China’s County Medical Communities Show U-Shaped Effect on Patient Retention

China’s flagship rural healthcare reform may follow a U-shaped path: a new study of tightly integrated county-level medical communities (CCMCs) finds a nonlinear relationship between how intensively these reform networks are implemented and the share of patients who stay in their home county for care. The research, led by Ting Shi and Lan Yang of Lanzhou University of Finance and Economics together with Yuehong Zhang of Gansu University of Political Science and Law, is set to appear in Frontiers in Public Health.

CCMCs are a key initiative of China’s healthcare reform, aiming to optimize the allocation of high-quality medical resources and advance a tiered healthcare delivery system. Yet evidence on whether the policy works has been mixed, and most previous studies treated CCMC construction as a simple yes-or-no intervention, ignoring how much reform has actually been carried out in a given province. To capture that variation, the researchers built a multidimensional CCMC Construction Intensity Index (CII) covering three dimensions: policy coverage, organizational integration, and institutional safeguards.

Rather than a dummy variable marking whether a reform exists, the index measures the degree of implementation on a continuous scale. The authors applied the index to provincial panel data spanning 2014 to 2024, using a two-way fixed-effects model that controls for unobserved differences across provinces and common time trends. The central finding is a U-shaped association between construction intensity and the relative scale of in-county visits.

At lower levels of implementation intensity, patient diversion toward county hospitals accompanies resource integration, drawing patients away from grassroots facilities rather than toward them. Once intensity crosses a threshold, however, substantive integration and long-term institutional mechanisms are gradually unleashed, and patient retention within counties begins to dominate. This pattern is mirrored by an inverted U-shaped trend in care-seeking outside the county, which first rises and then falls as local systems mature.

The authors are careful about the robustness of this curve. The association is attenuated under more demanding trend controls, and the location of the inflection point shows some sensitivity to sample composition. In other words, the threshold effect is a central tendency in the data rather than a precise, universal turning point, and readers should treat the exact crossover level with caution.

Heterogeneity analysis sharpens the policy picture. The U-shaped pattern is more pronounced in western regions of China and in areas with higher private hospital density, indicating that resource constraints and market structure shape how integration unfolds. Within grassroots facilities, the benefits of reform accrue primarily to township health centers, while community stations and village clinics lag behind, showing that the gains from integration have not yet reached the very bottom of the rural care hierarchy.

Theoretically, the study argues that CCMC development follows a threshold-dependent trajectory. By contributing Chinese evidence on nonlinear referral dynamics, the authors position the work within the international literature on integrated care and referral governance, which has focused largely on universal health systems and paid less attention to progressive reforms in developing countries facing resource constraints and patient outflows.

From a policy perspective, the findings support calibrating implementation intensity to local contexts rather than applying a uniform national tempo. They also suggest prioritizing township health centers as the main recipients of resource devolution, so as to mitigate early-stage patient diversion and support the gradual advancement of the tiered healthcare system.

Several caveats frame the conclusions. The association is attenuated under more demanding trend controls, and the inflection point’s sensitivity to sample composition means the threshold should be treated as indicative rather than exact. The analysis also rests on provincial-level panel data for a single country, so the patterns describe aggregate patient flows across China’s provinces during 2014 to 2024.

Even with those caveats, the study offers a quantitative account of how a large-scale integration reform unfolds unevenly over time, and a reminder that the same policy can look successful or failing depending on how far implementation has progressed.

Subject of Research: Nonlinear association between implementation intensity of China's county-level medical communities and patient flow patterns

Article Title: Nonlinear association between provincial implementation intensity of county-level medical communities and patient flow patterns

Article References: Shi, T., Yang, L., & Zhang, Y. (2026). Nonlinear association between provincial implementation intensity of county-level medical communities and patient flow patterns. Frontiers in Public Health, 14, Article 1966297. https://doi.org/10.3389/fpubh.2026.1966297

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1966297

Keywords: China healthcare reform, county-level medical communities, tiered healthcare system, integrated care, patient flow, patient retention, U-shaped relationship, implementation intensity, township health centers, health policy, panel data, referral governance