Pediatric cardiac surgery: machine learning models for postoperative complication prediction.

PURPOSE: Managing children undergoing cardiac surgery with cardiopulmonary bypass (CPB) presents a significant challenge for anesthesiologists. Machine Learning (ML)-assisted tools have the potential to enhance the recognition of patients at risk of complications and predict potential issues, ultimately improving outcomes. METHODS: We evaluated the prediction capacity of six models, ranging from logistic regression to support vector machine, using a dataset comprising 33 variables and 1364 subjects. The Area Under the Curve (AUC) and the F1 score served as the primary evaluation metrics. Our primary objectives were twofold: first, to develop an effective prediction model, and second, to create a user-friendly comprehensive model for identifying high-risk patients. RESULTS: The logistic regression model demonstrated the highest effectiveness, achieving an AUC of 83.65%, and an F1 score of 0.7296, with balanced sensitivity and specificity of 77.94% and 76.47%, respectively. In comparison, the comprehensive three-layer decision tree model achieved an AUC of 72.84%, with sensitivity (79.41%) comparable to more complex models. CONCLUSION: Our machine learning-assisted tools provide an additional perspective and enhance the predictive capabilities of traditional scoring methods. These tools can assist anesthesiologists in making well-informed decisions. Furthermore, we have successfully demonstrated the feasibility of creating a practical white-box model. The next steps involve conducting clinical validation and multicenter cross-validation. TRIAL REGISTRATION: NCT05537168.

Thermogravimetric, Morphological and Infrared Analysis of Blends Involving Thermoplastic Starch and Poly(ethylene-co-methacrylic acid) and Its Ionomer Form.

This study focuses on the thermal properties and structural features of blends consisting of thermoplastic starch (TPS) and poly(ethylene-co-methacrylic acid) copolymer (EMAA) or its ionomer form (EMAA-54Na). The aim is to investigate how carboxylate functional groups of the ionomer form intervene in blends compatibility at the interface of the two materials and how this impacts their properties. Two series of blends (TPS/EMAA and TPS/EMAA-54Na) were produced with an internal mixer, with TPS compositions between 5 and 90 wt%. Thermogravimetry shows two main weight losses, indicating that TPS and the two copolymers are primarily immiscible. However, a small weight loss existing at intermediate degradation temperature between those of the two pristine components reveals specific interactions at the interface. At a mesoscale level, scanning electron microscopy confirmed thermogravimetry results and showed a two-phase domain morphology, with a phase inversion at around 80 wt% TPS, but also revealed a different surface appearance evolution between the two series. Fourier-transformed infrared spectroscopy analysis also revealed discrepancies in fingerprint between the two series of blends, analysed in terms of additional interactions in TPS/EMAA-54Na coming from the supplementary sodium neutralized carboxylate functions of the ionomer.

Impact of collaborative physician-pharmacist stewardship strategies on prophylactic antibiotic practices: a quasi-experimental study.

BACKGROUND: An effective use of surgical antibiotic prophylaxis (SAP) appears essential to prevent the development of infections linked to surgery while inappropriate and excessive prescriptions of prophylactic antibiotics increase the risk of adverse effects, bacterial resistance and Clostridium difficile infections. In this study, we aimed to analyze SAP practices in an acute secondary hospital in Belgium during the years 2016-2021 in order to evaluate the impacts of combined stewardship interventions, implemented thanks to a physician-pharmacist collaboration. METHODS: A quasi-experimental study on SAP practices was conducted during 5 years (2016-2021) in a Belgian University Hospital. We first performed a retrospective observational transversal study on a baseline group (2016.1-2016.4). Then, we constituted a group of patients (2017.1-2017.4) to test a combined intervention strategy of stewardship which integrated the central role of a pharmacist in antibiotic stewardship team and in the pre-operative delivery of nominative kits of antibiotics adapted to patient factors. After this test, we collected patient data (2018.1-2018.4) to evaluate the sustained effects of stewardship interventions. Furthermore, we evaluated SAP practices (2019.1-2019.4) after the diffusion of a computerized decision support system. Finally, we analyzed SAP practices in the context of the COVID-19 pandemic (2020.1-2020.4 and 2021.1-2021.4). The groups were compared from year to year in terms of compliance to institutional guidelines, as evaluated from seven criteria (χ2 test). RESULTS: In total, 760 surgical interventions were recorded. The observational study within the baseline group showed that true penicillin allergy, certain types of surgery and certain practitioners were associated with non-compliance (p < 0.05). Compared with the baseline group, the compliance was significantly increased in the test group for all seven criteria assessed (p < 0.05). However, the effects were not fully sustained after discontinuation of the active interventions. Following the diffusion of the computerized decision support system, the compliance to guidelines was not significantly improved. Finally, the COVID-19 pandemic did not appear to affect the practices in terms of compliance to guidelines. CONCLUSIONS: This study shows that optimization of SAP practices is achievable within a proactive multidisciplinary approach including real-time pharmaceutical interventions in the operating area and in the care units practicing SAP.

Common clinical thresholds of intraoperative hypotension and 30-day mortality following surgery: A retrospective cohort study.

BACKGROUND: A wide range of thresholds define intraoperative hypotension and can be used to guide intraoperative blood pressure management. Many clinicians use the systolic blood pressure (SBP) <80 mmHg, the mean arterial pressure (MAP) <60 mmHg and the SBP percent drop from baseline (ΔSBP) >20% as alarming limits that should not be exceeded. Whether these common thresholds are valid limits that can inform clinicians on a possible increased risk of post-operative complications, particularly 30-day mortality, is currently unclear. METHODS: We performed a retrospective registry-based cohort study between January 2015 and July 2016 using departmental hospital databases and the National Death Registry. Uni- and multivariate analyses were performed to assess the association between each of these three thresholds and 30-day post-operative mortality. Six specific markers of hypotension were used. RESULTS: Of 11 304 patients, 86 (0.76%) died within 30 days following surgery. All intraoperative hypotension markers for SBP < 80 mmHg and MAP < 60 mmHg were significantly associated with 30-day mortality (P < .005). Markers of ΔSBP > 20% were not significant. After adjustment for age, gender, American Society of Anesthesiologists (ASA) score, emergency status and risk related to the type of surgery, both SBP < 80 mmHg and MAP < 60 mmHg (the per cent area under the threshold marker) showed the strongest associations with 30-day mortality, with odds ratios (ORs) of 3.02 (95% confidence interval (CI) 1.81-5.07) and 3.77 (95% CI 2.25-6.31) respectively. CONCLUSIONS: Commonly accepted thresholds of intraoperative hypotension, such as an SBP of 80 mmHg and an MAP of 60 mmHg, are valid alarming limits that are significantly and independently associated with 30-day mortality.

Anaesthesia care team improves outcomes in surgical patients compared with solo anaesthesiologist: An observational study.

BACKGROUND: In anaesthesiology, little attention has been drawn to the role of anaesthesia nurses as support personnel on quality of care. OBJECTIVES: To compare an anaesthesiologist alone (solo anaesthesiologist) with an anaesthesia care team (anaesthesiologist and anaesthesia nurse). DESIGN: An observational study. SETTING: A single centre study. PARTICIPANTS: Anaesthesiologists and anaesthesia nurses. INTERVENTION: Anaesthesia performed by solo anaesthesiologists compared with anaesthesia care teams. MAIN OUTCOME MEASURES: 30-day postoperative mortality and hospital length of stay. Propensity score matching was performed by logistic regression to adjust for baseline differences between the two groups and pairs of perfectly matched patients were formed. RESULTS: Anaesthesia was performed by solo anaesthesiologists in 2832 patients and by an anaesthesia care team in 2842 patients. Matching with 2095 pairs of perfectly matched patients was formed. The two groups were comparable in respect of sex and duration of anaesthesia but differed notably for age, American Society of Anesthesiologists' physical status score and type of surgery. Logistic regression showed a significantly lower 30-day mortality rate for the anaesthesia care teams compared with solo anaesthesiologists (0.76 vs. 1.56%, P = 0.0014). Length of hospital stay was also significantly reduced when an anaesthesia nurse was present (4.9 ±â€Š10.1 vs. 5.6 ±â€Š11.5 days, P = 0.0011). CONCLUSION: Anaesthesia given by teams of anaesthesiologists and anaesthesia nurses is associated with decreased 30-day postoperative mortality and shorter length of stay when compared with solo anaesthesiologists. Even without any demonstration of causality, this emphasises the benefits of the anaesthesia care team model. TRIAL REGISTRATION: CCB 325201730849.

In- and out-hospital mortality rate in surgical patients.

BACKGROUND: Death rates after surgery are increasingly analysed for clinical audit and quality assessment. Many studies commonly provide information only on deaths that occur during hospital stay, known as in-hospital death rates. By using hospital data set linked to death certificate registry, we recorded in- and out-hospital deaths within 30 and 60 post-operative days. METHODS: The study included all consecutive surgical procedures (denominator) under general or locoregional anaesthesia in adult patients admitted for elective or non-elective inpatient surgery. Patients undergoing planned day-case surgery or obstetrical procedures were excluded. The primary outcome was 30- and 60-day post-operative mortality rate (numerator) whether before or after discharge. RESULTS: The study material consisted of a sample of 36,494 surgical procedures corresponding to 28,202 patients. At 30-day, 384 (crude mortality rate of 1.1%) patients died, 314 (82%) during their hospitalisation and 70 (18%) after discharge. Factors that were associated with in-hospital mortality are ASA scores, emergency, duration of surgery and rate of admission to critical care unit. Within the 30-60 days interval, we recorded 231 supplemental deaths, 103 (45%) after discharge. CONCLUSION: In-hospital mortality alone is an incomplete measure of mortality even within 30 days of care. To identify the missing deaths, hospital records need to be linked to data from death certificate. This connection with the national death registry will allow obtaining the rate of in-hospital and out-hospital death.

Hypocapnia measured by end-tidal carbon dioxide tension during anesthesia is associated with increased 30-day mortality rate.

STUDY OBJECTIVE: To evaluate the relationship between intraoperative end-tidal carbon dioxide (etco2) values and clinical outcomes with special attention on 30-day postoperative mortality and secondarily on hospital length of stay (LOS). DESIGN: Retrospective, observational study. SETTING: Surgical theaters of the University Hospital Center of Charleroi. PATIENTS: Five thousand three hundred seventeen patients ASA I-IV undergoing various surgical procedures (except pediatric and cardiac surgery) under general anesthesia. INTERVENTIONS: No intervention on the patients. MEASUREMENTS: The mean etco2 level measured during anesthesia was secondarily extracted from an electronic information management system. Patients were divided into 2 separate groups based on etco2 values less than or greater than or equal to 35 mm Hg. The primary end point was the in- and outhospital mortality in the 30-day period after surgery. The second was the LOS more than 6 days. MAIN RESULTS: Hypocapnia occurred in 66% of the patients. Mortality rate at 30-day was 84 of 3554 (2.4%) in the low etco2 group vs 15 of 1763 (0.9%) in the other (odds ratio, 2.99 [1.69-5.28]; P<.001). In multivariate analysis, age and ASA scores had significant independent associations with mortality rate. Adjusting for these factors had an effect on the relative odds ratio of etco2 on mortality of 1.99 ([1.11-3.56]; P<.001). Patients with low etco2 experienced higher LOS (14.1±9.4 vs 13.1±8.9 days; P<.001). Thirty five percent of the patients in the low etco2 group were still hospitalized more than 6 days compared with 30% in the other (P<.001). CONCLUSION: Low etco2 level during anesthesia is associated with an increase in postoperative mortality rate and LOS. These results emphasize the importance of preventing hypocapnia during anesthesia to improve surgical outcomes.