Analysis Of Hospitalization Costs And Influencing Factors Of Neurosurgery Patients With Hospital Acquired Infections

Objective: To investigate the status of hospital acquired infection in neurosurgical patients,analyze the hospitalization cost of neurosurgical patients,and explore the risk factors of hospital acquired infection,and construct a prediction model to provide a realistic basis and scientific guidance for the prevention and control of hospital acquired infections in these patients.Method:(1)Data collection.A survey form was designed to collect patients data through the Hospital Information System and Hospital Infection Surverillance System.A total of 1440 neurosurgery patients who were admitted between January 1,2018,and December 31,2021 and met the inclusion/exclusion criteria were included in this study,with 251 patients in the infection group and 1189 patients in the control group.The survey content included patient basic information,surgery-related information,hospital acquired infection information,invasive procedures,and the hospitalization cost of neurosurgical patients.(2)Hospital acquired infection status.Quantitative data were described using mean ±standard deviation(x± SD)or median(M)and interquartile range(IQR),and qualitative data were described using the number(n)and percentage(%).The analysis content included the hospital aquired infection situation and hospitalization cost situation.(3)Analyze the hospitalization cost of neurosurgical patients.The description of hospitalization costs for the infection and control groups were based on the median.The MannWhitney U test was used to compare the costs between the two groups.The direct economic loss was the extra hospitalization expenses caused by hospital acquired infections,and the indirect economic loss was calculated based on the extended hospitalization time and the average daily income of Chinese residents.The translation of the above text should meet academic requirements.(4)Analysis of influence factors of hospital acquired infections.Univariate analysis used the chi-square test and t-test.A P-value <0.05 was considered statistically significant.Multivariate analysis used unconditional logistic regression to determine the independent factors influencing hospital acquired infections.The test was a two-tailed test with a significance level of α=0.05.(5)Construction of a hospital acquired infection model.In this study,based on machine learning methods,principal component analysis was used for data dimensionality reduction,and the logistic regression model in the strong classification model was used for model construction.The SMOTE algorithm was used to oversample the positive samples to form a new dataset.Then,the dataset was split into a training set and a test set at a ratio of 8:2.In the training set,five-fold cross-validation was used,and the hyperparameters of the logistic regression model were adjusted using grid search to prevent overfitting.The trained model was applied to the test set to obtain a confusion matrix,and the recall rate,accuracy,precision,and F2-score were calculated to evaluate the model results.The receiver operating characteristic curve was plotted,and the area under the curve was calculated to evaluate the model’s performance.The translation of the above text should meet academic requirements.Result:(1)The situation of hospital acquired infection in neurosurgery patients.In this study,there were 251 cases of infection,with an infection rate of 17.43% and an infection incidence rate of 18.47%.The top three rates of hospital acquired infections were lower respiratory tract infections 53.39%(134 cases),organ(or cavity)infections 23.90%(60 cases)and sepsis 4.38%(11 cases).A total of 493 patients underwent microbiological examination and 188 strains of pathogenic bacteria were detected,with a detection rate of 38.13%,of which 73.40%(138 strains)and 26.60%(50 strains)were Gram-negative and positive bacteria,respectively,The top three detection rates of pathogenic bacteria were Acinetobacter baumannii at 39.36%(74 strains),Staphylococcus aureus at 17.02%(32 strains)and Klebsiella pneumoniae at 9.04%(17 strains).(2)Medical expense status of hospital acquired infection in neurosurgery patients.Among the1440 study subjects,the proportion of patients who paid medical expenses through the New Rural Cooperative Medical Care was the highest at 49.86%,while the proportion of patients covered by the Basic Medical Insurance for Urban Residents was the lowest at 8.13%.The median total hospitalization cost was CNY48,136.48(26,828.77 to 74,137.57),with the top three expenses in order of CNY11,334.57(1,764.38 to 20,051.69)for consumables costs,CNY7,180.50(3,050.5 to11,054.75)for surgical treatment fee and CNY2,920.00(1,610.00 to 4,581.75)for imaging diagnostic fees,and a minimum of CNY58(CNY18.00 to 196.00)for rehabilitation fees.(3)Analyze the hospitalization cost of neurosurgery patients.Patients with hospital acquired infection were used as the infection group and those without hospital acquired infection as the control group to compare the hospitalization costs using the Mann-Whitney U test.The results showed that the differences in total hospitalization costs,general medical service fees,general treatment service fees,nursing fees,laboratory diagnostic fees,imaging diagnostic fees,clinical diagnostic project fees,non-surgical treatment fees,surgical treatment fees,rehabilitation fees,traditional Chinese medicine treatment fees,antimicrobial drug costs,and consumables costs were all statistically significant(P<0.001).The direct economic burden of hospital acquired infection was CNY 50,777.69,with a growth rate of 117%.The top three direct economic burdens were antimicrobial drug costs(CNY 7,360.57),nursing fees(CNY 7,090.00),and general treatment service fees(CNY 4,634.20).By comparing the growth rates of various expenses,the top three expenses with the highest increase were antimicrobial drug costs(1080%),non-surgical treatment fees(506%),and nursing fees(423%).The average length of hospital stay for the infection group(28.98±15.72 days)was significantly longer than that of the control group(16.16±10.03 days)with a statistically significant difference(P<0.001).The average extended hospitalization days caused by hospital infections in neurosurgery patients was 12.82 days,charge for loss of working time is CNY 603.64.(4)Risk factors for hospital acquired infections in neurosurgery patients.Univariate analysis showed statistically significant results for time of admission,gender,hypertension,tumour history,route of admission,GCS score,antimicrobial use,length of stay,days with fever,doctor in charge,ASA,cleanliness of incision,duration of surgery,type of surgery,indwelling central venous cannula,mechanical ventilation and indwelling catheter.Seven independent risk factors were screened by unconditional logistic regression,doctor in charge(OR=1.035,95% CI: 1.010 to1.060),indwelling catheter(OR=2.769,95% CI: 1.459 to 5.255),days with fever(OR=1.129,95% CI: 1.083 to 1.176),length of stay(OR=1.052,95% CI: 1.036 to 1.068),hypertension(OR=1.482,95% CI: 1.030 to 2.132),tumour history(OR=2.195,95% CI: 1.124 to 4.283),ASA=III(OR=3.393,95% CI.1.475 to 7.806).(5)Construction of a hospital acquired infection prediction model.In this study,principal component analysis was used to reduce the dimensionality of the categorical variables with statistical significance(P < 0.05)in the univariate analysis.The results showed that there were 5principal components that satisfied the condition of feature value > 1.Principal components F1-F5 and days of fever and length of stay were selected as the model indicators.A logistic regression model was constructed using the strong classification model’s logic regression model.In this study,there were 251 positive samples(infection group)and 1189 negative samples(control group).Due to the imbalance of the sample distribution,the SMOTE algorithm was used to oversample the positive samples.A new dataset with a sample size of 2,378 was obtained,which was divided into a training set and a test set in an 8:2 ratio.The model was validated using 5-fold cross-validation on the training set,and the hyperparameters of the model were adjusted using grid search to prevent overfitting.The final parameters selected were alpha=0 and lambda=0.05.The training set model was applied to the test set to obtain the confusion matrix.Recall rate,accuracy,precision,and F2-score were used to evaluate the model’s results,with recall rate of 0.77,accuracy of 0.77,precision of 0.72,and F2-score of 0.75.The logistic(P)model was 0.138+0.639×feverdays+0.761×LOS+0.512×F1-0.219×F2+0.241×F3+0.113×F4+0.077×F5,where feverdays is days of fever,and LOS is the length of stay.Conclusion:(1)Lower respiratory tract infections account for the highest proportion of hospital infections in neurosurgery patients,with the most commonly detected strains being Acinetobacter baumannii,Staphylococcus aureus,and Klebsiella pneumoniae;(2)Hospital infections can lead to prolonged hospitalization and increased hospitalization cost for neurosurgery patients;(3)Length of stay,days of fever,indwelling catheter,doctor in charge,hypertension,tumours history,and ASA=III are independent risk factors for hospital acquired infections in neurosurgery patients;(4)Based on machine learning methods,a logistic regression prediction model that was constructed by variables of length of stay,days of fever and so on,has good predictive ability for hospital acquired infections in neurosurgery patients.