A study on the regional distribution of 30-day readmission patients with stroke in Ningxia and its correlation with medical and health resources
宁夏脑卒中30天再入院患者区域分布及其与医疗卫生资源的相关性研究

摘要

目的 分析宁夏22个县区脑卒中患者30天再入院率的区域分布特征，探讨其与医疗卫生资源的相关性，为优化资源配置提供依据。方法 收集2022年宁夏二级及以上综合医疗机构病案首页数据，计算各县区脑卒中患者30天再入院率。采用Shapiro-Wilk检验评估数据正态性，对偏态资料进行自然对数转换。采用Pearson相关分析各卫生资源指标与转换后再入院率的相关性，进一步采用逐步多元线性回归筛选独立影响因素。结果 22个县区原始再入院率呈偏态分布，经对数转换后满足正态性要求（P=0.212）。单因素分析显示，每万人口医疗机构数与转换后再入院率呈显著正相关（r=-0.499，P=0.018）；其余指标无统计学意义。逐步回归仅保留每万人口医疗机构数进入模型（B=-0.090，P=0.004），模型可解释再入院率变异的34.0%。结论 宁夏脑卒中患者30天再入院率与医疗机构密度呈显著正相关，医疗资源集中区域的再入院率更高，可能反映该类地区收治患者病情复杂程度较高。建议在评价再入院率时充分考虑区域医疗中心的功能定位，并加强基层服务能力建设。

Abstract

Objective To analyze the regional distribution characteristics of 30-day readmission rates of stroke patients in 22 counties and districts of Ningxia, and to explore their correlation with medical and health resources, in order to provide a basis for optimizing resource allocation.
Methods The data of the first page of medical records from secondary and above general medical institutions in Ningxia in 2022 were collected. The 30-day readmission rates of stroke patients in each county were calculated. The Shapiro-Wilk test was used to assess the normality of the data, and the natural logarithm transformation was performed on the skewed data. Pearson correlation analysis was used to explore the correlation between each health resource indicator and the transformed readmission rate. Further, stepwise multiple linear regression was used to screen independent influencing factors.
Results The original 30-day readmission rates in 22 counties showed a skewed distribution (median 3.65%, range 1.71%-8.82%), and after logarithmic transformation, they met the normality requirements (P = 0.212). Univariate analysis showed that the number of medical institutions per 10,000 population was significantly negatively correlated with the transformed readmission rate (r =0.583, P = 0.004); the number of primary medical institutions per 10,000 population was also significantly positively correlated with the transformed readmission rate (r=0.499, P=0.018). Stepwise regression retained only the number of medical institutions per 10,000 population in the final model (B=0.090, P=0.004), and the model could explain 30.7% of the variation in the readmission rate.
Conclusion   The 30-day readmission rate of stroke patients in Ningxia is significantly positively correlated with the density of medical institutions. Regions with concentrated medical resources have higher readmission rates, which may reflect the higher complexity of patients admitted in such areas. It is recommended to fully consider the functional positioning of regional medical centers when evaluating readmission rates, and to strengthen the construction of primary service capabilities.