Development of a surgical site infection prediction model in orthopaedic trauma: The Denver Health Model

-A machine learning model to predict surgical site infection after surgery of lower extremity fractures

PURPOSE

This study aimed to develop machine learning algorithms for identifying predictive factors associated with the risk of postoperative surgical site infection in patients with lower extremity fractures.

METHODS

A machine learning analysis was conducted on a dataset comprising 1,579 patients who underwent surgical fixation for lower extremity fractures to create a predictive model for risk stratification of postoperative surgical site infection. We evaluated different clinical and demographic variables to train four machine learning models (neural networks, boosted generalised linear model, naïve bayes, and penalised discriminant analysis). Performance was measured by the area under the curve score, Youdon's index and Brier score. A multivariate adaptive regression splines (MARS) was used to optimise predictor selection.

RESULTS

The final model consisted of five predictors. (1) Operating room time, (2) ankle region, (3) open injury, (4) body mass index, and (5) age. The best-performing machine learning algorithm demonstrated a promising predictive performance, with an area under the ROC curve, Youdon's index, and Brier score of 77.8%, 62.5%, and 5.1%-5.6%, respectively.

CONCLUSION

The proposed predictive model not only assists surgeons in determining high-risk factors for surgical site infections but also empowers patients to closely monitor these factors and take proactive measures to prevent complications. Furthermore, by considering the identified predictors, this model can serve as a reference for implementing preventive measures and reducing postoperative complications, ultimately enhancing patient outcomes. However, further investigations involving larger datasets and external validations are required to confirm the reliability and applicability of our model.

Development and Validation of a Model for Predicting Surgical Site Infection After Pelvic Organ Prolapse Surgery

IMPORTANCE

Surgical site infection (SSI) is a common and costly complication. Targeted interventions in high-risk patients may lead to a reduction in SSI; at present, there is no method to consistently identify patients at increased risk of SSI.

OBJECTIVE

The aim of this study was to develop and validate a model for predicting risk of SSI after pelvic organ prolapse surgery.

STUDY DESIGN

Women undergoing surgery between 2011 and 2017 were identified using Current Procedural Terminology codes from the Centers for Medicare and Medicaid Services 5% Limited Data Set. Surgical site infection ≤90 days of surgery was the primary outcome, with 41 candidate predictors identified, including demographics, comorbidities, and perioperative variables. Generalized linear regression was used to fit a full specified model, including all predictors and a reduced penalized model approximating the full model. Model performance was measured using the c-statistic, Brier score, and calibration curves. Accuracy measures were internally validated using bootstrapping to correct for bias and overfitting. Decision curves were used to determine the net benefit of using the model.

RESULTS

Of 12,334 women, 4.7% experienced SSI. The approximated model included 10 predictors. Model accuracy was acceptable (bias-corrected c-statistic [95% confidence interval], 0.603 [0.578-0.624]; Brier score, 0.045). The model was moderately calibrated when predicting up to 5-6 times the average risk of SSI between 0 and 25-30%. There was a net benefit for clinical use when risk thresholds for intervention were between 3% and 12%.

CONCLUSIONS

This model provides estimates of probability of SSI within 90 days after pelvic organ prolapse surgery and demonstrates net benefit when considering prevention strategies to reduce SSI.

Machine-Learning Models for Predicting Surgical Site Infections using Patient Pre-Operative Risk and Surgical Procedure Factors

BACKGROUND

Surgical site infections (SSIs) are a significant healthcare problem as they can cause increased medical costs and increased morbidity and mortality. Assessing a patient's pre-operative risk factors can improve risk stratification and help guide the surgical decision-making process. Previous efforts to use pre-operative risk factors to predict the occurrence of SSIs have relied upon traditional statistical modeling approaches. The aim of this paper is to develop and validate, using state-of-the-art machine learning (ML) approaches, classification models for the occurrence of SSI to improve upon previous models.

METHODS

In this work, using the American College of Surgeons' National Surgical Quality Improvement Program (ACS NSQIP) database, the performances (e.g., prediction accuracy) of seven different ML approaches (Logistic Regression (LR), Naïve Bayesian (NB), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), Artificial Neural Network (ANN), and Deep Neural Network (DNN)) were compared. The performance of these models was evaluated using the area under the curve, accuracy, precision, sensitivity, and F1-score metrics.

RESULTS

Overall, 2,882,526 surgical procedures were identified in the study for the SSI predictive models' development. The results indicate that the DNN model offers the best predictive performance with 10-fold compared to the other six approaches considered (area under the curve = 0.8518, accuracy = 0.8518, precision = 0.8517, sensitivity = 0.8527, F1-score = 0.8518). Emergency case surgeries, American Society of Anesthesiologists (ASA) Index of 4 (ASA_4), BMI, Vascular surgeries, and general surgeries were most significant influencing features towards developing an SSI.

CONCLUSION

Equally important is that the commonly used LR approach for SSI prediction displayed mediocre performance. The results are encouraging as they suggest that the prediction performance for SSIs can be improved using modern ML approaches.

Risk Factors for Gram-Negative Fracture-Related Infection

In this study, we evaluated risk factors for gram-negative fracture-related infection in a mixed cohort of gram-positive and gram-negative fracture-related infections to guide perioperative antibiotic prophylaxis for surgical fixation of fractures. We performed a retrospective review of all patients with fracture who were treated at an urban academic level I trauma center between February 1, 2012, and June 30, 2017. Inclusion criteria were as follows: (1) open or closed fracture with internal fixation; (2) deep, acute to subacute (<6 weeks), culture-positive fracture-related infection; and (3) age 18 years or older. Infections were classified as gram positive, gram negative, or polymicrobial. Demographic, surgical, and postoperative characteristics were compared among groups. Of 3360 patients, 43 (1.3%) had a fracture-related infection (15 gram negative, 14 gram positive, and 14 polymicrobial). Risk factors for gram-negative infection included initial external fixation (P=.038), the need for soft tissue coverage of an open fracture site (P=.039), lower albumin level at the time of infection (P=.005), and hospitalization for longer than 10 days (P=.018). Perioperative gram-negative antibiotic prophylaxis for fracture fixation surgery should be considered for those who have been staged with external fixation, require soft tissue coverage, are at risk for malnutrition in the postoperative period, and have prolonged inpatient hospitalization. [Orthopedics. 2022;45(2):91-96.].

Risk factors for treatment failure of fracture-related infections

OBJECTIVE

Infection after fracture fixation is a potentially devastating outcome, and surgical management is frequently unsuccessful at clearing these infections. The purpose of this study is to determine if factors can be identified that are associated with treatment failure after operative management of a deep surgical site infection.

METHODS

We retrospectively reviewed the billing system at a Level I trauma center between March 2006 and December 2015. We identified 451 patients treated for deep surgical site infection after fracture fixation at our center. A multivariate regression analysis was then performed to evaluate for factors associated with treatment failure.

RESULTS

Mean follow-up was 2.3 years. One hundred fifty-six patients (35%) failed initial surgical management. Risk factors associated with treatment failure included initial culture results positive for polymicrobial organisms (odds ratio [OR], 1.6; 95% confidence interval [CI], 1.0-2.4), removal of implants (OR, 1.9; 95% CI, 1.2-2.9), or Gustilo-Anderson IIIB/IIIC injury (OR, 2.0; 95% CI, 1.1-3.7). Increased body mass index and fulfilling the criteria to have a methicillin-resistant Staphylococcus aureus (MRSA) nasal swab screening showed a trend toward increased risk of failure.

CONCLUSION

Treatment failure after deep surgical site infection was relatively common. Three distinct factors (polymicrobial infection, removal of implants, and IIIB/C fracture) were associated with failure to eradicate the infection in the first series of surgeries and antibiotics. These data might help guide clinicians as they counsel patients on the risk of treatment failure and might focus efforts to improve treatment toward patients at higher risk of treatment failure.

Surgical Site Infections After Routine Syndesmotic Screw Removal: A Systematic Review

OBJECTIVES

To investigate the incidence of surgical site infections (SSIs) after routine removal of syndesmotic screws (SSs) placed to stabilize syndesmotic injuries.

DATA SOURCES

A systematic literature search was performed in the PubMed, Cochrane, and EMBASE databases for studies published online before February 2020, using the key words and synonyms of "syndesmotic screw" ("ankle fractures" or "syndesmotic injury") and "implant removal."

STUDY SELECTION

Studies were eligible for inclusion when they described >10 adult patients undergoing elective/scheduled removal of the SS.

DATA EXTRACTION

The 15 included articles were assessed for quality and risk of bias using the Newcastle-Ottawa Scale. Baseline characteristics of the studies, the study population, the intervention, the potential confounders, and the primary outcome (% of SSIs) were extracted using a customized extraction sheet.

DATA SYNTHESIS

The primary outcome was presented as a proportion of included patients and as a weighted mean, using inverse variance, calculated in RStudio. Furthermore, potential confounders were identified.

CONCLUSIONS

The percentage of SSIs ranged from 0% to 9.2%, with a weighted mean of 4%. The largest proportion of these infections were superficial (3%, 95% confidence interval: 2-5), compared with 2% deep infections (95% confidence interval: 1-4). These rates were comparable to those of other foot/ankle procedures indicating that the individual indication for SS removal (SSR) should be carefully considered. Future studies should focus on valid indications for SSR, the influence of prophylactic antibiotics on an SSI after SSR, and complications of retaining the SS to enable a fair benefits/risks comparison of routine versus on-demand removal of the SS.

LEVEL OF EVIDENCE

Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

Fatores de risco para infecção de sítio cirúrgico em cirurgias traumato-ortopédicas

Introdução: A Infecção do Sítio Cirúrgico ocupa a terceira posição entre as infecções adquiridas em serviços de saúde, configurando-se como um importante problema de saúde pública, presente em cerca de 15% daquelas encontradas em pacientes hospitalizados, e um dos tipos de cirurgias de maior probabilidade para esse evento são as traumato-ortopédicas. Objetivo: Verificar a associação entre os fatores de risco e a presença de Infecção de Sítio Cirúrgico em cirurgias traumato-ortopédicas. Materiais e métodos: Estudo prospectivo descritivo, quantitativo, realizado entre agosto a outubro de 2017 com 84 pacientes. Os dados referentes ao paciente, ao procedimento e os sinais de Infecções de Sítio Cirúrgico, encontrados no terceiro e décimo dia de pós-operatório, foram coletados por meio de um formulário. Resultados: Das variáveis estudadas, comportaram-se como fatores de risco de Infecção de Sítio Cirúrgico: tabagismo, diabetes e idade acima de 50 anos. A idade foi o principal fator de risco relacionado ao paciente, presente em 31, 36,9% dos casos. Dor, edema e hiperemia foram os sinais de infecção mais prevalentes. Discussão e conclusões: Os fatores de risco relacionados ao paciente possuem maior relevância na associação com infecção do sítio cirúrgico em comparação com os relacionados ao procedimento.

Surgical site infection following traumatic orthopaedic surgeries in geriatric patients: Incidence and prognostic risk factors

Geriatric population is increasing rapidly worldwide, and fragility fracture and complication following orthopaedic surgery in elderly people have now become major challenges for surgeons. Further studies are required to identify potentially modifiable factors associated with surgical site infection (SSI) in geriatric patients. This retrospective, multicenter study was conducted at four level I hospitals in China. During the 31-month study period, a total of 2341 patients (65 years or older) underwent orthopaedic surgery and complete data were recorded from September 2015 to April 2018. Demographics information, medications and additional comorbidities, surgery-related variables, and laboratory indexes were extracted and analysed. Receiver-operating characteristic analysis was performed to detect the optimum threshold of continuous variables. Independent risk factors of SSI were identified by univariate and multivariate analyses. Finally, 63 patients suffered from wound infection within the follow-up period, indicating a 2.7% incidence rate of SSI. Statistical results showed that open injury (odds ratio [OR], 9.5; 95% confidence interval [CI], 5.4-16.7), American Society of Anesthesiologists classified III-IV score (OR, 2.2; 95% CI, 1.3-3.8), surgical duration of >132 minutes (OR, 2.9; 95% CI, 1.1-5.0), serum albumin (ALB) of <36.4 mg/L (OR, 2.0; 95% CI, 1.6-3.4), and blood glucose (GLU) of >118 mg/dL (OR, 3.1; 95% CI, 1.1-5.3) were independent risk factors of postoperative SSI. With the application of sensitive and modifiable variables such as surgical duration and the levels of ALB and GLU, more geriatric patients with sub-high risk of postoperative SSI could be identified.

Microbiological profile and susceptibility pattern of surgical site infections related to orthopaedic trauma

BACKGROUND

Understanding the epidemiology of microorganisms associated with surgical site infections related to orthopaedic trauma (SSI-ROT) is important in establishing treatment protocols. The aim of this study was to evaluate the etiology and susceptibility pattern of SSIs related to orthopaedic trauma in a Brazilian reference hospital for trauma.

METHODS

Patients with SSI-ROT in a Brazilian reference hospital for trauma were retrospectively analyzed. All patients with orthopaedic trauma who underwent a surgical procedure and developed SSI within one year were included. All patients had culture samples from the surgical site obtained from biopsy of bone or soft tissue. Clinical and epidemiological data of the patients were collected.

RESULTS

A total of 147 patients with trauma-related infection were included in the analysis. The mean time to infection was 55.5 days, and the mean duration of hospitalization was 20.0 days. The in-hospital mortality rate after infection was 5.4%. Cultures were obtained from all patients, with 104 samples obtained from soft tissues and 43 samples from bone. The positivity rate was 93.2%. Among the isolates, 56.5% (77 patients) were gram-negative bacteria and 43.8% (60 patients) were gram-positive bacteria. Staphylococcus aureus was identified in 34%, Enterobacter spp. in 14.9%, and Pseudomonas aeruginosa in 11.6%. Staphylococcus aureus presented a higher positivity in bone samples (odds ratio, 1.29; 95% CI, 1.01-1.70; p = 0.04). Few microorganisms were multi-resistant.

CONCLUSION

SSI in orthopaedic trauma can be associated with gram-negative bacilli, the susceptibility profile of which suggested that most infections occur after discharge. Staphylococcus aureus infections were commonly caused by methicillin-susceptible isolates, and this susceptibility to oral antibiotic options helps in the dehospitalization of patients.