Identification Of Surgical Site Infections Research Articles

An explainable long short-term memory network for surgical site infection identification

BackgroundCurrently, surgical site infection surveillance relies on labor-intensive manual chart review. Recently suggested solutions involve machine learning to identify surgical site infections directly from the medical record. Deep learning is a form of machine learning that has historically performed better than traditional methods while being harder to interpret. We propose a deep learning model, a long short-term memory network, for the identification of surgical site infection from the medical record with an attention layer for explainability. MethodsWe retrieved structured data and clinical notes from the University of Utah Health System’s electronic health care record for operative events randomly selected for manual chart review from January 2016 to June 2021. Surgical site infection occurring within 30 days of surgery was determined according to the National Surgical Quality Improvement Program definition. We trained the long short-term memory model along with traditional machine learning models for comparison. We calculated several performance metrics from a holdout test set and performed additional analyses to understand the performance of the long short-term memory, including an explainability analysis. ResultsSurgical site infection was present in 4.7% of the total 9,185 operative events. The area under the receiver operating characteristic curve and sensitivity of the long short-term memory was higher (area under the receiver operating characteristic curve: 0.954, sensitivity: 0.920) compared to the top traditional model (area under the receiver operating characteristic curve: 0.937, sensitivity: 0.736). The top 5 features of the long short-term memory included 2 procedure codes and 3 laboratory values. ConclusionSurgical site infection surveillance is vital for the reduction of surgical site infection rates. Our explainable long short-term memory achieved a comparable area under the receiver operating characteristic curve and greater sensitivity when compared to traditional machine learning methods. With explainable deep learning, automated surgical site infection surveillance could replace burdensome manual chart review processes.

Natural Language Processing for the Identification of Surgical Site Infections in Orthopaedics.

The identification of surgical site infections for infection surveillance in hospitals depends on the manual abstraction of medical records and, for research purposes, depends mainly on the use of administrative or claims data. The objective of this study was to determine whether automating the abstraction process with natural language processing (NLP)-based models that analyze the free-text notes of the medical record can identify surgical site infections with predictive abilities that match the manual abstraction process and that surpass surgical site infection identification from administrative data. We used surgical site infection surveillance data compiled by the infection prevention team to identify surgical site infections among patients undergoing orthopaedic surgical procedures at a tertiary care academic medical center from 2011 to 2017. We compiled a list of keywords suggestive of surgical site infections, and we used NLP to identify occurrences of these keywords and their grammatical variants in the free-text notes of the medical record. The key outcome was a binary indicator of whether a surgical site infection occurred. We estimated 7 incremental multivariable logistic regression models using a combination of administrative and NLP-derived variables. We split the analytic cohort into training (80%) and testing data sets (20%), and we used a tenfold cross-validation approach. The main analytic cohort included 172 surgical site infection cases and 200 controls that were repeatedly and randomly selected from a pool of 1,407 controls. For Model 1 (variables from administrative data only), the sensitivity was 68% and the positive predictive value was 70%; for Model 4 (with NLP 5-grams [distinct sequences of 5 contiguous words] from the medical record), the sensitivity was 97% and the positive predictive value was 97%; and for Model 7 (a combination of Models 1 and 4), the sensitivity was 97% and the positive predictive value was 97%. Thus, NLP-based models identified 97% of surgical site infections identified by manual abstraction with high precision and 43% more surgical site infections compared with models that used administrative data only. Models that used NLP keywords achieved predictive abilities that were comparable with the manual abstraction process and were superior to models that used administrative data only. NLP has the potential to automate and aid accurate surgical site infection identification and, thus, play an important role in their prevention. This study examines NLP's potential to automate the identification of surgical site infections. This automation can potentially aid the prevention and early identification of these surgical complications, thereby reducing their adverse clinical and economic impact.

Identification of surgical site infections using electronic health record data

The objective of this study was to develop an algorithm for identifying surgical site infections (SSIs) using independent variables from electronic health record data and outcomes from the American College of Surgeons National Surgical Quality Improvement Program to supplement manual chart review. We fit 3 models to data from patients undergoing operations at the University of Colorado Hospital between 2013 and 2015: a similar model reported previously in the literature, a comprehensive model with 136 possible predictors, and a combination of those. All models used a generalized linear model with a lasso penalty. Several techniques for handling imbalance in the outcome were also used: Youden's J statistic to optimize the probability cutoff and sampling techniques combined with Youden's J. The models were then tested on data from patients undergoing operations during 2016. Two hundred thirty of 6,840 patients (3.4%) had an SSI. The comprehensive model fit to the full set of training data performed the best, achieving 90% specificity, 80% sensitivity, and an area under the receiver operating characteristic curve of 0.89. We identified a model that accurately identified SSIs. The framework presented can be easily implemented by other American College of Surgeons National Surgical Quality Improvement Program-participating hospitals to develop models for enhancing surveillance of SSIs.

Development of a single, practical measure of surgical site infection (SSI) for patient report or observer completion.

Background:Surgical site infections (SSIs) are the third most common hospital-associated infection and can lead to significant patient morbidity and healthcare costs. Identification of SSIs is key to surveillance and research but reliable assessment is challenging, particularly after hospital discharge when most SSIs present. Existing SSI measurement tools have limitations and their suitability for post-discharge surveillance is uncertain.Aims:This study aimed to develop a single measure to identify SSI after hospital discharge, suitable for patient or observer completion.Methods:A three-phase mixed methods study was undertaken: Phase 1, an analysis of existing tools and semi-structured interviews with patients and professionals to establish the content of the measure; Phase 2, development of questionnaire items suitable for patients and professionals; Phase 3, pre-testing the single measure to assess acceptability and understanding to both stakeholder groups. Interviews and pre-testing took place over 12 months in 2014–2015 with patients and professionals from five specialties recruited from two UK hospital Trusts.Findings:Analyses of existing tools and interviews identified 19 important domains for assessing SSIs. Domains were developed into provisional questionnaire items. Pre-testing and iterative revision resulted in a final version with 16 items that were understood and easily completed by patients and observers (healthcare professionals).Conclusion:A single patient and observer measure for post-discharge SSI assessment has been developed. Further testing of the validity, reliability and accuracy of the measure is underway.

Incidence of healthcare-associated infections in a tertiary care hospital: results from a three-year period of electronic surveillance

Point prevalence surveillance is widely used to monitor healthcare-associated infections (HAIs). Incidence surveillance offers more accurate information than point prevalence surveillance, but is more time consuming. Electronic surveillance systems may allow for more widespread incidence surveillance. To determine the incidence of HAI in a tertiary care hospital in Finland over a three-year period from 2011 to 2013 using an automated electronic IS programme, linked to all of the hospital's electronic databases. The programme identified cases of HAI prospectively from initiation of antibiotic treatment. All of the cases were verified manually after discharge. In order to evaluate the sensitivity of the electronic method for the identification of surgical site infections, the medical records of all patients who underwent cardiac surgery or total joint arthroplasty were evaluated retrospectively. In total, 78,211 patients, covering 321,974 patient-days, were admitted to the wards during the study period, and 29,694 antibiotic treatment initiations were registered. After manual review, 5089 (17.1%) of these were found to be for HAIs. The total time needed to undertake the surveillance for the whole hospital (353 hospital beds) was 255 days, which is the approximate total annual working time for one nurse. Sensitivity evaluation showed that the number of surgical site infections identified by manual and electronic methods was identical. The three-year incidence of HAI was 15.8 per 1000 patient-days and 4.9% of all discharged patients. Continuous electronic incidence surveillance based on initiation of antibiotic treatment may be a practical means of measuring hospital-wide incidence of HAI, but this method still requires personnel resources.

FRI0284 Preoperative Use of Biologic Agents is not an Independent Risk Factor for SSI and Delayed Wound Healing in Patients with Rheumatoid Arthritis

BackgroundThe perioperative use of biologic DMARDs in patients with rheumatoid arthritis (RA) has potential risks for delayed wound healing and infectious complications. Surgeons are recommended to discontinue biologic DMARDs to...

Intérêt de l’utilisation des données du Programme médicalisé des systèmes d’information (PMSI) pour la surveillance des infections nosocomiales aux Hospices Civils de Lyon

Background Surveillance is an effective element in the fight against nosocomial infections, but the monitoring methods are often cumbersome and time consuming. The detection of infection in computerized databases is a means to alleviate the workload of health care teams. The objective of this study was to evaluate the performance of using discharge summaries in medico-administrative databases (PMSI) for the identification of nosocomial infections in surgery, intensive care and obstetrics. Methods The retrospective assessment study included patients who were hospitalized in general surgery, intensive care and obstetrics at different periods of time in 2006 and 2007 depending on the wards. Patients were monitored according to standard protocols which are coordinated at the regional level by the Southeast coordinating centre (CCLIN). The performance of identifying cases of nosocomial infection from discharge diagnoses coded by using the International Classification of Diseases (tenth revision) was evaluated by a study of sensitivity, specificity, positive and negative predictive values with their 95% confidence intervals. Results Using a limited number of diagnostic codes, the sensitivity and specificity were, respectively, 26.3% (95% CI 13.2–42.1) and 99.5% (95% 98.8–100.0) for the identification of surgical site infections. By expanding the number of diagnostic codes, the sensitivity and specificity were 78.9% (95% CI 65.8–92.1) and 65.7% (95% CI 61.0–70.3). The sensitivity and specificity for case identification of nosocomial infections in intensive care were 48.8% (95% CI 42.6–55.0) and 78.4% (95% CI 76.1–80.1), and were 42.9% (95% CI 25.0–60.7) and 87.3% (95% CI 85.2–89.3) for identification of postpartum infections. Conclusion The PMSI is not a sufficiently efficient method in terms of sensitivity to be used in surveillance of nosocomial infections. A reassessment of the PMSI must be considered, with changes in coding of comorbidity that occurred in 2009.

Risk Factors for Surgical Site Infections Following Spinal Fusion Surgery at a Children's Hospital

BACKGROUND/OBJECTIVES: Postoperative infections especially surgical site infections (SSIs) are serious complications of spine fusions requiring extensive medical care and frequently additional surgical intervention. Between 1996 and 1999 at a children's hospital, a significant increase in the SSI rate was identified in patients with no known risk factors for infection preoperatively. The objectives of the study were to identify specific risk factors for SSIs in patients undergoing spinal fusion surgery and to analysis the characteristics of these infections. METHODS: A retrospective cohort study with medical chart review of all patients (ages 6 months to 21 years of age) who underwent spine fusion surgery between Jan. 1999 and Dec. 2003. Risk factors for SSI were determined using univariate analysis and multivariate logistic regression. SSIs were analyzed using CDC and NNSI guidelines. RESULTS: During the 5-year study, 234 spine fusion surgeries were performed with 19 (8%) patients developing a SSI. Univariate analysis identified diagnosis (idiopathic, neuromuscular, or structural scolotic curves), a secondary condition (such as cerebral palsy, myelodysplasia, an autoimmune condition, or genetic syndromes), and ASA score were significant preoperative risks with cut time, blood loss, blood replacement (type and amount), and bone graft used significant procedural risks. Postoperatively, hospital length of stay, antibiotic use, transfer to the ICU, and other non-infectious postoperative problems were significant. In the multivariate logistic regression, only cut time remained significant. Of the 19 SSIs, 11 (58%) occurred ≤ 30 days; 3 (16% occurred between 31 to 365 days, and 5 (26%) were identified more than 366 days postoperatively. Sixty three percent were classified as organ/space infections, requiring extensive medical intervention. Comparison of National Nosocomial Infection Surveillance risk index and study population infection rates Number of Risk Factor NNIS Rate Study Population Rate 0 1.2% 1.3% 1 3.1% 5.1% 2,3 7.2% 1.7% CONCLUSIONS: Diagnosis, cut time and ASA were found to be significant in this study. The identification of SSIs after 366 days (2 to 5 years) postoperatively is becoming an increasing concern requiring surveillance for infection to continue for several years to determine the actual SSI rate for this population of patients.

Enhanced Identification of Postoperative Infections among Outpatients

We investigated using administrative claims data to identify surgical site infections (SSI) after breast surgery and cesarean section. Postoperative diagnosis codes, procedure codes, and pharmacy information were automatically scanned and used to identify claims suggestive of SSI ("indicators") among 426 (22%) of 1,943 breast procedures and 474 (10%) of 4,859 cesarean sections. For 104 breast procedures with indicators explained in available medical records, SSI were confirmed for 37%, and some infection criteria were present for another 27%. Among 204 cesarean sections, SSI were confirmed for 40%, and some criteria were met for 27%. The extrapolated infection rates of 2.8% for breast procedures and 3.1% for cesarean section were similar to those reported by the National Nosocomial Infection Surveillance program but differ in representing predominantly outpatient infections. Claims data may complement other data sources for identification of surgical site infections following breast surgery and cesarean section.