Reducing inappropriate medical practices: a scenario-based experiment on fiduciary duty and audit enforcement in China’s healthcare sector

China’s recent anti-corruption efforts in the healthcare sector have focused on senior hospital administrators and on procurement processes for pharmaceuticals and medical supplies, with comparatively little attention paid to frontline clinical physicians. Yet the everyday practices of frontline physicians—including excessive testing, overtreatment, and overprescription—constitute a major source of cost inflation and have contributed to a substantial erosion of patient trust. Drawing on principal–agent theory and on responsive regulation theory, this study examines whether introducing a statutory fiduciary duty between physicians and patients, enforced through emerging medical audit mechanisms, is associated with reductions in physicians’ self-reported willingness to engage in inappropriate practices. We conducted a sequential exploratory mixed-methods study comprising (i) a qualitative phase of eight semi-structured interviews with practicing clinical physicians in Beijing and Shenyang (March–July 2023), analyzed using reflexive thematic analysis, and (ii) a between-subjects scenario-based questionnaire experiment with 120 practicing physicians recruited from hospitals in Jinan, Qingdao, Shenyang, and Beijing. Participants in the quantitative phase were systematically allocated to an experimental condition (a proposed statutory fiduciary duty enforced through medical audits and graduated sanctions) or a control condition (the existing ethics-based framework). The primary outcome was self-reported willingness to reduce inappropriate medical practices, measured on a five-point Likert scale. Statistical analyses comprised an independent-samples t test, a Mann–Whitney U test, multiple linear regression with HC3 heteroscedasticity-robust standard errors, and a three-step hierarchical moderation analysis with mean-centered predictors and simple-slope decomposition. Participants in the experimental condition reported a significantly higher willingness to reduce inappropriate medical practices than those in the control condition (M = 4.07, SD = 0.73 vs. M = 1.93, SD = 0.69; t(118) = 16.46, p < 0.001; mean difference = 2.13, 95% CI [1.88, 2.39]; Cohen’s d = 3.01). Group assignment remained a robust predictor in a multiple regression model adjusting for department, years of practice, perceived medical audit credibility, and perceived institutional frequency of inappropriate practices (B = 2.45, SE = 0.14, β = 0.96, p < 0.001; R² = 0.846). Hierarchical moderation analysis showed that perceived audit credibility significantly moderated the experimental effect (interaction B = 1.15, SE = 0.12, t = 9.44, p < 0.001, 95% CI [0.91, 1.39]; ΔR² = 0.064). Predicted simple-slope group differences increased monotonically from 0.89 points at audit level 2 to 4.33 points at audit level 5. Three converging qualitative themes—institutional tolerance of inappropriate practices, structural rationalization of misconduct, and sensitivity to enforceable sanctions—triangulated with the quantitative findings. The findings provide preliminary, exploratory evidence that a statutory fiduciary duty enforced through credible medical audit mechanisms is associated with substantially greater reported willingness to reduce inappropriate medical practices than ethics-based exhortation alone. The widening gap between conditions as audit credibility increased is consistent with the central prediction of responsive regulation theory: the credibility of escalation at the apex of the regulatory pyramid is what makes cooperative norms at its base operative. Limitations include a small non-probability sample, reliance on self-reported behavioral intentions, the contrast in stimulus specificity between conditions, and the absence of pre-registration. The findings should be regarded as hypothesis-generating, providing a basis for larger pre-registered studies rather than definitive evidence of causal effects.