Objective To build a multi‐level hierarchical structure model of the influencing factors for hierarchical medical system ,to identify the role relationship between all the factors and transmission pathways ,and to recommend on developing China′s hierarchical medical system . Methods Thirty influencing factors were identified in a screening based on literature review for the hierarchical medical system .On such basis ,16 influencing factors were identified by three health policy experts ,which affect operations of the current system .Interpretative structural modeling was called into play in the end to analyze the hierarchy relationship between various influencing factors and the conduction loops .Results There exist among the 16 factors a 3‐level hierarchical structureand two conduction loops .The factor directly limiting the hierarchical medical system is two‐way referral,and most internal core drivers arehuman resources development and governance mechanism.By means of self‐growth and external constraints ,they exert their influence on the operation of hierarchical medical system .Conclusions There are interactive hierarchical effects among the factors ,merging into three node elements of functional role,inter‐entity relationshipand patient participation.

Objective:To investigate the effect of gap junction protein Cx43 inhibitor carbenoxolone (CBX) on cognitive function and its possible mechanism in epileptic rats.Methods:One hundred and twenty Wistar rats were randomly divided into sham-operated group, epilepsy group, epilepsy+solvent group, and epilepsy+CBX group ( n=30). The models of temporal lobe epilepsy in the later three groups were prepared by injection of kainic acid in the hippocampus. Intraperitoneal injection of CBX (20 mg/kg) or equal amount of normal saline were given to the rats in the epilepsy+CBX group and epilepsy+solvent group 30 min before modeling. Western blotting was used to detect the protein expressions of phosphorylated (p)-Cx43 and microtubule associated protein light chain 3 (LC3) in the hippocampus 6, 12, and 24 h after modeling; the protein localization of p-Cx43 and LC3 in the hippocampus and optical density of their positive cells were detected by immunohistochemistry 24 h after modeling; the learning and memory abilities of rats were tested by Morris water maze experiment 30 d after modeling. Results:Western blotting results showed that as compared with those in the sham-operated group, p-CX43 and LC3 protein expressions in the hippocampal CA3 regions of epilepsy group and epilepsy+solvent group were significantly increased at 6, 12 and 24 h after modeling ( P<0.05); as compared with the epilepsy group and epilepsy+solvent group, the epilepsy+CBX group had statistically decreased p-CX43 and LC3 protein expressions in the hippocampal CA3 regions at each time point ( P<0.05). Immunohistochemical staining showed that p-CX43 was localized at the cell membrane and cytoplasm of hippocampal astrocytes; LC3 was located at the cytoplasm of hippocampal neurons. As compared with those in the sham-operated group, the optical density values of p-CX43 and LC3 positive cells in hippocampal CA3 regions of epilepsy group and epilepsy+solvent group were increased ( P<0.05). As compared with those in the epilepsy group and the epilepsy+solvent group, the optical density values of p-CX43 and LC3 positive cells in the hippocampal CA3 regions of the epilepsy+CBX group were significantly decreased ( P<0.05). Morris water maze test results showed that as compared with that in the sham-operated group, the escape latency in the epilepsy group and epilepsy+solvent group was significantly prolonged ( P<0.05); as compared with that in the epilepsy group and epilepsy+solvent group, the latency in the epilepsy+CBX group was significantly shortened ( P<0.05). Conclusion:CBX can weaken the neuronal autophagy and reduce the damage to cognitive function by inhibiting the p-Cx43 protein expression in the astrocytes of the hippocampal CA3 regions.