| Objective:To establish predictive models for the common postoperative complications morbidity and mortality of hip fracture in elderly patients,and to evaluate its predictive value in the surgical risk assessment.Method:1.to establish the data collection table of 4 cases,including postoperative pulmonary infection,postoperative cognitive dysfunction,postoperative deep venous thrombosis and postoperative death;2.collect the clinical data of elderly patients with hip fracture who were operated in the Affiliated Hospital of Southwest Medical University from January 2012 to October 2016;3.and then use Epidata3.1 software to establish the corresponding 4databases and clinical data was entered into the database;the data was imported into spss19.0 software for statistical analysis: variables were analyzed with t and?2 test in order to get statistically significant variables when P<0.05(Taking a=0.05 as the test level).4.On the basis of physiological and operative severity score for the enumeration of mortality and morbidity(POSSUM),we divided thesevariables into physiological index and operative severity index in order to establish a scoring system for postoperative pulmonary infection,postoperative cognitive dysfunction,postoperative deep venous thrombosis and postoperative mortality in elderly patients with hip fracture.We get four models for predicting the incidences of postoperative pulmonary infection,postoperative cognitive dysfunction,postoperative deep venous thrombosis of lower extremity,and postoperative mortality by Logistic regression analysis.5.Finally,we used the ratio between the actual value and the predicted value,the receiver operating characteristic curve(ROC)and the Hosmer-Lemeshow test to evaluate the predictive value of these new models.Results:1.The group of postoperative pulmonary infection : 1)some physiological indexes such as age,WBC count,ASA classification,COPD,heart function classification,number of comorbidities and some operative severity indexes such as preoperative preparation time,operative time,intraoperative blood loss and anesthesia were risk factors of postoperative pulmonary infection in elderly patients with hip fracture.2)Prediction model of risk of postoperative pulmonary infection in elderly patients with hip fracture: Ln[R/(1-R)]=-7.187+0.226×PS+0.161×OS ? 3)The mean predictive value of postoperative pulmonary infection rate was 8.93%.The actual value is9.89%.The ratio between the actual value and the predicted value is1.11.There was no significant difference between the two(?2=0.279,P=0.673 > 0.05).ROC had 82.7% in sensitivity,72.4% in specificity,27.6% in misdiagnosis rate,17.3% in missed diagnosis rate.The area under ROC curve is0.814,which shows that the prediction accuracy is good.The goodness of fit of the model for predicting the risk of postoperative pulmonary infection in elderly patients with hip fracture was analyzed by Hosmer-Lemeshow test.The results show that the prediction model gives a good fit to the data set(H2=7.707,df=8,P=0.463>0.05).2.The group of postoperative deep venous thrombosis of lower extremity : 1)some physiological indexes such as age,FIB,serum triglycerides,BMI,varicose veins,hypertension,coronary heart disease,diabetes,stroke,infection and some operative severity indexes such as preoperative preparation time,operative time,intraoperative blood loss and anesthesia were risk factors of postoperative pulmonary infection in elderly patients with hip fracture.2)Prediction model of risk of postoperative deep venous thrombosis of lower extremity in elderly patients with hip fracture:Ln[R/(1-R)]=-11.493+0.347×PS+0.327×OS ? 3)The mean predictive value of postoperative deep venous thrombosis of lower extremity rate was12.57%.The actual value is 13.38%.The ratio between the actual value and the predicted value is1.06.There was no significant difference between the two(?2=0.144,P= 0.776>0.05).ROC had 74.20% in sensitivity,86.20% in specificity,13.80% in misdiagnosis rate,25.80% in missed diagnosis rate.The area under ROC curve is 0.87,which shows that the prediction accuracy is good.The goodness of fit of the model for predicting the risk of postoperative deep venous thrombosis of lower extremity in elderly patients with hip fracture was analyzed by Hosmer-Lemeshow test.The results show that the prediction model gives a good fit to the data set(H2= 3.309,df=8,P=0.914>0.05).3.The group of postoperative cognitive dysfunction: 1)some physiological indexes such as age,blood pressure(systolic blood pressure),albumin,oxygen partial pressure,number of complications,COPD,stroke and some operative severity indexes such as operative time,intraoperative blood loss and anesthesia were risk factors of postoperative pulmonary infection in elderly patients with hip fracture.2)Prediction model of risk of postoperative cognitive dysfunction in elderly patients with hip fracture: Ln[R/(1-R)]=-6.88+0.191×PS+0.302×OS ? 3)The mean predictive value of postoperative cognitive dysfunction rate was 12.38%.The actual value is 14.28%.The ratio between the actual value and the predicted value is 1.15.There was no significant difference between the two(?2=0.330,P=0.667>0.05).ROC had 53.3% in sensitivity,90% in specificity,10% in misdiagnosis rate,46.7% in missed diagnosis rate.The area under ROC curve is 0.759,which shows that the predictionaccuracy is good.The goodness of fit of the model for predicting the risk of postoperative cognitive dysfunction in elderly patients with hip fracture was analyzed by Hosmer-Lemeshow test.The results show that the prediction model gives a good fit to the data set(H2=7.707,df=8,P=0.463>0.05).4.The group of postoperative mortality : 1)some physiological indexes such as age,white blood cell,albumin,blood pressure(systolic blood pressure),creatinine,ASA classification,cardiac function classification,number of complications,COPD,stroke,diabetes and some operative severity indexes such as preoperative preparation time,operative time,intraoperative blood loss and surgical approach were risk factors of postoperative mortality in elderly patients with hip fracture.2)Prediction model of risk of postoperative mortality in elderly patients with hip fracture: Ln[R/(1-R)]=-11.565 +0.265× PS+0.121×OS.3)The mean predictive value of postoperative mortality was 3.99%.The actual value is 5.18%.The ratio between the actual value and the predicted value is 1.3.There was no significant difference between the two(?2=0.820,P=0.451>0.05).ROC had96.2% in sensitivity,88.8% in specificity,11.2% in misdiagnosis rate,3.8% in missed diagnosis rate.The area under ROC curve is 0.967,which shows that the prediction accuracy is good.The goodness of fit of the model for predicting the risk of postoperative mortality in elderly patients with hip fracture was analyzed by Hosmer-Lemeshowtest.The results show that the prediction model gives a good fit to the data set(H2=10.869,df=8,P=0.209> 0.05).Conclusion : In this study,we established scoring systems and predictive models for the common postoperative complications morbidity and mortality of hip fracture in elderly patients.1)Prediction model of risk of postoperative pulmonary infection in elderly patients with hip fracture: Ln[R/(1-R)]=-7.187+0.226×PS+0.161×OS?2)Prediction model of risk of postoperative deep venous thrombosis of lower extremity in elderly patients with hip fracture:Ln[R/(1-R)]=-11.493 +0.347×PS+0.327×OS?3)Prediction model of risk of postoperative cognitive dysfunction in elderly patients with hip fracture: Ln[R/(1-R)]=-6.88 +0.191× PS + 0.302×OS?4)Prediction model of risk of postoperative mortality in elderly patients with hip fracture: Ln[R/(1-R)]=-11.565+ 0.265×PS+0.121 ×OS?we used the ratio between the actual value and the predicted value,the receiver operating characteristic curve(ROC)and the Hosmer-Lemeshow test to evaluate the predictive value of these new models.The results show that the four models have good accuracy.In order to solve the problem of a wide range of postoperative complications,the postoperative complications of senile hip fractureswere subdivided.We analyzed the common complications and obtained the risk factors of surgical operation,and excluded some risk factors unrelated to the complications.However,the sample size is not enough,and the results of this study are not compared with the PPOSSUM and the POSSUM system in orthopaedic surgery,so further study is needed. |
PDF Full Text Request