RESUMEN
BACKGROUND: Despite low mortality for elective procedures in the United States and developed countries, some patients have unexpected care escalations (UCE) following post-anesthesia care unit (PACU) discharge. Studies indicate patient risk factors for UCE, but determining which factors are most important is unclear. Machine learning (ML) can predict clinical events. We hypothesized that ML could predict patient UCE after PACU discharge in surgical patients and identify specific risk factors. METHODS: We conducted a single center, retrospective analysis of all patients undergoing non-cardiac surgery (elective and emergent). We collected data from pre-operative visits, intra-operative records, PACU admissions, and the rate of UCE. We trained a ML model with this data and tested the model on an independent data set to determine its efficacy. Finally, we evaluated the individual patient and clinical factors most likely to predict UCE risk. RESULTS: Our study revealed that ML could predict UCE risk which was approximately 5% in both the training and testing groups. We were able to identify patient risk factors such as patient vital signs, emergent procedure, ASA Status, and non-surgical anesthesia time as significant variable. We plotted Shapley values for significant variables for each patient to help determine which of these variables had the greatest effect on UCE risk. Of note, the UCE risk factors identified frequently by ML were in alignment with anesthesiologist clinical practice and the current literature. CONCLUSIONS: We used ML to analyze data from a single-center, retrospective cohort of non-cardiac surgical patients, some of whom had an UCE. ML assigned risk prediction for patients to have UCE and determined perioperative factors associated with increased risk. We advocate to use ML to augment anesthesiologist clinical decision-making, help decide proper disposition from the PACU, and ensure the safest possible care of our patients.
Asunto(s)
Aprendizaje Automático , Humanos , Estudios Retrospectivos , Femenino , Masculino , Persona de Mediana Edad , Factores de Riesgo , Anciano , Periodo Perioperatorio , Adulto , Signos Vitales , Complicaciones Posoperatorias/epidemiología , Medición de Riesgo/métodos , Periodo de Recuperación de la AnestesiaRESUMEN
The use of artificial intelligence (AI) and machine learning (ML) in anesthesiology and perioperative medicine is quickly becoming a mainstay of clinical practice. Anesthesiology is a data-rich medical specialty that integrates multitudes of patient-specific information. Perioperative medicine is ripe for applications of AI and ML to facilitate data synthesis for precision medicine and predictive assessments. Examples of emergent AI models include those that assist in assessing depth and modulating control of anesthetic delivery, event and risk prediction, ultrasound guidance, pain management, and operating room logistics. AI and ML support analyzing integrated perioperative data at scale and can assess patterns to deliver optimal patient-specific care. By exploring the benefits and limitations of this technology, we provide a basis of considerations for evaluating the adoption of AI models into various anesthesiology workflows. This analysis of AI and ML in anesthesiology and perioperative medicine explores the current landscape to understand better the strengths, weaknesses, opportunities, and threats (SWOT) these tools offer.
RESUMEN
There is no current consensus on the follow up of kidney function in patients undergoing cardiopulmonary bypass (CPB). The main objectives of this pilot study is to collect preliminary data on kidney function decline encountered on the first postoperative visit of patients who have had CPB and to identify predictors of kidney function decline post hospital discharge. Design: Retrospective chart review. Adult patients undergoing open heart procedures utilizing CPB. Patient demographics, type of procedure, pre-, intra-, and postoperative clinical, hemodynamic echocardiographic, and laboratory data were abstracted from electronic medical records. Acute kidney disease (AKD), and chronic kidney disease (CKD) were diagnosed based on standardized criteria. Interval change in medications, hospital admissions, and exposure to contrast, from hospital discharge till first postoperative visit were collected. AKD, and CKD as defined by standardized criteria on first postoperative visit. 83 patients were available for analysis. AKD occurred in 27 (54%) of 50 patients and CKD developed in 12 (42%) out of 28 patients. Older age was associated with the development of both AKD and CKD. Reduction in right ventricular cardiac output at baseline was associated with AKD (OR: 0.5, 95% CI: 0.3, 0.79, P = 0.01). Prolongation of transmitral early diastolic filling wave deceleration time was associated with CKD (OR: 1.02, 95% CI: 1.01, 1.05, P = 0.03). In-hospital acute kidney injury (AKI) was a predictor of neither AKD nor CKD. AKD and CKD occur after CPB and may not be predicted by in-hospital AKI. Older age, right ventricular dysfunction and diastolic dysfunction are important disease predictors. An adequately powered longitudinal study is underway to study more sensitive predictors of delayed forms of kidney decline after CPB.
Asunto(s)
Lesión Renal Aguda , Insuficiencia Renal Crónica , Adulto , Humanos , Proyectos Piloto , Estudios Retrospectivos , Estudios Longitudinales , Puente Cardiopulmonar/efectos adversos , Riñón , Lesión Renal Aguda/diagnóstico , Lesión Renal Aguda/etiología , Insuficiencia Renal Crónica/etiología , Factores de Riesgo , Enfermedad AgudaRESUMEN
Magnetoencephalography (MEG) measures magnetic fluctuations in the brain generated by neural processes, some of which, such as cardiac signals, are generally removed as artifacts and discarded. However, heart rate variability (HRV) has long been regarded as a biomarker related to autonomic function, suggesting the cardiac signal in MEG contains valuable information that can provide supplemental health information about a patient. To enable access to these ancillary HRV data, we created an automated extraction tool capable of capturing HRV directly from raw MEG data with artificial intelligence. Five scans were conducted with simultaneous MEG and electrocardiogram (ECG) acquisition, which provides a ground truth metric for assessing our algorithms and data processing pipeline. In addition to directly comparing R-peaks between the MEG and ECG signals, this work explores the variation of the corresponding HRV output in time, frequency, and non-linear domains. After removing outlier intervals and aligning the ECG and derived cardiac MEG signals, the RMSE between the RR-intervals of each was RMSE1 = 2 ms, RMSE2 = 2 ms, RMSE3 = 8 ms, RMSE4 = 4 ms, RMSE5 = 13 ms. The findings indicate that cardiac artifacts from MEG data carry sufficient signal to approximate an individual's HRV metrics.
RESUMEN
BACKGROUND: Implementation of an anesthesiology-led cardiac implantable electronic device (CIED) service can be viewed to have economic and efficiency challenges. This study evaluates the cost savings of an anesthesiology-led CIED service. METHODS: A total of 830 patients presented in the pre-implementation period from 1 March 2016 to 31 December 2017, and 1981 patients presented in the post-implementation period from 1 January 2018 to 31 October 2021. Interrupted time-series analysis for single-group comparisons was used to evaluate the cost savings resulting from reduction in operating room (OR) start delays for patients with CIEDs. RESULTS: OR start-time delay was reduced by 10.6 min (95%CI: -20.5 to -0.83), comparing pre- to post-implementation. For an OR cost of USD 45/min, we estimated the direct cost to the department to be USD 1.68/min. The intervention translated into a total cost reduction during the intervention period of USD 250,000 (USD 18,000 to USD 470,000) per year for the institution and USD 9800 (USD 730 to USD 17,000) per year for the department. The yearly cost of employing a full-time team of CIED specialists would have been USD 135,456. The service triggered electrophysiology consultation on 13 device malfunctions. CONCLUSIONS: An anesthesiology-led CIED service resulted in substantial cost savings, increased OR efficiency and patient safety.
RESUMEN
In this Pro-Con commentary article, we discuss whether or not code sharing should be mandatory for scientific publications. Scientific programming is an increasingly prevalent tool in research. However, there are not unified guidelines for code availability requirements. Some journals require code sharing. Others require code descriptions. Yet others have no policies around code sharing. The Pro side presented here argues that code sharing should be mandatory for all scientific publications involving code. This Pro argument comes in 2 parts. First, any defensible reason for not sharing code is an equally valid a reason for the manuscript itself not being published. Second, lack of code sharing requirements creates 2 tiers of science: one where reproducibility is required and one where it is not. Additionally, the Pro authors suggest that a debate over code sharing is itself a decade out-of-date due to the emerging availability of containerization and virtual environment sharing software. The Pro argument concludes with an appeal that authors release code to make their work more understandable by other researchers. The Con side presented here argues that computer source codes of medical technology equipment should not be subject to mandatory public disclosure. The source code is a crucial part of what makes a particular device unique and allows that device to outperform its competition. The Con authors believe that public disclosure of this proprietary information would destroy all incentives for businesses to develop new and improved technologies. Competition in the free marketplace is what drives companies to constantly improve their products, to develop new and better medical devices. The open disclosure of these "trade secret" details would effectively end that competitive drive. Why invest time, money, and energy developing a "better mousetrap" if your competitors can copy it and produce it the next day?
Asunto(s)
Comercio , Reproducibilidad de los ResultadosRESUMEN
As implementation of artificial intelligence grows more prevalent in perioperative medicine, a clinician's ability to distinguish differentiating aspects of these algorithms is critical. There are currently numerous marketing and technical terms to describe these algorithms with little standardization. Additionally, the need to communicate with algorithm developers is paramount to actualize effective and practical implementation. Of particular interest in these discussions is the extent to which the output or predictions of algorithms and tools are understandable by medical practitioners. This work proposes a simple nomenclature that is intelligible to both clinicians and developers for quickly describing the interpretability of model results. There are three high-level categories: transparent, translucent, and opaque. To demonstrate the applicability and utility of this terminology, these terms were applied to the artificial intelligence and machine-learning-based products that have gained Food and Drug Administration approval. During this review and categorization process, 22 algorithms were found with perioperative utility (in a database of 70 total algorithms), and 12 of these had publicly available citations. The primary aim of this work is to establish a common nomenclature that will expedite and simplify descriptions of algorithm requirements from clinicians to developers and explanations of appropriate model use and limitations from developers to clinicians.
RESUMEN
Community structure, including relationships between and within groups, is foundational to our understanding of the world around us. For dissimilarity-based data, leveraging social concepts of conflict and alignment, we provide an approach for capturing meaningful structural information resulting from induced local comparisons. In particular, a measure of local (community) depth is introduced that leads directly to a probabilistic partitioning conveying locally interpreted closeness (or cohesion). A universal choice of threshold for distinguishing strongly and weakly cohesive pairs permits consideration of both local and global structure. Cases in which one might benefit from use of the approach include data with varying density such as that arising as snapshots of complex processes in which differing mechanisms drive evolution locally. The inherent recalibrating in response to density allows one to sidestep the need for localizing parameters, common to many existing methods. Mathematical results together with applications in linguistics, cultural psychology, and genetics, as well as to benchmark clustering data have been included. Together, these demonstrate how meaningful community structure can be identified without additional inputs (e.g., number of clusters or neighborhood size), optimization criteria, iterative procedures, or distributional assumptions.
Asunto(s)
Modelos Teóricos , Características de la Residencia , Ciencias Sociales , Algoritmos , HumanosAsunto(s)
Encéfalo/fisiología , Circulación Cerebrovascular/fisiología , Homeostasis/fisiología , Anciano , Anciano de 80 o más Años , Presión Sanguínea/fisiología , Estudios de Factibilidad , Femenino , Humanos , Cuidados Intraoperatorios/métodos , Masculino , Persona de Mediana Edad , Estudios Retrospectivos , Espectroscopía Infrarroja Corta/métodosRESUMEN
Developing an accurate and interpretable model to predict an individual's risk for Coronavirus Disease 2019 (COVID-19) is a critical step to efficiently triage testing and other scarce preventative resources. To aid in this effort, we have developed an interpretable risk calculator that utilized de-identified electronic health records (EHR) from the University of Alabama at Birmingham Informatics for Integrating Biology and the Bedside (UAB-i2b2) COVID-19 repository under the U-BRITE framework. The generated risk scores are analogous to commonly used credit scores where higher scores indicate higher risks for COVID-19 infection. By design, these risk scores can easily be calculated in spreadsheets or even with pen and paper. To predict risk, we implemented a Credit Scorecard modeling approach on longitudinal EHR data from 7,262 patients enrolled in the UAB Health System who were evaluated and/or tested for COVID-19 between January and June 2020. In this cohort, 912 patients were positive for COVID-19. Our workflow considered the timing of symptoms and medical conditions and tested the effects by applying different variable selection techniques such as LASSO and Elastic-Net. Within the two weeks before a COVID-19 diagnosis, the most predictive features were respiratory symptoms such as cough, abnormalities of breathing, pain in the throat and chest as well as other chronic conditions including nicotine dependence and major depressive disorder. When extending the timeframe to include all medical conditions across all time, our models also uncovered several chronic conditions impacting the respiratory, cardiovascular, central nervous and urinary organ systems. The whole pipeline of data processing, risk modeling and web-based risk calculator can be applied to any EHR data following the OMOP common data format. The results can be employed to generate questionnaires to estimate COVID-19 risk for screening in building entries or to optimize hospital resources.
RESUMEN
OBJECTIVES: Renal hypoperfusion is a common mechanism of cardiac surgery-related acute kidney injury (CS-AKI). However, the optimal amount of volume resuscitation to correct systemic hypoperfusion and prevent the postoperative development of CS-AKI has been a subject of debate. The goal of this study was to assess the association of volume responsiveness determined by stroke volume variation using the passive leg raise test (PLRT) at chest closure, with the development of CS-AKI according to the Kidney Disease Improving Global Outcomes criteria. DESIGN: Single-center, prospective observational study. SETTING: Tertiary hospital. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 131 patients were studied from January 2015 until May 2017. All patients underwent cardiac surgery that required cardiopulmonary bypass. Volume responsiveness was assessed at chest closure using the PRLT. Stroke volume variation from the sitting to the recumbent positions was measured by transesophageal echocardiography. Fluid responsiveness was defined as an increase of >12% of stroke volume from sitting to recumbent positions. A total of 82 (68.3%) patients were fluid-responsive versus 38 (31.6%) who were fluid-unresponsive. CS-AKI occurred in 30% of patients. There was no difference in CS-AKI between fluid-responsive and fluid-nonresponsive groups. However, CS-AKI was associated independently with an increases in body mass index and preoperative diastolic blood pressure. CS-AKI also was associated with prolonged intensive care unit length of stay. CONCLUSION: End-of-procedure volume responsiveness is not associated with a high risk for postoperative CS-AKI.
Asunto(s)
Lesión Renal Aguda , Procedimientos Quirúrgicos Cardíacos , Lesión Renal Aguda/diagnóstico , Lesión Renal Aguda/epidemiología , Lesión Renal Aguda/etiología , Procedimientos Quirúrgicos Cardíacos/efectos adversos , Puente Cardiopulmonar/efectos adversos , Humanos , Pierna , Complicaciones Posoperatorias/diagnóstico por imagen , Complicaciones Posoperatorias/epidemiología , Estudios Prospectivos , Factores de RiesgoRESUMEN
Thrombin is a key component for chemotherapeutic and antithrombotic therapy development. As the physiologic and pathologic roles of the light chain still remain vague, here, we continue previous efforts to understand the impacts of the disease-associated single deletion of LYS9 in the light chain. By combining supervised and unsupervised machine learning methodologies and more traditional structural analyses on data from 10 µs molecular dynamics simulations, we show that the conformational ensemble of the ΔK9 mutant is significantly perturbed. Our analyses consistently indicate that LYS9 deletion destabilizes both the catalytic cleft and regulatory functional regions and result in some conformational changes that occur in tens to hundreds of nanosecond scaled motions. We also reveal that the two forms of thrombin each prefer a distinct binding mode of a Na+ ion. We expand our understanding of previous experimental observations and shed light on the mechanisms of the LYS9 deletion associated bleeding disorder by providing consistent but more quantitative and detailed structural analyses than early studies in literature. With a novel application of supervised learning, i.e. the decision tree learning on the hydrogen bonding features in the wild-type and ΔK9 mutant forms of thrombin, we predict that seven pairs of critical hydrogen bonding interactions are significant for establishing distinct behaviors of wild-type thrombin and its ΔK9 mutant form. Our calculations indicate the LYS9 in the light chain has both localized and long-range allosteric effects on thrombin, supporting the opinion that light chain has an important role as an allosteric effector.
Asunto(s)
Aprendizaje Automático , Simulación de Dinámica Molecular , Mutación , Trombina/química , Trombina/genética , Regulación Alostérica , Humanos , Enlace de Hidrógeno , Unión Proteica , Conformación Proteica , Sodio/metabolismo , Trombina/metabolismoRESUMEN
Here we present a time-dependent correlation method that provides insight into how long a system takes to grow into its equal-time (Pearson) correlation. We also show a usage of an extant time-lagged correlation method that indicates the time for parts of a system to become decorrelated, relative to equal-time correlation. Given a completed simulation (or set of simulations), these tools estimate (i) how long of a simulation of the same system would be sufficient to observe the same correlated motions, (ii) if patterns of observed correlated motions indicate events beyond the timescale of the simulation, and (iii) how long of a simulation is needed to observe these longer timescale events. We view this method as a decision-support tool that will aid researchers in determining necessary sampling times. In principle, this tool is extendable to any multidimensional time series data with a notion of correlated fluctuations; however, here we limit our discussion to data from molecular-dynamics simulations.
RESUMEN
The concept of a cluster or community in a network context has been of considerable interest in a variety of settings in recent years. In this paper, employing random walks and geodesic distance, we introduce a unified measure of cluster-based proximity between nodes, relative to a given subset of interest. The inherent simplicity and informativeness of the approach could make it of value to researchers in a variety of scientific fields. Applicability is demonstrated via application to clustering for a number of existent data sets (including multipartite networks). We view community detection (i.e. when the full set of network nodes is considered) as simply the limiting instance of clustering (for arbitrary subsets). This perspective should add to the dialogue on what constitutes a cluster or community within a network. In regards to health-relevant attributes in social networks, identification of clusters of individuals with similar attributes can support targeting of collective interventions. The method performs well in comparisons with other approaches, based on comparative measures such as NMI and ARI.
Asunto(s)
Análisis por Conglomerados , Modelos Teóricos , Algoritmos , Animales , Gatos , Red Nerviosa/citología , Red Nerviosa/fisiología , Red SocialRESUMEN
Correlated motion analysis provides a method for understanding communication between and dynamic similarities of biopolymer residues and domains. The typical equal-time correlation matrices-frequently visualized with pseudo-colorings or heat maps-quickly convey large regions of highly correlated motion but hide more subtle similarities of motion. Here we propose a complementary method for visualizing correlations within proteins (or general biopolymers) that quickly conveys intuition about which residues have a similar dynamic behavior. For grouping residues, we use the recently developed non-parametric clustering algorithm HDBSCAN. Although the method we propose here can be used to group residues using correlation as a similarity matrix-the most straightforward and intuitive method-it can also be used to more generally determine groups of residues which have similar dynamic properties. We term these latter groups "Dynamic Domains", as they are based not on spatial closeness but rather closeness in the column space of a correlation matrix. We provide examples of this method across three human proteins of varying size and function-the Nf-Kappa-Beta essential modulator, the clotting promoter Thrombin and the mismatch repair protein (dimer) complex MutS-alpha. Although the examples presented here are from all-atom molecular dynamics simulations, this visualization technique can also be used on correlations matrices built from any ensembles of conformations from experiment or computation.
Asunto(s)
Algoritmos , Simulación de Dinámica Molecular , Movimiento (Física) , Proteínas/química , Programas Informáticos , Proteínas/genéticaRESUMEN
Understanding the efficacy of and creating delivery mechanisms for therapeutic nucleic acids requires understanding structural and kinetic properties which allow these polymers to promote the death of cancerous cells. One molecule of interest is a 10 mer of FdUMP (5-fluoro-2'-deoxyuridine-5'-O-monophosphate) - also called F10. Here we investigate the structural and kinetic behavior of F10 in intracellular and extracellular solvent conditions along with non-biological conditions that may be efficacious in in vitro preparations of F10 delivery systems. From our all-atom molecular dynamics simulations totaling 80 microseconds, we predict that F10's phosphate groups form close-range interactions with calcium and zinc ions, with calcium having the highest affinity of the five ions investigated. We also predict that F10's interactions with magnesium, potassium and sodium are almost exclusively long-range interactions. In terms of intramolecular interactions, we find that F10 is least structured (in terms of hydrogen bonds among bases) in the 150 mM NaCl (extracellular-like solvent conditions) and most structured in 150 mM ZnCl2. Kinetically, we see that F10 is unstable in the presence of magnesium, sodium or potassium, finding stable kinetic traps in the presence of calcium or zinc.
Asunto(s)
ADN/química , Simulación de Dinámica Molecular , Calcio/química , Análisis por Conglomerados , ADN/metabolismo , Desoxiuridina/análogos & derivados , Desoxiuridina/química , Enlace de Hidrógeno , Iones/química , Cadenas de Markov , Conformación de Ácido Nucleico , Análisis de Componente Principal , Zinc/químicaRESUMEN
Thrombin is a multifunctional enzyme that plays an important role in blood coagulation, cell growth, and metastasis. Depending upon the binding of sodium ions, thrombin presents significantly different enzymatic activities. In the environment with sodium ions, thrombin is highly active in cleaving the coagulated substrates and this is referred to as the "fast" form; in the environment without sodium ions, thrombin turns catalytically less active and is in the "slow" form. Although many experimental studies over the last two decades have attempted to reveal the structural and kinetic differences between these two forms, it remains vague and disputed how the functional switch between the "fast" and "slow" forms is mediated by Na+ cations. In this work, we employ microsecond-scale all-atom molecular dynamics simulations to investigate the differences in the structural ensembles in sodium-bound/unbound and potassium-bound/unbound thrombin. Our calculations indicate that the regulatory regions, including the 60s, γ loops, and exosite I and II, are primarily affected by both the bound and unbound cations. Conformational free energy surfaces, estimated from principal component analysis, further reveal the existence of multiple conformational states. The binding of a cation introduces changes in the distribution of these states. Through comparisons with potassium-binding, the binding of sodium ions appears to shift the population toward conformational states that might be catalytically favorable. Our study of thrombin in the presence of sodium/potassium ions suggests Na+-mediated generalized allostery is the mechanism of thrombin's functional switch between the "fast" and "slow" forms.
Asunto(s)
Coagulación Sanguínea , Simulación de Dinámica Molecular , Trombina/fisiología , Sitios de Unión , Cinética , Potasio , Conformación ProteicaRESUMEN
Given their increasingly frequent usage, understanding the chemical and structural properties which allow therapeutic nucleic acids to promote the death of cancer cells is critical for medical advancement. One molecule of interest is a 10-mer of FdUMP (5-fluoro-2'-deoxyuridine-5'-O-monophosphate) also called F10. To investigate causes of structural stability, we have computationally restored the 2' oxygen on each ribose sugar of the phosphodiester backbone, creating FUMP[10]. Microsecond time-scale, all-atom, simulations of FUMP[10] in the presence of 150 mM MgCl2 predict that the strand has a 45% probability of folding into a stable hairpin-like secondary structure. Analysis of 16 µs of data reveals phosphate interactions as likely contributors to the stability of this folded state. Comparison with polydT and polyU simulations predicts that FUMP[10]'s lowest order structures last for one to 2 orders of magnitude longer than similar nucleic acid strands. Here we provide a brief structural and conformational analysis of the predicted structures of FUMP[10], and suggest insights into its stability via comparison to F10, polydT, and polyU.