骨科患者压疮危险因素调查

Investigate on pressure ulcer risk factors of orthopedic patients

目的探讨骨科患者的压疮相关危险因素，为压疮的防治等提供借鉴。方法以Braden量表为评分标准，分别采用Logistic回归分析和BP神经网络，评价201例骨科患者压疮危险情况，以分析其危险因素。结果不同入院诊断、意识水平、手术状态及是否糖尿病、石膏固定患者Braden分级比较，差异均无统计学意义（P〉0．05），而不同年龄、营养状况、牵引情况、手术情况及是否吸烟、偏瘫、截瘫、人血白蛋白、大便失禁、小便失禁、外固定支架、穿防旋鞋患者Braden等级不同，差异有统计学意义（x2分别为21．82，13．32，39．53，8．08，6．97，14．25，31．82，23．26，5．83，15．46，27．76，9．40；P〈0．05）；Logistic回归分析显示对Braden分级进行判别的总正确率为86．1％（x2=133．31，P〈0．01）；BP神经网络的总正确率为90％（X2=155．67，P〈0．01）。结论结合Logistic回归分析与BP神经网络，年龄、偏瘫、外固定支架及骨牵引与Braden分级的对数优势呈线性关系，其余自变量则偏向于是非线性的。

Objective To discuss the risk factors for pressure ulcer of orthopedic patients and provide reference for prevention and treatment of pressure ulcer. Methods Risk factors for pressure ulcer of 201 orthopedic patients were evaluated by using logistic regression and BP neural network respectively, with Braden rating scale as evaluation standards. Results There was no significant difference of Braden rating among patients of different levels of consciousness, different surgical situations, diabetes or not and different plaster immobilizations （ P 〉 0.05 ）. There were significant differences of Braden rating among patients of different ages, nutritional status, traction situations, surgical outcomes, smoking habits, hemiplegic paralysis, paraplegia, serum albumin, fecal incontinence, urinary incontinence, external fixation, and anti-spin shoes （x2 =21.82,13.32, 39.53,8.08,6.97,14.25,31.82,23.26,5.83,15.46,27.76,9.40, respectively; P 〈 0.05）. Logistic regression analysis showed that the total correct rate of discrimination of the Braden classifications was 86.1% （X2 = 133.31, P 〈 O. O1 ） , while BP neural network＇ s total correct rate was 90% （ X2 = 155. 67, P 〈 O. 01 ）. Conclusions Logistic regression analysis combined with BP neural network shows, age, hemiplegia, external fixation and bone traction have a linear relationship with the logarithm advantage of the Braden classifications and the others have a non-linear relationship.