Abstract:

[Purpose/Significance] Online medical communities (OMCs) have become an indispensable infrastructure that allows patients to access professional knowledge and enables physicians to expand their service, However, persistent complaints about information quality, physician disengagement and patient attrition reveal that the involved parties often fall into a non-cooperative trap. Prior literature has dominantly examined either the static motivational profiles of individual physicians or patients, or the dyadic interactions between one knowledge contributor and one knowledge seeker. Our understanding of how value co-creation unfolds in a service-ecosystem setting, where platforms, physicians, and patients simultaneously adjust their strategies, is hampered by the absence of a holistic, dynamic, and multi-actor perspective. Therefore, this paper shifts the analytical lens from isolated behavior or bilateral exchange to an evolutionary game among three interdependent stakeholders - the platform, the physician and the patient - within the knowledge service ecosystem. By embedding regulatory cost-benefit logic, trust mechanism and perceived-value-loss arguments into an evolutionary framework, the study unpacks the conditions under which collective knowledge sharing can be sustained and identifies the critical levers that can nudge the system towards a virtuous equilibrium. The findings will advance service-dominant logic and knowledge-sharing theory by revealing how micro-level strategic adaptations aggregate to create macro-level ecosystem viability. The findings will provide actionable insights for platform governance aimed at mitigating the real-world crises such as physician burnout and patient dissatisfaction. [Method/Process] Drawing on evolutionary game theory, we constructed a tripartite model in which the platform chooses between active regulation and passive regulation, the physician between active contribution (including both knowledge transfer and affective/extra-role support) and passive contribution, and the patient between active participation (information search, feedback and self-management behavior) and passive participation. Utility functions were specified to capture the net payoffs of each actor under eight possible strategy combinations, incorporating extra benefits, additional costs, perceived value losses, platform incentives and trust-based moderators. Using replicator dynamics, we derived the evolutionary stable strategies (ESS) for each actor and the joint ESS for the system. MATLAB simulations were then employed to trace the trajectory of strategy adjustment under varying parameter values, with sensitivity analyses performed for regulatory cost, physician reward, community trust and patient effort cost. Parameter ranges were anchored in prior empirical evidence and refined to ensure convergence to feasible equilibria. [Results/Conclusions] The analytical and simulation results converge on three main insights. First, the system possesses a unique pareto-dominant equilibrium-the triad (active regulation, active contribution, and active participation)-that emerges when the product of each actor's trust-adjusted net benefit exceeds the corresponding threshold. Second, the transition path is highly sensitive to the relative magnitude of marginal benefits and costs: lowering the platform's regulatory expenditures or increasing its incremental revenue will accelerate convergence to active regulation; enhancing physicians' reputational and intrinsic rewards or reducing their affective labor cost will markedly elevate cooperative contribution; and compressing patients' cognitive and privacy cost while enlarging their health outcome gain will propel active participation. Third, community trust operates as a critical moderator: when trust is high, physicians are willing to contribute even if the perceived value loss from non-contribution is modest, whereas low trust neutralizes the effect of potential gains and locks the system into a low-effort trap. From a managerial perspective, the study recommends that platforms deploy AI-assisted tools to relieve physicians of repetitive tasks, calibrate incentive budgets to prevent overspending, and establish fairness-enhancing governance practices to foster trust. However, limitations include the omission of knowledge flows between physicians and between patients as well as the reliance on stylized parameters. Future research could extend the model to a multi-layer network incorporating professional sub-communities and patient peer groups, and calibrate the payoff structure with field data could enhance the model's external validity.

Key words: service ecosystem, value co-creation, online medical community, knowledge sharing, evolutionary game