Harnessing Machine Learning to Improve Patient Outcomes in Pulmonary and Critical Care Medicine.
Ontology highlight
ABSTRACT:
SUBMITTER: Levenson AR
PROVIDER: S-EPMC7528791 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:The theory and computational tools developed to interpret and explore energy landscapes in molecular science are applied to the landscapes defined by local minima for neural networks. These machine learning landscapes correspond to fits of training data, where the inputs are vital signs and laboratory measurements for a database of patients, and the objective is to predict a clinical outcome. In this contribution, we test the predictions obtained by fitting to single measurements, and then to combinations of between 2 and 10 different patient medical data items. The effect of including measurements over different time intervals from the 48 h period in question is analysed, and the most recent values are found to be the most important. We also compare results obtained for neural networks as a function of the number of hidden nodes, and for different values of a regularization parameter. The predictions are compared with an alternative convex fitting function, and a strong correlation is observed. The dependence of these results on the patients randomly selected for training and testing decreases systematically with the size of the database available. The machine learning landscapes defined by neural network fits in this investigation have single-funnel character, which probably explains why it is relatively straightforward to obtain the global minimum solution, or a fit that behaves similarly to this optimal parameterization.
| S-EPMC5541539 | biostudies-other
Project description:Background: Little is known about historical and recent application trends for pulmonary critical care medicine (PCCM) or pulmonary medicine (PM) fellowship programs. Describing trends in and characteristics of PCCM and PM applications, applicants, and fellowship programs can help program directors and medical educators understand trainees' interest in and application patterns for these fellowship programs. Objective: The objective of this study was to use National Residency Match Program data to assess recent trends in PCCM and PM fellowship applications and compare characteristics of applicants and fellowship programs. Methods: In 2019, we used National Residency Match Program data to evaluate applicant ranking and matching in PCCM and PM fellowship programs and to compare applicant and fellowship program characteristics. Results: From 2008 through 2019, the majority of applicants (59.1%) matched into PCCM were graduates of U.S. allopathic or osteopathic medical schools, whereas 87% of PM fellows were non-U.S. graduates. PCCM was the preferred specialty for 90.8% of matched applicants versus only 31.6% of matched PM applicants (P < 0.001). The match rate for PCCM applicants was 67.2% versus 23.8% for PM applicants (P < 0.001). Of PCCM applicants, 36.6% matched into their top choice versus 10.8% of PM applicants (P < 0.001). There are far fewer PM fellowship positions (n = 23) and programs (n = 12) than PCCM positions (n = 450) and programs (n = 131). The mean fill rates from the 2004 through 2016 appointment years are 94.1% in PCCM and 97.4% in PM (P = 0.009). Conclusion: PCCM is a prevailing specialty choice over PM among residency graduates, with matched applicants more likely to list PCCM than PM as their preferred specialty. Further exploration into applicants' interest in critical care compared with PM may prove beneficial in guiding applicants to programs that will best meet their career goals.
| S-EPMC8043272 | biostudies-literature
Project description:GDF-15 (growth differentiation factor 15) acts both as a stress-induced cytokine with diverse actions at different body sites and as a cell-autonomous regulator linked to cellular senescence and apoptosis. For multiple reasons, this divergent transforming growth factor-β molecular superfamily member should be better known to pulmonary researchers and clinicians. In ambulatory individuals, GDF-15 concentrations in peripheral blood are an established predictive biomarker of all-cause mortality and of adverse cardiovascular events. Concentrations upon admission of critically ill patients (without or with sepsis) correlate with organ dysfunction and independently predict short- and long-term mortality risk. GDF-15 is a major downstream mediator of p53 activation, but it can also be induced independently of p53, notably by nonsteroidal antiinflammatory agents. GDF-15 blood concentrations are markedly elevated in adults and children with pulmonary hypertension. Concentrations are also increased in chronic obstructive pulmonary disease, in which they contribute to mucus hypersecretion, airway epithelial cell senescence, and impaired antiviral defenses, which together with murine data support a role for GDF-15 in chronic obstructive pulmonary disease pathogenesis and progression. This review summarizes biological and clinical data on GDF-15 relevant to pulmonary and critical care medicine. We highlight the recent discovery of a central nervous system receptor for GDF-15, GFRAL (glial cell line-derived neurotrophic factor family receptor-α-like), an important advance with potential for novel treatments for obesity and cachexia. We also describe limitations and controversies in the existing literature, and we delineate research questions that must be addressed to determine whether GDF-15 can be therapeutically manipulated in other clinical settings.
| S-EPMC6543753 | biostudies-literature
Project description:Examples of fully integrated machine learning models that drive clinical care are rare. Despite major advances in the development of methodologies that outperform clinical experts and growing prominence of machine learning in mainstream medical literature, major challenges remain. At Duke Health, we are in our fourth year developing, piloting, and implementing machine learning technologies in clinical care. To advance the translation of machine learning into clinical care, health system leaders must address barriers to progress and make strategic investments necessary to bring health care into a new digital age. Machine learning can improve clinical workflows in subtle ways that are distinct from how statistics has shaped medicine. However, most machine learning research occurs in siloes, and there are important, unresolved questions about how to retrain and validate models post-deployment. Academic medical centers that cultivate and value transdisciplinary collaboration are ideally suited to integrate machine learning in clinical care. Along with fostering collaborative environments, health system leaders must invest in developing new capabilities within the workforce and technology infrastructure beyond standard electronic health records. Now is the opportunity to break down barriers and achieve scalable growth in the number of high-impact collaborations between clinical researchers and machine learning experts to transform clinical care.
| S-EPMC6354017 | biostudies-other
Project description:BackgroundEndotracheal intubation in the intensive care unit (ICU) is a high-risk procedure. Competence in endotracheal intubation is a requirement for Pulmonary and Critical Care Medicine (PCCM) training programs, but fellow experience as the primary operator in intubating ICU patients has not been described on a large scale.ObjectiveWe hypothesized that significant variation surrounding endotracheal intubation practices in medical ICUs exists in United States (US) PCCM training programs.MethodsWe administered a survey to a convenience sample of US PCCM fellows to elicit typical intubation practices in the medical ICU.Results89 discrete US PCCM and Internal Medicine CCM training programs (77% response rate) were represented. At 43% of programs, the PCCM fellow was "always or almost always" designated the primary operator for intubation of a medical ICU patient, whereas at 21% of programs, the PCCM fellow was "rarely or never" the primary operator responsible for intubating in the ICU. Factors influencing this variation included time of day, hospital policies, attending skill or preference, ICU census and acuity, and patient factors. There was an association between location of the training program, but not program size, and whether the PCCM fellow was the primary operator.ConclusionThere is significant variation in whether PCCM fellows are the primary operators to intubate medical ICU patients during training. Further work should explore how this variation affects fellow career development and competence in intubation.
| S-EPMC7959255 | biostudies-literature
Project description: Not available
| S-EPMC8357065 | biostudies-literature
Project description: Not available
| S-EPMC7706163 | biostudies-literature
Project description:Spurred by advances in processing power, memory, storage, and an unprecedented wealth of data, computers are being asked to tackle increasingly complex learning tasks, often with astonishing success. Computers have now mastered a popular variant of poker, learned the laws of physics from experimental data, and become experts in video games - tasks that would have been deemed impossible not too long ago. In parallel, the number of companies centered on applying complex data analysis to varying industries has exploded, and it is thus unsurprising that some analytic companies are turning attention to problems in health care. The purpose of this review is to explore what problems in medicine might benefit from such learning approaches and use examples from the literature to introduce basic concepts in machine learning. It is important to note that seemingly large enough medical data sets and adequate learning algorithms have been available for many decades, and yet, although there are thousands of papers applying machine learning algorithms to medical data, very few have contributed meaningfully to clinical care. This lack of impact stands in stark contrast to the enormous relevance of machine learning to many other industries. Thus, part of my effort will be to identify what obstacles there may be to changing the practice of medicine through statistical learning approaches, and discuss how these might be overcome.
| S-EPMC5831252 | biostudies-literature
Project description:Degradation of intracellular proteins in Gram-negative bacteria regulates various cellular processes and serves as a quality control mechanism by eliminating damaged proteins. To understand what causes the proteolytic machinery of the cell to degrade some proteins while sparing others, we employed a quantitative pulsed-SILAC (Stable Isotope Labeling with Amino acids in Cell culture) method followed by mass spectrometry analysis to determine the half-lives for the proteome of exponentially growing Escherichia coli, under standard conditions. We developed a likelihood-based statistical test to findactively degraded proteins, and identified dozens of novel proteins that are fast-degrading. Finally, we used structural, physicochemical and protein-protein interaction network descriptorsto train a machine-learning classifier to discriminate fast-degrading proteins from the rest of the proteome. Our combined computational-experimental approach provides means for proteomic-based discovery of fast degrading proteins in bacteria and the elucidation of the factors determining protein half-livesand have implications for protein engineering. Moreover, as rapidly degraded proteins may play an important role in pathogenesis, our findings could identify new potential antibacterial drug targets
2021-01-13 | PXD022112 | Pride
Project description:Critical thinking, the capacity to be deliberate about thinking, is increasingly the focus of undergraduate medical education, but is not commonly addressed in graduate medical education. Without critical thinking, physicians, and particularly residents, are prone to cognitive errors, which can lead to diagnostic errors, especially in a high-stakes environment such as the intensive care unit. Although challenging, critical thinking skills can be taught. At this time, there is a paucity of data to support an educational gold standard for teaching critical thinking, but we believe that five strategies, routed in cognitive theory and our personal teaching experiences, provide an effective framework to teach critical thinking in the intensive care unit. The five strategies are: make the thinking process explicit by helping learners understand that the brain uses two cognitive processes: type 1, an intuitive pattern-recognizing process, and type 2, an analytic process; discuss cognitive biases, such as premature closure, and teach residents to minimize biases by expressing uncertainty and keeping differentials broad; model and teach inductive reasoning by utilizing concept and mechanism maps and explicitly teach how this reasoning differs from the more commonly used hypothetico-deductive reasoning; use questions to stimulate critical thinking: "how" or "why" questions can be used to coach trainees and to uncover their thought processes; and assess and provide feedback on learner's critical thinking. We believe these five strategies provide practical approaches for teaching critical thinking in the intensive care unit.
| S-EPMC5461985 | biostudies-literature