Older adults in three Chinese cities are more willing to keep using community smart services when they trust the platform, find the rules clear and correctable, and feel confident and supported in handling digital tasks, according to a three-wave survey of 520 people published in Frontiers in Public Health. The study, led by Guanxi Chen of NingboTech University and Jiajun Chen of Joongbu University, applied the Theory of Planned Behavior to continuance intention, the willingness to choose a digital service again after prior use, and found that attitude, subjective norm, and perceived behavioral control were each positively associated with that intention.
The research focused on community smart services, digital channels through which residents handle everyday public transactions such as appointment booking, repair and grievance requests, utility payments, information inquiries, health follow-up reminders, and community activity registration. For older users, difficulty navigating these channels can restrict access to resources relevant to healthy aging, which is why continued use matters for digital inclusion. The authors argued that repeated engagement raises questions about the fairness, manageability, and reliability of digitally mediated public services beyond the mere availability of technology.
To capture these experiences, the team surveyed older adults who had used at least one community smart service in the preceding 30 days across eight communities in Zhejiang Province: three in Ningbo, three in Hangzhou, and two in Shaoxing. Data were collected between March and August 2026 in three time-lagged waves. At the first wave, 735 participants rated six contextual perceptions: procedural fairness, accessibility fairness, digital self-efficacy, digital burden, perceived risk, and platform trust. At the second wave, 680 participants rated attitude, subjective norm, and perceived behavioral control. At the third wave, 627 completed the continuance intention measures. After screening, 520 complete three-wave matched cases were retained, a cumulative follow-up attrition rate of 14.69 percent. The final sample included 200 participants from Ningbo, 192 from Hangzhou, and 128 from Shaoxing.
The participants were mostly aged 60 to 75, with 90.00 percent falling in that range, and 53.08 percent were women. Most were retired, and half reported only occasional trial use of the services. Half said they could basically complete tasks independently, 40 percent occasionally required assistance, and 10 percent frequently needed help from others. Trained research assistants administered paper questionnaires at community sites, and anonymous codes on paper cards linked responses across waves without attaching identifying information to the analytical dataset.
The analysis used partial least squares structural equation modeling in SmartPLS 4.1, with bootstrap confidence intervals from 10,000 resamples and Benjamini-Hochberg false discovery rate adjustment across 39 structural and indirect tests. The measurement model performed well: Cronbach’s alpha ranged from 0.794 to 0.860, composite reliability from 0.879 to 0.914, and all 45 heterotrait-monotrait ratios fell below the conservative 0.85 threshold, the highest being 0.597 between platform trust and attitude. The model explained 36.4 percent of the variance in attitude, 24.9 percent in subjective norm, 24.1 percent in perceived behavioral control, and 28.5 percent in continuance intention.
The structural results showed that all three theory-of-planned-behavior pathways to continuance intention held after adjustment: attitude carried a standardized coefficient of 0.253, subjective norm 0.214, and perceived behavioral control 0.196, each with p below 0.001. Platform trust was the strongest contextual correlate, showing a positive association with attitude of 0.372, with subjective norm of 0.243, and with perceived behavioral control of 0.125. Procedural fairness was positively associated with all three components, with coefficients of 0.187 for attitude, 0.164 for subjective norm, and 0.138 for perceived behavioral control. Accessibility fairness, by contrast, was associated with attitude at 0.136 and subjective norm at 0.125 but showed no significant link to perceived behavioral control, where the coefficient was a negligible 0.012.
Digital capability and burden told a complementary story. Digital self-efficacy was positively associated with perceived behavioral control at 0.250, its strongest association, and also with attitude at 0.181 and subjective norm at 0.171. Digital burden, which captures the time, effort, and uncertainty of multistep procedures and repeated verification, was negatively associated with perceived behavioral control at minus 0.209 and with subjective norm at minus 0.092, while its association with attitude did not reach significance. Perceived risk showed a small negative association with perceived behavioral control at minus 0.081 but was not significantly associated with attitude or subjective norm, and none of its three specific indirect paths to continuance intention met the adjusted significance threshold.
The mediation analysis estimated 18 specific indirect associations from the six contextual perceptions to continuance intention through the three planned-behavior components. Platform trust carried the largest indirect association through attitude, at 0.094, followed by significant paths through subjective norm at 0.052 and perceived behavioral control at 0.025. Digital self-efficacy was significant through all three components, with the strongest path through perceived behavioral control at 0.049. Procedural fairness was significant through attitude at 0.047, subjective norm at 0.035, and perceived behavioral control at 0.027, while accessibility fairness was significant through attitude at 0.034 and subjective norm at 0.027 but not through perceived behavioral control. Digital burden’s negative indirect association was concentrated in the perceived behavioral control pathway at minus 0.041.
Supplementary analyses added further texture. A feedforward neural network predicting continuance intention from the nine construct scores did not outperform a ridge regression benchmark on held-out data, with mean held-out RMSE of 0.604 for the network against 0.563 for the linear model. Permutation importance ranked attitude highest at a normalized 100 percent, followed by subjective norm at 71.7 percent and perceived behavioral control at 53.8 percent; among the contextual antecedents, platform trust led at 36.0 percent, with perceived risk lowest at 2.6 percent. Necessary condition analysis identified moderate, statistically supported ceiling relationships for attitude, subjective norm, and perceived behavioral control, with effect sizes of 0.1995, 0.1654, and 0.1553 respectively, all p below 0.001, while the six contextual antecedents showed no comparable necessity evidence.
Twelve follow-up interviews illustrated how these statistical patterns appear in daily life. One participant said that as long as the process is clearly explained and someone is available to ask when something goes wrong, the service feels usable over the long term. Another observed that someone helping to get in does not mean being able to use it alone next time, a distinction that mirrors the gap between accessibility fairness and perceived control. Complaints about repeated verification and fear of tapping the wrong thing reflected digital burden and perceived risk, while reassurance came from knowing a channel genuinely belongs to the community and that someone will deal with problems. The authors concluded that continued engagement depends on more than technological access, pointing to age-friendly design, transparent and correctable procedures, capability support, responsive assistance, and reliable offline alternatives. They cautioned that the study covered existing users in eight communities, did not measure actual service use over time, and lacked baseline records for participants lost to follow-up, leaving room for further research on non-users and changing service conditions.
Subject of Research: Older adults' continuance intention for community smart services in China
Article Title: Fairness, trust, and digital inclusion in older adults’ intention to continue using community smart services in China
Article References: Chen, G., & Chen, J. (2026). Fairness, trust, and digital inclusion in older adults’ intention to continue using community smart services in China. Frontiers in Public Health, 14, Article 2004189. https://doi.org/10.3389/fpubh.2026.2004189
Image Credits: AI Generated
DOI: 10.3389/fpubh.2026.2004189
Keywords: community smart services, older adults, digital inclusion, Theory of Planned Behavior, platform trust, procedural fairness, digital self-efficacy, digital burden, perceived risk, continuance intention, PLS-SEM, China
News Source: Ophelia Keating. (November 4, 2026). Trust and fair procedures keep older adults using community smart services in China. Scienmag.
