), the importance-based model produced a non-positive-definite latent covariance matrix, and the perceived-performance model produced an inadmissible estimate (Heywood case); structural coefficients were therefore not interpreted substantively. Item-level IPA identified three attributes, all related to community-based and continuing care, with relatively high importance but relatively low perceived performance. In this exploratory sample, the attributes respondents considered important were distinguishable from their perceptions of actual service performance, and IPA identified candidate areas for further service design inquiry. The modified SERVQUAL structure was not validated in this sample and requires refinement and independent validation before it can support structural or comparative conclusions. Because the sample was not representative of users of Taiwan’s public healthcare system, the findings provide a starting point for stakeholder engagement, co-design, and measurement refinement rather than evidence of causal effects or established intervention priorities.