Fernandez Arjona M, Herruzo Cabrera R, Gomez-Sancha F, Calero Rey J
Department of preventive medicine, Hospital la Paz, Universidad autonoma de Madrid, Spain.
Minerva Med. 1996 May;87(5):189-94.
We carried out a four-year prospective study in the rehabilitation and traumatology center of the Hospital La Paz, where a total of 5260 patients were included. The objective of the study was to know the main risk factors that could influence the development of an infection of surgical wound, and get a predictive equation of infection of surgical wound in our patients. We utilized logistic regression for it, following the patient by means of active epidemiological surveillance; the main risk factors found were: immunodeficiency (OR = 8.3) wrong scaring (OR = 14.4), more than one intervention (OR = 3.5), type of surgical intervention (OR = 4.8) and incorrect use of prophylaxis (OR = 6.3). We considered that this knowledge could permit us to diminish the incidence of infections, specially if there is some of these factors, like the antimicrobial prophylaxis, that could be easily modified. Finally, upon applying to our patients the gotten equation, we get a ROC curve with 85% of the area under it, and if we take the point of cut of greater sensibility and specificity in this situation of very low prevalence of illness, it determines we could use it better in order to mark patients that don't suffer infection (high negative predictive value), making this possible to improve the efficacy of our work.
我们在拉巴斯医院的康复与创伤中心开展了一项为期四年的前瞻性研究,共纳入5260例患者。该研究的目的是了解可能影响手术伤口感染发生的主要危险因素,并得出我们患者手术伤口感染的预测方程。我们为此采用逻辑回归分析,通过主动流行病学监测对患者进行随访;发现的主要危险因素有:免疫缺陷(比值比=8.3)、缝合不当(比值比=14.4)、不止一次干预(比值比=3.5)、手术干预类型(比值比=4.8)以及预防措施使用不当(比值比=6.3)。我们认为,这些知识能够使我们降低感染发生率,特别是在存在某些如抗菌预防措施等可轻易改变的因素时。最后,将所得方程应用于我们的患者时,我们得到了一条曲线下面积为85%的受试者工作特征曲线,并且在疾病患病率极低的这种情况下,如果我们选取具有更高敏感性和特异性的截断点,就可以确定我们能更好地利用它来标记未发生感染的患者(高阴性预测值),从而提高我们工作的效率。