The hierarchical medical system in China aims to guide patients toward primary healthcare institutions for first-contact care. Nevertheless, the proportion of visits at the primary level has continued to decline. The system's actual effectiveness and the mechanisms underlying changes in primary care service volume remain insufficiently understood.

To examine trends in primary care service volume and identify key driving factors after implementation of the hierarchical medical system, thereby providing empirical evidence for policy optimization.

Data were drawn from the China Health Statistical Yearbook series spanning 2005 to 2023. Interrupted time series (ITS) analysis was used to evaluate policy effects, and the logarithmic mean divisia index (LMDI) method was employed to decompose and quantify the contribution of each driving factor.

The ITS model showed that total primary care visits increased at an annual rate of 196 million (P=0.005), yet the proportion declined by 0.67 percentage points per year (P<0.001). At the point of policy intervention (2016), primary care visits fell abruptly by approximately 435 million (P<0.001), with the proportion dropping by 2.14 percentage points (P<0.001). After policy implementation, growth slowed markedly; the annual increase was 168 million fewer than in the pre-intervention period (P<0.001). LMDI results indicated a cumulative change of 2 351 million in primary care visits from 2005 to 2023. The per capita investment effect (EP) and institutional scale effect (SI) were the main positive drivers, while the capital intensity effect (IE, -12 409 million) and physician service productivity effect (PD, -1 551 million) posed major impediments to growth. In the post-implementation period, PD shifted from positive to negative, reflecting weakened growth momentum.

Primary care growth in China exhibits characteristics of an extensive model marked by high input and low efficiency. Resource and workforce productivity have not kept pace with investment increases, which may constrain the patient-diversion capacity of the hierarchical medical system.

分级诊疗制度旨在推动患者首诊下沉至基层医疗卫生机构，但我国基层诊疗服务量占比仍呈下降趋势，政策实际效果及驱动机制有待评估。

探究分级诊疗政策实施后基层医疗服务量的变化趋势及驱动因素，为政策优化提供实证依据。

本研究数据来源于2005—2023年《中国卫生健康统计年鉴》系列资料，采用中断时间序列（ITS）分析和对数平均迪氏指数分解法（LMDI）评估政策效应并量化因素贡献。

ITS模型结果显示，基层总诊疗人次年均增速为1.96亿人次（P=0.005），但占比以年均0.67%的速度下降（P<0.001）。在政策干预时点（2016年），基层总诊疗人次瞬时下降约4.35亿人次（P<0.001），占比下降2.14%（P<0.001）。政策实施后，基层总诊疗人次增长趋势减缓，年增长速度较政策干预前下降1.68亿人次（P<0.001）。LMDI分解结果显示，2005—2023年基层服务量总体变化量为23.51亿人次，人均投入效应（EP）与机构规模效应（SI）为主要正向驱动因素，但资本密集度效应（IE）为-124.09亿人次、医师服务产出率效应（PD）为-15.51亿人次，构成明显增长阻力。分阶段来看，政策实施后PD由正值转为负值，增长动力较政策实施前有所减弱。

当前基层服务增长呈现出一定的高投入、低效率的特征，资源与人员效率尚未同步提升，可能制约了分级诊疗的分流效果。