Machine Learning Techniques to Identify Antimicrobial Resistance in the Intensive Care Unit.

The presence of bacteria with resistance to specific antibiotics is one of the greatest threats to the global health system. According to the World Health Organization, antimicrobial resistance has already reached alarming levels in many parts of the world, involving a social and economic burden for the patient, for the system, and for society in general. Because of the critical health status of patients in the intensive care unit (ICU), time is critical to identify bacteria and their resistance to antibiotics. Since common antibiotics resistance tests require between 24 and 48 h after the culture is collected, we propose to apply machine learning (ML) techniques to determine whether a bacterium will be resistant to different families of antimicrobials. For this purpose, clinical and demographic features from the patient, as well as data from cultures and antibiograms are considered. From a population point of view, we also show graphically the relationship between different bacteria and families of antimicrobials by performing correspondence analysis. Results of the ML techniques evidence non-linear relationships helping to identify antimicrobial resistance at the ICU, with performance dependent on the family of antimicrobials. A change in the trend of antimicrobial resistance is also evidenced.

Prevention of Ventilator-Associated Pneumonia: The Multimodal Approach of the Spanish ICU "Pneumonia Zero" Program.

OBJECTIVES: The "Pneumonia Zero" project is a nationwide multimodal intervention based on the simultaneous implementation of a comprehensive evidence-based bundle measures to prevent ventilator-associated pneumonia in critically ill patients admitted to the ICU. DESIGN: Prospective, interventional, and multicenter study. SETTING: A total of 181 ICUs throughout Spain. PATIENTS: All patients admitted for more than 24 hours to the participating ICUs between April 1, 2011, and December 31, 2012. INTERVENTION: Ten ventilator-associated pneumonia prevention measures were implemented (seven were mandatory and three highly recommended). The database of the National ICU-Acquired Infections Surveillance Study (Estudio Nacional de Vigilancia de Infecciones Nosocomiales [ENVIN]) was used for data collection. Ventilator-associated pneumonia rate was expressed as incidence density per 1,000 ventilator days. Ventilator-associated pneumonia rates from the incorporation of the ICUs to the project, every 3 months, were compared with data of the ENVIN registry (April-June 2010) as the baseline period. Ventilator-associated pneumonia rates were adjusted by characteristics of the hospital, including size, type (public or private), and teaching (postgraduate) or university-affiliated (undergraduate) status. MEASUREMENTS AND MAIN RESULTS: The 181 participating ICUs accounted for 75% of all ICUs in Spain. In a total of 171,237 ICU admissions, an artificial airway was present on 505,802 days (50.0% of days of stay in the ICU). A total of 3,474 ventilator-associated pneumonia episodes were diagnosed in 3,186 patients. The adjusted ventilator-associated pneumonia incidence density rate decreased from 9.83 (95% CI, 8.42-11.48) per 1,000 ventilator days in the baseline period to 4.34 (95% CI, 3.22-5.84) after 19-21 months of participation. CONCLUSIONS: Implementation of the bundle measures included in the "Pneumonia Zero" project resulted in a significant reduction of more than 50% of the incidence of ventilator-associated pneumonia in Spanish ICUs. This reduction was sustained 21 months after implementation.

Validity of the current risk assessment scale for pressure ulcers in intensive care (EVARUCI) and the Norton-MI scale in critically ill patients.

OBJECTIVE: The objective of this study was to evaluate the validity of risk detection scales EVARUCI and Norton-MI (modified by INSALUD) to detect critical adult patients with the risk of developing pressure ulcers (PU) in an intensive care unit (ICU). DESIGN: The authors have conducted a descriptive, prospective study at the ICU in their hospital from 2008 to 2014. The evaluations of both scales were registered daily by nurses from the unit. PATIENTS: Adult patients admitted into the ICU. MAIN OUTCOMES MEASURE: The study measured the sensitivity, specificity, positive predictive values (PPV) and negative predictive values (NPV) of each of the scores for both scales and areas under curve (AUC) of receiver operating characteristics (ROC). MAIN RESULTS: The authors have evaluated a total of 2534 patients. For the cut-off point recommended by the authors in the scale Norton-MI (PC 14), a sensitivity of 94,05% (93,28-94,82) was obtained, specificity of 40,47% (39,72-41,22), VPP 26,22% and VPN 96,80%. For EVARUCI (CP 10) a sensitivity of 80,43% (79,15-81,72), specificity 64,41 (63,68-65,14), VPP 33,71% and VPN of 93,60%. The ABC-COR was 0,774 with a 95% CI of 0,766 to 0,781 for the scale of Norton-MI and 0.756 with a 95% CI of 0,749 to 0,764 for EVARUCI. CONCLUSION: Both scales are valid to help predict the risk of developing PU in critical patients. The sensitivity and ABC-COR are very similar for EVARUCI and Norton-Mi. The authors state they do not have any financial interests linked to this article.

Dimensionality reduction and ensemble of LSTMs for antimicrobial resistance prediction.

Bacterial resistance to antibiotics has been rapidly increasing, resulting in low antibiotic effectiveness even treating common infections. The presence of resistant pathogens in environments such as a hospital Intensive Care Unit (ICU) exacerbates the critical admission-acquired infections. This work focuses on the prediction of antibiotic resistance in Pseudomonas aeruginosa nosocomial infections at the ICU, using Long Short-Term Memory (LSTM) artificial neural networks as the predictive method. The analyzed data were extracted from the Electronic Health Records (EHR) of patients admitted to the University Hospital of Fuenlabrada from 2004 to 2019 and were modeled as Multivariate Time Series. A data-driven dimensionality reduction method is built by adapting three feature importance techniques from the literature to the considered data and proposing an algorithm for selecting the most appropriate number of features. This is done using LSTM sequential capabilities so that the temporal aspect of features is taken into account. Furthermore, an ensemble of LSTMs is used to reduce the variance in performance. Our results indicate that the patient's admission information, the antibiotics administered during the ICU stay, and the previous antimicrobial resistance are the most important risk factors. Compared to other conventional dimensionality reduction schemes, our approach is able to improve performance while reducing the number of features for most of the experiments. In essence, the proposed framework achieve, in a computationally cost-efficient manner, promising results for supporting decisions in this clinical task, characterized by high dimensionality, data scarcity, and concept drift.

Extracorporeal membrane oxygenation (ECMO) and beyond in near fatal asthma: A comprehensive review.

The treatment of choice in severe asthma exacerbations with respiratory failure includes ventilatory support, both invasive and/or non-invasive, along with different kinds of asthma medication. Of note, the rate of mortality of patients with asthma has decreased substantially in recent years mainly due to significant advances in pharmacological treatment and other management strategies. However, the risk of death in patients with severe asthma who require invasive mechanical ventilation has been estimated between 6.5% and 10.3%. When conventional measures fail, rescue strategies, such as extracorporeal membrane oxygenation (ECMO) or extracorporeal CO2 removal (ECCO2R) may need to be implemented. While ECMO does not constitute a definitive treatment per se, it can minimize further ventilator associated lung injury (VALI) and can enable diagnostic-therapeutic maneuvers that cannot be performed without ECMO such as bronchoscopy and transfer for diagnostic imaging. Asthma is one of the diseases that is associated with excellent outcomes for patients with refractory respiratory failure requiring ECMO support, as shown by the Extracorporeal Life Support Organization (ELSO) registry. Moreover, in such situations, the use of ECCO2R for rescue has been described and utilized in both children and adults and is more widely spread in different hospitals than ECMO. In this article, we aim to review the evidence for the usefulness of extracorporeal respiratory support measures in the management of severe asthma exacerbations that lead to respiratory failure.

Artificial Intelligence to Get Insights of Multi-Drug Resistance Risk Factors during the First 48 Hours from ICU Admission.

Multi-drug resistance (MDR) is one of the most current and greatest threats to the global health system nowadays. This situation is especially relevant in Intensive Care Units (ICUs), where the critical health status of these patients makes them more vulnerable. Since MDR confirmation by the microbiology laboratory usually takes 48 h, we propose several artificial intelligence approaches to get insights of MDR risk factors during the first 48 h from the ICU admission. We considered clinical and demographic features, mechanical ventilation and the antibiotics taken by the patients during this time interval. Three feature selection strategies were applied to identify statistically significant differences between MDR and non-MDR patient episodes, ending up in 24 selected features. Among them, SAPS III and Apache II scores, the age and the department of origin were identified. Considering these features, we analyzed the potential of machine learning methods for predicting whether a patient will develop a MDR germ during the first 48 h from the ICU admission. Though the results presented here are just a first incursion into this problem, artificial intelligence approaches have a great impact in this scenario, especially when enriching the set of features from the electronic health records.

Data and Network Analytics for COVID-19 ICU Patients: A Case Study for a Spanish Hospital.

The COVID-19 pandemic presents unprecedented challenges to the healthcare systems around the world. In 2020, Spain was among the countries with the highest Intensive Care Unit (ICU) hospitalization and mortality rates. This work analyzes data of COVID-19 patients admitted to a Spanish ICU during the first wave of the pandemic. The patients in our study either died (deceased patients) or were discharged from the ICU (non-deceased patients) and underwent the following landmarks: beginning of symptoms; arrival at the emergency department; beginning of the hospital stay; and ICU admission. Our goal is to create a graph-based data-science methodology to find associations among patients' comorbidities, previous medication, symptoms, and the COVID-19 treatment, and to analyze their evolution across landmarks. Towards that end, we first perform a hypothesis test based on bootstrap to identify discriminative features among deceased and non-deceased patients. Then, we leverage graph-based representations and network analytics to determine pairwise associations and complex relations among clinical features. The descriptive statistical analysis confirms that deceased patients exhibit multiple comorbidities with stronger levels of association and are treated with a wider range of drugs during the ICU stay. We also observe that the most common treatment was the simultaneous administration of lopinavir/ritonavir with hydroxychloroquine, regardless of the patients' outcome. Our results illustrate how graph tools and representations yield insights on the relations among comorbidities, drug treatments, and patients' evolution. All in all, the approach puts forth a new data-analysis tool for clinicians that can be applied to analyze (post-COVID) symptom/patient evolution.

[Usefulness of failure mode and effects analysis to improve patient safety during the process of incorporating new nurses in an intensive care unit]. / Utilidad del análisis modal de fallos y efectos para la mejora de la seguridad de los pacientes, en el proceso de incorporación de nuevo personal de enfermería a un servicio de medicina intensiva.

OBJECTIVE: To analyze proactively the process of incorporating new nurses in the intensive care unit (ICU) in order to detect risk areas and establish improvements that increase critical patient safety. MATERIAL AND METHODS: Once the risk area was defined, the different phases of failure mode and effects analysis (FMEA) were applied: work team selection; process design; process phases definition; failure modes, possible causes and effects analysis; risk priority for each failure, and development of ameliorating and corrective actions. The proposed actions consisted of an orientation and training program (theoretical and practical) for new nurses, a supervision plan, a progressive responsibility program and ICU participation in personnel recruitment. RESULTS: Twelve nurses began to work in the ICU during the first 18 months of the program's implementation. Of these, only one nurse had full experience in critical care and three had partial experience. Participation of the ICU in personnel recruitment was nil. All the nurses with no or partial experience followed the orientation program (nursing supervisor interview, test of previous knowledge, handing over of the employee handbook, etc.), the theoretical and practical training program (supervision and tutorship) and the progressive responsibility program. More than half (63.6%) of the new nurses had another nurse duplicating their jobs during the training period and 54.5% of the new nurses attended the critical care course for nurses. Nurses participating in the orientation and training program expressed a high level of satisfaction. These measures helped nurses to decrease their stress and anxiety, increase and consolidate their knowledge, and provide safer care to critical patients. CONCLUSIONS: FMEA is a useful tool for improving ICU processes, even those involving human resources. The improvements implemented to decrease clinical risk related to the incorporation of new nurses in the ICU, based on previous training, will increase the safety of critical patient care by decreasing human errors due to inexperience.

[Perception of safety culture in Spanish intensive care units]. / Percepción de la cultura de seguridad en los servicios de medicina intensiva españoles.

BACKGROUND AND OBJECTIVE: Promoting a safety culture in intensive care units (ICUs) is a basic strategy to improve patient safety. The aim of this study was to measure the safety culture in Spanish ICUs. METHOD: We drafted a questionnaire based on the Safety Climate Survey (SCS) and the Safety Attitude Questionnaire-ICU model (SAQ-ICU). A translation-back translation method was employed together with focus group discussions. A questionnaire was designed that analyzed six dimensions: teamwork climate, safety climate, perceptions of management, job satisfaction, working conditions, and stress recognition. The survey was delivered to 22 Spanish ICUs. The results were analyzed to detect strengths and weaknesses in the ICU safety culture. RESULTS: The internal consistency of the survey was 0.92. The response rate was 29.8%. The distribution of positive responses by dimension was as follows: job satisfaction: 65.2%, teamwork climate: 62.1%, safety climate: 50.7%, perceptions of management: 30.3%, working conditions: 43.3%, and stress recognition: 68.8%. Some strengths were detected, such as the percentages who responded affirmatively to the statements "I like my job" (95.1%) and "briefings are important for patient safety" (86.8%). We found significant differences by hospital size: attitudes were better in ICU staff in smaller hospitals than in large hospitals. CONCLUSIONS: The safety culture is poor in Spanish ICUs, but awareness is greater in smaller hospitals. Nevertheless, some strengths were identified, such as communication. Tools to promote free reporting of errors and incident reports should be provided.

Percepción de la cultura de seguridad en los servicios de medicina intensiva españoles / Perception of Safety Culture in Spanish Intensive Care Units

Introducción y objetivo: Potenciar la cultura de seguridad en los servicios de medicina intensiva (SMI) es una estrategia fundamental para mejorar la seguridad del paciente. El objetivo de este estudio es medir la percepción que sobre la cultura de seguridad (CS) se tiene en los SMI españoles. Método: Se elaboró una encuesta basada en las encuestas estadounidenses Safety Climate Survey (SCS) y Safety Attitude Questionnaire ICU version (SAQ-ICU). Se utilizó el método “traducción-retrotraducción” junto con la discusión por grupos focales. Se redactó un cuestionario que analizaba seis dimensiones: ambiente laboral, relaciones laborales, organización del servicio y el hospital, condiciones laborales, clima de seguridad y reconocimiento del nivel de estrés. El cuestionario se repartió en 22 SMI. Los resultados se analizaron para detectar las fortalezas y las debilidades de la cultura de seguridad de los SMI. Resultados: La coherencia interna del cuestionario fue 0,92. La tasa de participación fue del 29,8%. La distribución de las respuestas positivas por dimensiones fue: ambiente, el 65,2%; relaciones laborales, el 62,1%; organización, el 30,3%; condiciones laborales, el 43,3%; clima de seguridad, el 50,7%, y reconocimiento del nivel de estrés, el 68,8%. Se detectaron fortalezas como “me gusta mi trabajo” (95,1%) o “las sesiones son importantes” (86,8%). Se encontraron diferencias estadísticamente significativas a favor de los hospitales pequeños. Conclusiones: La CS en los SMI españoles es insuficiente, aunque se percibe mejor en los hospitales peque- ños. Sin embargo, encontramos fortalezas en el ámbito de la comunicación. Deberíamos proporcionar herramientas para potenciar una comunicación libre de errores y la declaración de los eventos adversos (AU)

Background and objective: Promoting a safety culture in intensive care units (ICUs) is a basic strategy to improve patient safety. The aim of this study was to measure the safety culture in Spanish ICUs. Method: We drafted a questionnaire based on the Safety Climate Survey (SCS) and the Safety Attitude Questionnaire-ICU model (SAQ-ICU). A translation-back translation method was employed together with focus group discussions. A questionnaire was designed that analyzed six dimensions: teamwork climate, safety climate, perceptions of management, job satisfaction, working conditions, and stress recognition. The survey was delivered to 22 Spanish ICUs. The results were analyzed to detect strengths and weaknesses in the ICU safety culture. Results: The internal consistency of the survey was 0.92. The response rate was 29.8%. The distribution of positive responses by dimension was as follows: job satisfaction: 65.2%, teamwork climate: 62.1%, safety climate: 50.7%, perceptions of management: 30.3%, working conditions: 43.3%, and stress recognition: 68.8%. Some strengths were detected, such as the percentages who responded affirmatively to the statements “I like my job” (95.1%) and “briefings are important for patient safety” (86.8%). We found significant differences by hospital size: attitudes were better in ICU staff in smaller hospitals than in large hospitals (AU)

Utilidad del análisis modal de fallos y efectos para la mejora de la seguridad de los pacientes, en el proceso de incorporación de nuevo personal de enfermería a un servicio de medicina intensiva / Usefulness of failure mode and effects analysis to improve patient safety during the process of incorporating new nurses in an intensive care unit

Objetivo: Analizar proactivamente el proceso de incorporación laboral del personal de enfermería a un servicio de medicina intensiva (SMI) para detectar áreas de riesgo y establecer acciones de mejora que aumenten la seguridad de los pacientes críticos. Material y métodos: Tras detectar el área de riesgo, se aplicó la metodología AMFE en sus diferentes fases: selección del grupo de trabajo, diseño del proceso, identificación de las fases, análisis de fallos, posibles causas y efectos, priorización del riesgo de cada fallo y desarrollo de las acciones de prevención y mejora. Las acciones propuestas fueron: plan de acogida y formación (teórico-práctica), tutela y supervisión, plan de responsabilidad progresiva (PRP) y participación del SMI en la selección del personal. Resultados: Se incorporaron 12 enfermeras en los primeros 18 meses de funcionamiento del plan; de ellas, sólo 1 tenía experiencia plena en UCI y 3, experiencia parcial. La participación del SMI en la selección de personal fue nula. El 100% de las enfermeras sin experiencia o con experiencia parcial siguieron el plan de acogida (entrevista con la supervisora, evaluación de conocimientos previos, entrega del Manual de Acogida), el plan de formación teórico-práctico (tutela y supervisión) y el PRP. El 63,6% de las incorporaciones tuvo “doblaje” de su puesto durante su incorporación. El 54,5% asistió al Curso de Cuidados Críticos para enfermería. Las enfermeras que recibieron el plan de acogida manifiestan un alto grado de satisfacción, y estas medidas contribuyeron a disminuir su estrés y su ansiedad, ampliar y afianzar conocimientos y practicar una asistencia más segura a los pacientes críticos. Conclusiones: El AMFE es una herramienta útil en la mejora de los procesos de los SMI que se puede aplicar también a aquellos en los que están implicados los recursos humanos. Las mejoras implantadas para disminuir el riesgo asistencial en relación con la incorporación laboral de nuevas enfermeras, basadas en la formación previa, contribuirán a dar una asistencia más segura a los enfermos, al disminuir los errores humanos derivados de la inexperiencia (AU)

Objective: To analyze proactively the process of incorporating new nurses in the intensive care unit (ICU) in order to detect risk areas and establish improvements that increase critical patient safety. Material and methods: Once the risk area was defined, the different phases of failure mode and effects analysis (FMEA) were applied: work team selection; process design; process phases definition; failure modes, possible causes and effects analysis; risk priority for each failure, and development of ameliorating and corrective actions. The proposed actions consisted of an orientation and training program (theoretical and practical ) for new nurses, a supervision plan, a progressive responsibility program and ICU participation in personnel recruitment. Results: Twelve nurses began to work in the ICU during the first 18 months of the program’s implementation. Of these, only one nurse had full experience in critical care and three had partial experience. Participation of the ICU in personnel recruitment was nil. All the nurses with no or partial experience followed the orientation program (nursing supervisor interview, test of previous knowledge, handing over of the employee handbook, etc.), the theoretical and practical training program (supervision and tutorship) and the progressive responsibility program. More than half (63.6%) of the new nurses had another nurse duplicating their jobs during the training period and 54.5% of the new nurses attended the critical care course for nurses. Nurses participating in the orientation and training program expressed a high level of satisfaction. These measures helped nurses to decrease their stress and anxiety, increase and consolidate their knowledge, and provide safer care to critical patients. Conclusions: FMEA is a useful tool for improving ICU processes, even those involving human resources. The improvements implemented to decrease clinical risk related to the incorporation of new nurses in the ICU, based on previous training, will increase the safety of critical patient care by decreasing human errors due to inexperience (AU)