目的 考察金黄色葡萄球菌耐药率与抗菌药物使用强度（AUD）的相关性，建立可用于预测金黄色葡萄球菌耐药率的回归模型。方法 回顾性分析2014年第1季度至2020年第4季度信阳市中心医院金黄色葡萄球菌耐药率监测数据以及同期抗菌药物AUD数据。采用2014年第1季度至2019年第4季度数据作为试验样本数据，采用线性相关性分析法分析两者的相关性。以耐药率为因变量，抗菌药物AUD为自变量，对相关性显著的变量建立线性回归模型。以2020年第1季度至2020年第4季度数据为考核样本，应用所建立的回归模型预测金黄色葡萄球菌耐药率，并与同期耐药率实际值对比以验证模型预测的有效性。结果 金黄色葡萄球菌对大部分抗菌药物耐药率随时间呈下降趋势（P<0.05），对红霉素和阿奇霉素耐药率随时间无显著变化趋势。金黄色葡萄球菌对庆大霉素耐药率与头孢唑林AUD呈正相关（r=0.431，P=0.036），与阿奇霉素AUD呈正相关（r=0.523，P=0.009），与左氧氟沙星AUD呈正相关（r=0.606，P=0.002）。对左氧氟沙星耐药率与头孢唑林AUD呈正相关（r=0.447，P=0.029），与左氧氟沙星AUD呈正相关（r=0.482，P=0.017）。耐甲氧西林金黄色葡萄球菌（MRSA）检出率与左氧氟沙星AUD呈正相关（r=0.639，P=0.001）。分别对庆大霉素耐药率、左氧氟沙星耐药率、MRSA检出率建立一元或多元线性回归模型，结果显示模型均通过显著性检验（F₁=7.416，P₁=0.004；F₂=6.317，P₂=0.007；F₃=11.65，P₃=0.002）。将2020年第1季度至2020年第4季度抗菌药物AUD数据代入相应的回归模型，得到的耐药率预测值在实际值上下浮动，耐药率实际值均处于95%预测区间以内。结论 信阳市中心医院金黄色葡萄球菌耐药率与抗菌药物AUD存在相关性，所建立的线性回归模型可初步反映两者的量化关系，可对耐药率进行有效预测。

Objective The correlation between Staphylococcus aureus drug resistance rate and antibiotic use intensity (AUD) was investigated, and a regression model for predicting staphylococcus aureus drug resistance rate was established. Methods The monitoring data of Staphylococcus aureus drug resistance rate in Xinyang Central Hospital from the first quarter of 2014 to the fourth quarter of 2020 and AUD data of antibiotics in the same period were retrospectively analyzed. The data from the first quarter of 2014 to the fourth quarter of 2019 were used as the test sample data, and the correlation between them was analyzed by linear correlation analysis. Taking drug resistance rate as dependent variable and antibiotic AUD as independent variable, a linear regression model was established for variables with significant correlation. Taking the data from the first quarter of 2020 to the fourth quarter of 2020 as the evaluation sample, the established regression model was used to predict the drug resistance rate of Staphylococcus aureus, and compared with the actual drug resistance rate of the same period to verify the validity of the model prediction. Results The drug resistance rates of Staphylococcus aureus to most antibacterial drugs showed a decreasing trend over time (P < 0.05), while the drug resistance rates to erythromycin and azithromycin showed no significant trend over time. The drug resistance rate of Staphylococcus aureus to gentamicin was positively correlated with cefazolin AUD (r=0.431, P=0.036), azithromycin AUD (r=0.523, P=0.009), levofloxacin AUD (r=0.606, P=0.002). The drug resistance rate to levofloxacin was positively correlated with cefazolin AUD (r=0.447, P=0.029), and with levofloxacin AUD (r=0.482, P=0.017). The detection rate of methicillin- resistant Staphylococcus aureus (MRSA) was positively correlated with levofloxacin AUD (r=0.639, P=0.001). Unitary or multiple linear regression models were established for gentamicin resistance rate, levofloxacin resistance rate, and MRSA detection rate, respectively, and the results showed that the models passed the significance test (F₁=7.416, P₁=0.004; F₂=6.317, P₂=0.007; F₃=11.65, P₃=0.002). The AUD data of antibiotics from the first quarter of 2020 to the fourth quarter of 2020 were substituted into the corresponding regression model, and the predicted values of drug resistance rate obtained fluctuated around the actual values, and the actual values of drug resistance rate were all within the 95% prediction range. Conclusion The drug resistance rate of Staphylococcus aureus in Xinyang Central Hospital was correlated with AUD. The established linear regression model could preliminarily reflect the quantitative relationship between the two and effectively predict the drug resistance rate.

R978.1