| Over the last ten years,Acute Care Surgery has grown rapidly as an emerging specialty.The three subspecialties that have emerged are emergency general surgery(EGS),trauma surgery,and critical surgical care.EGS disciplines are currently facing challenges similar to those that trauma surgery faced 50 years ago.The Trauma Society of the American College of Surgeons published the Manual of Trauma Optimization Diagnosis and Treatment in 1976,which defined the risk assessment system for trauma surgery,formulated standardized diagnosis and treatment measures,and stated that trauma centers should constantly optimize the quality of diagnosis and treatment,train trauma surgery specialists,and develop specialized trauma medical systems(such as operating rooms,imaging radiology centers,trauma-related equipment,etc.).These trauma surgery-based practices have significantly improved outcomes in the trauma patient population.The EGS discipline is rapidly evolving in today’s world.With the disease burden of EGS rising year after year,such issues in the EGS discipline construction system must be addressed as soon as possible.This study used clinical research to fill in the gaps in the EGS discipline’s existing disciplinary risk assessment system.Based on single-center EGS,the first part of this study looked at developing and clinically validating a nomogram tool for predicting adverse events in postoperative hospitalization in patients with sepsis in abdominal-related diseases.The purpose of the literature review is to: define the abdominal-related emergency general surgery disease spectrum and comorbid sepsis;retrospectively include patients who met the inclusion criteria and define study-related demographic parameters;preoperative comorbidity parameters;pre-surgical physiological-related parameters;intra-operative surgical-related parameters;and poor postoperative outcomes(in-hospital deaths or postoperative hospital stays more than 15 days,with no death).The indicators with statistically significant differences in univariate analysis(P < 0.05)were included in multivariate logarithmic probability regression analysis.According to the study’s specific implementation,independent risk factors related to poor postoperative hospitalization prognosis were screened.Four variables were analyzed and determined,including age,AAST EGS grade,SOFA score,and surgical approach,and a nomogram model was created.In the validation group,the predictive power of the nomogram model was assessed using the area under the receiver operating characteristic curve,the calibration curve,and the decision benefit curve.The nomogram model’s predictive power in the validation group was compared to that of other tools(SOFA score,APACHE II score,AAST EGS grading).The findings show that the constructed nomogram model can predict the likelihood of a poor postoperative prognosis,suggesting some ideas and practices for developing risk assessment tools for EGS diseases.The second part of this study examines multi-center,large-sample clinical data of patients undergoing emergency surgery for EGS abdominal-related diseases by employing five machine learning algorithms(random forest,log probability regression,support vector machine,limit gradient boosting,naive Bayes)to develop a predictive model for in-hospital death and four postoperative complications(postoperative pneumonia,postoperative incision infection,postoperative thrombosis,and postoperative ventilator use > 48h).According to the area under the curve,sensitivity,specificity,accuracy,precision,and F1 value indicators,the random forest algorithm had the best predictive ability for the five outcome events.Compared to the ESAS score,APACHE II score,and ASA classification,the constructed model performed well.An online calculator based on this model is being developed to provide quantifiable risk probability predictions for clinical diagnosis and treatment of such patients to improve diagnosis and treatment quality. |
PDF Full Text Request