Antibiofilm Agents for the Treatment and Prevention of Bacterial Vaginosis: A Systematic Narrative Review.

BACKGROUND: Bacterial vaginosis (BV) is difficult to eradicate due to BV biofilms protecting BV bacteria (Gardnerella, Prevotella, and other genera). With the growing understanding of biofilms, we systematically reviewed the current knowledge on the efficacy of anti-BV biofilm agents. METHODS: We searched literature in the Scopus, Medline, and Embase databases for empirical studies investigating substances for the treatment of BV biofilms or prevention of their recurrence and their efficacy and/or safety. RESULTS: Of 201 unique titles, 35 satisfied the inclusion criteria. Most studies (89%) reported on preclinical laboratory research on the efficacy of experimental antibiofilm agents (80%) rather than their safety. Over 50% were published within the past 5 years. Agents were classified into 7 groups: antibiotics, antiseptics, cationic peptides, enzymes, plant extracts, probiotics, and surfactants/surfactant components. Enzymes and probiotics were most commonly investigated. Earlier reports of antibiotics having anti-BV biofilm activity have not been confirmed. Some compounds from other classes demonstrated promising anti-BV biofilm efficacy in early studies. CONCLUSIONS: Further research is anticipated on successful antibiofilm agents. If confirmed as effective and safe in human clinical trials, they may offer a breakthrough in BV treatment. With rising antibiotic resistance, antibiofilm agents will significantly improve the current standard of care for BV management.

Determining the Accuracy of Intestinal Ultrasound Scores as a Prescreening Tool in Crohn's Disease Clinical Trials.

INTRODUCTION: High rates of screen failure for the minimum Simple Endoscopic Score for Crohn's Disease (SES-CD) plague Crohn's disease (CD) clinical trials. We aimed to determine the accuracy of segmental intestinal ultrasound (IUS) parameters and scores to detect segmental SES-CD activity. METHODS: A single-center, blinded, cross-sectional cohort study of children and young adult patients with CD undergoing IUS and ileocolonoscopy, comparing segmental IUS bowel wall thickness (BWT), hyperemia (modified Limberg score [MLS]), and scores to detect segmental SES-CD activity: (i) SES-CD ≤2, (ii) SES-CD ≥6, and (iii) SES-CD ≥4 in the terminal ileum (TI) only. Primary outcome was accuracy of BWT, MLS, and IUS scores to detect SES-CD ≤2 and SES-CD ≥6. Secondary outcomes were accuracy of TI BWT, MLS, and IUS scores to detect SES-CD ≥4 and correlation with the SES-CD. RESULTS: Eighty-two patients (median [interquartile range] age 16.5 [12.9-20.0] years) underwent IUS and ileocolonoscopy of 323 bowel segments. Segmental BWT ≤3.1 mm had a similar high accuracy to detect SES-CD ≤2 as IUS scores (area under the receiver operating curve [AUROC] 0.833 [95% confidence interval 0.76-0.91], 94% sensitivity, and 73% specificity). Segmental BWT ≥3.6 mm and ≥4.3 mm had similar high accuracy to detect SES-CD ≥6 (AUROC 0.950 [95% confidence interval 0.92-0.98], 89% sensitivity, 93% specificity) in the colon and an SES-CD ≥4 in the TI (AUROC 0.874 [0.79-0.96], 80% sensitivity, and 91% specificity) as IUS scores. Segmental IUS scores strongly correlated with the SES-CD. DISCUSSION: Segmental IUS BWT is highly accurate to detect moderate-to-severe endoscopic inflammation. IUS may be the ideal prescreening tool to reduce unnecessary trial screen failures.

Prevalence and prognostic impact of complex percutaneous coronary intervention among octogenarians.

BACKGROUND: The number of octogenarians referred to percutaneous coronary interventions (PCI) is rising steadily. The prevalence and prognostic impact of complex PCI (CPCI) in this vulnerable population has not been fully evaluated. METHODS: Patients ≥80 years old who underwent PCI between 2012 and 2019 at Mount Sinai Hospital were included. Patients were categorized based on PCI complexity, defined as the presence of at least one of the following criteria: use of atherectomy, total stent length ≥60 mm, ≥3 stents implanted, bifurcation treated with at least 2 stents, PCI involving ≥3 vessels, ≥3 lesions, left main, saphenous vein graft or chronic total occlusion. The primary outcome was major adverse cardiovascular events (MACE), a composite of all-cause death, myocardial infarction (MI), or target-vessel revascularization (TVR), within 1 year after PCI. Secondary outcomes included major bleeding. RESULTS: Among 2657 octogenarians, 1387 (52%) underwent CPCI and were more likely to be men and to have cardiovascular risk factors or comorbidities. CPCI as compared with no-CPCI was associated with a higher 1-year risk of MACE (16.6% vs. 11.1%, adjusted HR 1.3, 95% CI 1.06-1.77, p value 0.017), due to an excess of MI and TVR, and major bleeding (10% vs. 5.8%, adjusted HR 1.64, 95% CI 1.20-2.55, p value 0.002). CONCLUSIONS: Among octogenarians, CPCI was associated with a significantly higher 1-year risk of MACE, due to higher rates of MI and TVR but not of all-cause death, and of major bleeding. Strategies to reduce complications should be implemented in octogenarians undergoing CPCI.

Development and Validation of a Natural Language Processing Model to Identify Low-Risk Pulmonary Embolism in Real Time to Facilitate Safe Outpatient Management.

STUDY OBJECTIVE: This study aimed to (1) develop and validate a natural language processing model to identify the presence of pulmonary embolism (PE) based on real-time radiology reports and (2) identify low-risk PE patients based on previously validated risk stratification scores using variables extracted from the electronic health record at the time of diagnosis. The combination of these approaches yielded an natural language processing-based clinical decision support tool that can identify patients presenting to the emergency department (ED) with low-risk PE as candidates for outpatient management. METHODS: Data were curated from all patients who received a PE-protocol computed tomography pulmonary angiogram (PE-CTPA) imaging study in the ED of a 3-hospital academic health system between June 1, 2018 and December 31, 2020 (n=12,183). The "preliminary" radiology reports from these imaging studies made available to ED clinicians at the time of diagnosis were adjudicated as positive or negative for PE by the clinical team. The reports were then divided into development, internal validation, and temporal validation cohorts in order to train, test, and validate an natural language processing model that could identify the presence of PE based on unstructured text. For risk stratification, patient- and encounter-level data elements were curated from the electronic health record and used to compute a real-time simplified pulmonary embolism severity (sPESI) score at the time of diagnosis. Chart abstraction was performed on all low-risk PE patients admitted for inpatient management. RESULTS: When applied to the internal validation and temporal validation cohorts, the natural language processing model identified the presence of PE from radiology reports with an area under the receiver operating characteristic curve of 0.99, sensitivity of 0.86 to 0.87, and specificity of 0.99. Across cohorts, 10.5% of PE-CTPA studies were positive for PE, of which 22.2% were classified as low-risk by the sPESI score. Of all low-risk PE patients, 74.3% were admitted for inpatient management. CONCLUSION: This study demonstrates that a natural language processing-based model utilizing real-time radiology reports can accurately identify patients with PE. Further, this model, used in combination with a validated risk stratification score (sPESI), provides a clinical decision support tool that accurately identifies patients in the ED with low-risk PE as candidates for outpatient management.

Development of a Human Immunodeficiency Virus Risk Prediction Model Using Electronic Health Record Data From an Academic Health System in the Southern United States.

BACKGROUND: Human immunodeficiency virus (HIV) pre-exposure prophylaxis (PrEP) is underutilized in the southern United States. Rapid identification of individuals vulnerable to diagnosis of HIV using electronic health record (EHR)-based tools may augment PrEP uptake in the region. METHODS: Using machine learning, we developed EHR-based models to predict incident HIV diagnosis as a surrogate for PrEP candidacy. We included patients from a southern medical system with encounters between October 2014 and August 2016, training the model to predict incident HIV diagnosis between September 2016 and August 2018. We obtained 74 EHR variables as potential predictors. We compared Extreme Gradient Boosting (XGBoost) versus least absolute shrinkage selection operator (LASSO) logistic regression models, and assessed performance, overall and among women, using area under the receiver operating characteristic curve (AUROC) and area under precision recall curve (AUPRC). RESULTS: Of 998 787 eligible patients, 162 had an incident HIV diagnosis, of whom 49 were women. The XGBoost model outperformed the LASSO model for the total cohort, achieving an AUROC of 0.89 and AUPRC of 0.01. The female-only cohort XGBoost model resulted in an AUROC of 0.78 and AUPRC of 0.00025. The most predictive variables for the overall cohort were race, sex, and male partner. The strongest positive predictors for the female-only cohort were history of pelvic inflammatory disease, drug use, and tobacco use. CONCLUSIONS: Our machine-learning models were able to effectively predict incident HIV diagnoses including among women. This study establishes feasibility of using these models to identify persons most suitable for PrEP in the South.

Risk of in-hospital Deterioration for Children with Single Ventricle Physiology.

Children with single ventricle physiology (SV) are at high risk of in-hospital morbidity and mortality. Identifying children at risk for deterioration may allow for earlier escalation of care and subsequently decreased mortality.We conducted a retrospective chart review of all admissions to the pediatric cardiology non-ICU service from 2014 to 2018 for children < 18 years old. We defined clinical deterioration as unplanned transfer to the ICU or inpatient mortality. We selected children with SV by diagnosis codes and defined infants as children < 1 year old. We compared demographic, vital sign, and lab values between infants with and without a deterioration event. We evaluated vital sign and medical therapy changes before deterioration events.Among infants with SV (129 deterioration events over 225 admissions, overall 25% with hypoplastic left heart syndrome), those who deteriorated were younger (p = 0.001), had lower baseline oxygen saturation (p = 0.022), and higher baseline respiratory rate (p = 0.022), heart rate (p = 0.023), and hematocrit (p = 0.008). Median Duke Pediatric Early Warning Score increased prior to deterioration (p < 0.001). Deterioration was associated with administration of additional oxygen support (p = 0.012), a fluid bolus (p < 0.001), antibiotics (p < 0.001), vasopressor support (p = 0.009), and red blood cell transfusion (p < 0.001).Infants with SV are at high risk for deterioration. Integrating baseline and dynamic patient data from the electronic health record to identify the highest risk patients may allow for earlier detection and intervention to prevent clinical deterioration.

Development and Validation of Machine Learning Models to Predict Admission From Emergency Department to Inpatient and Intensive Care Units.

STUDY OBJECTIVE: This study aimed to develop and validate 2 machine learning models that use historical and current-visit patient data from electronic health records to predict the probability of patient admission to either an inpatient unit or ICU at each hour (up to 24 hours) of an emergency department (ED) encounter. The secondary goal was to provide a framework for the operational implementation of these machine learning models. METHODS: Data were curated from 468,167 adult patient encounters in 3 EDs (1 academic and 2 community-based EDs) of a large academic health system from August 1, 2015, to October 31, 2018. The models were validated using encounter data from January 1, 2019, to December 31, 2019. An operational user dashboard was developed, and the models were run on real-time encounter data. RESULTS: For the intermediate admission model, the area under the receiver operating characteristic curve was 0.873 and the area under the precision-recall curve was 0.636. For the ICU admission model, the area under the receiver operating characteristic curve was 0.951 and the area under the precision-recall curve was 0.461. The models had similar performance in both the academic- and community-based settings as well as across the 2019 and real-time encounter data. CONCLUSION: Machine learning models were developed to accurately make predictions regarding the probability of inpatient or ICU admission throughout the entire duration of a patient's encounter in ED and not just at the time of triage. These models remained accurate for a patient cohort beyond the time period of the initial training data and were integrated to run on live electronic health record data, with similar performance.

Renoprotective and Immunomodulatory Effects of GDF15 following AKI Invoked by Ischemia-Reperfusion Injury.

BACKGROUND: Gdf15 encodes a TGF-ß superfamily member that is rapidly activated in response to stress in multiple organ systems, including the kidney. However, there has been a lack of information about Gdf15 activity and effects in normal kidney and in AKI. METHODS: We used genome editing to generate a Gdf15nuGFP-CE mouse line, removing Gdf15 at the targeted allele, and enabling direct visualization and genetic modification of Gdf15-expressing cells. We extensively mapped Gdf15 expression in the normal kidney and following bilateral ischemia-reperfusion injury, and quantified and compared renal responses to ischemia-reperfusion injury in the presence and absence of GDF15. In addition, we analyzed single nucleotide polymorphism association data for GDF15 for associations with patient kidney transplant outcomes. RESULTS: Gdf15 is normally expressed within aquaporin 1-positive cells of the S3 segment of the proximal tubule, aquaporin 1-negative cells of the thin descending limb of the loop of Henle, and principal cells of the collecting system. Gdf15 is rapidly upregulated within a few hours of bilateral ischemia-reperfusion injury at these sites and new sites of proximal tubule injury. Deficiency of Gdf15 exacerbated acute tubular injury and enhanced inflammatory responses. Analysis of clinical transplantation data linked low circulating levels of GDF15 to an increased incidence of biopsy-proven acute rejection. CONCLUSIONS: Gdf15 contributes to an early acting, renoprotective injury response, modifying immune cell actions. The data support further investigation in clinical model systems of the potential benefit from GDF15 administration in situations in which some level of tubular injury is inevitable, such as following a kidney transplant.

Calcitriol loading before total parathyroidectomy with autotransplant in patients with end-stage kidney disease: does it prevent postoperative hypocalcaemia?

BACKGROUND: Hungry bone syndrome (HBS) is one of the most serious complications following parathyroidectomy for severe hyperparathyroidism. There is a lack of literature informing the treatment and risk factors for this condition and the ideal pre-operative strategy for prevention. AIMS: The primary aims were to examine the incidence of HBS with pre-operative calcitriol loading for 10 days and to determine the risk factors for HBS. The secondary aims were to determine the rate of intravenous calcium replacement in those with HBS and to assess whether cinacalcet removal has increased rates of parathyroidectomy in the end-stage kidney disease population. METHODS: We performed a retrospective study from 2011 to 2018 on 45 patients with end-stage kidney disease undergoing total parathyroidectomy with autotransplantation for severe hyperparathyroidism. This was based at the John Hunter and Newcastle Private Hospitals in New South Wales. RESULTS: 28.3% of patients with calcitriol loading undergoing parathyroidectomy fulfilled criteria for HBS. Pre-operative variables that were associated with HBS were elevated parathyroid hormone (P = 0.028) and longer duration of renal replacement therapy (P = 0.033). Rates of total parathyroidectomy were higher after the removal of calcimimetics from the Pharmaceutical Benefits Scheme (P = 0.0024). CONCLUSIONS: HBS remains a common complication of parathyroidectomy, even with prolonged high-dose calcitriol loading. This emphasises the need for further trials investigating other targeted therapies, such as bisphosphonates, to prevent HBS. Those most at risk of HBS are patients with high bone turnover and prolonged renal replacement therapy.

First Principles Calculation of the Entropy of Liquid Aluminum.

The information required to specify a liquid structure equals, in suitable units, its thermodynamic entropy. Hence, an expansion of the entropy in terms of multi-particle correlation functions can be interpreted as a hierarchy of information measures. Utilizing first principles molecular dynamics simulations, we simulate the structure of liquid aluminum to obtain its density, pair and triplet correlation functions, allowing us to approximate the experimentally measured entropy and relate the excess entropy to the information content of the correlation functions. We discuss the accuracy and convergence of the method.

Tutorial: Parallel Computing of Simulation Models for Risk Analysis.

Simulation models are widely used in risk analysis to study the effects of uncertainties on outcomes of interest in complex problems. Often, these models are computationally complex and time consuming to run. This latter point may be at odds with time-sensitive evaluations or may limit the number of parameters that are considered. In this article, we give an introductory tutorial focused on parallelizing simulation code to better leverage modern computing hardware, enabling risk analysts to better utilize simulation-based methods for quantifying uncertainty in practice. This article is aimed primarily at risk analysts who use simulation methods but do not yet utilize parallelization to decrease the computational burden of these models. The discussion is focused on conceptual aspects of embarrassingly parallel computer code and software considerations. Two complementary examples are shown using the languages MATLAB and R. A brief discussion of hardware considerations is located in the Appendix.

Topology-based Clustering of Functional Brain Networks in an Alzheimer's Disease Cohort.

Alzheimer's disease is a progressive neurodegenerative disease with many identifying biomarkers for diagnosis. However, whole-brain phenomena, particularly in functional MRI modalities, are not fully understood nor characterized. Here we employ the novel application of topological data analysis (TDA)-based methods of persistent homology to functional brain networks from ADNI-3 cohort to perform a subtyping experiment using unsupervised clustering techniques. We then investigate variations in QT-PAD challenge features across the identified clusters. Using a Wasserstein distance kernel with a variety of clustering algorithms, we found that the 0th-homology Wasserstein distance kernel and spectral clustering yielded clusters with significant differences in whole brain and medial temporal lobe (MTL) volume, thus demonstrating an intrinsic link between whole brain functional topology and brain morphometric structure. These findings demonstrate the importance of MTL in functional connectivity and the efficacy of using TDA-based machine learning methods in network neuroscience and neurodegenerative disease subtyping.

Second-generation everolimus-eluting intracoronary stents: a comprehensive review of the clinical evidence.

Percutaneous coronary intervention with implantation of second-generation drug-eluting stents (DES) has emerged as a mainstay for the treatment of obstructive coronary artery disease given its beneficial impact on clinical outcomes in these patients. Everolimus-eluting stents (EES) are one of the most frequently implanted second-generation DES; their use for the treatment of a wide range of patients including those with complex coronary lesions is supported by compelling evidence. Although newer stent platforms such as biodegradable polymer DES may lower local vessel inflammation, their efficacy and safety have not yet surpassed that of Xience stents. This article summarizes the properties of the Xience family of EES and the evidence supporting their use across diverse patient demographics and coronary lesion morphologies.

Patients with coronary artery disease (CAD) often require treatment for symptoms caused by blockages in coronary arteries. In addition to medical therapy, available procedure options include either coronary artery bypass grafting, a major heart surgery or percutaneous coronary intervention (PCI) with stenting. PCI is a minimally invasive procedure where a metallic stent (a mesh made up of fine metallic network in a tube shape used to keep vessels open) is advanced over a wire through an artery to open the coronary artery blockage. Over the past few decades, improvements in procedure technique and stent material have made PCI a highly safe and efficacious procedure. A newer generation of stents, known as drug-eluting stents (DES), have been developed in which metallic struts are covered with a highly biocompatible polymer (a thin material coating over the metallic mesh) that releases drugs at the blockage site to prevent local cell growth in the vessel wall. Among the second-generation DES, Xience everolimus-eluting stents (EES) have shown better outcomes compared with earlier generations of stents. Another version of DES with biodegradable polymer coating is emerging but their advantage over EES remains uncertain. Currently, Xience EES are one of the most commonly used stents to treat CAD. This manuscript covers an in-depth review of clinical evidence on the performance of Xience stents in a diverse range patient populations.

Before and after COVID-19: Changes in symptoms and diagnoses in 13,033 adults.

BACKGROUND: Most patients with COVID-19 report experiencing one or more symptoms after acute infection subsides, known as post-acute sequelae of SARS-CoV-2 infection (PASC). Though research has examined PASC after acute COVID-19, few studies have examined PASC over a longer follow-up duration or accounted for rates of symptoms and diagnoses before COVID-19 infection, and included those not actively seeking treatment for PASC. To determine what symptoms and diagnoses are occurring at higher rates after acute COVID-19 infection from a more inclusive sample, we extracted electronic hospital records (EHR) data from 13,033 adults with previously known diagnoses and symptoms. METHODS: The sample was comprised of patients who had a positive PCR test for SARS-CoV-2 between March 1, 2020, and December 31, 2020, and follow-up was conducted through November 29, 2021. All patients in the sample had medical appointments ≥4 weeks before and ≥4 weeks after their positive PCR test. At these appointments, all ICD-10 codes recorded in the EHR were classified into 21 categories based on the literature and expert review. Conditional logistic regression models were used to quantify the odds of these symptoms and diagnostic categories following COVID-19 infection relative to visits occurring before infection. The sample was comprised of 28.0% adults over 65 and was 57.0% female. After the positive PCR test, the most recorded diagnoses and symptoms were dyspnea and respiratory failure, myositis, musculoskeletal pain/stiffness, anxiety, and depression. RESULTS: Results from regression analyses showed increased odds of diagnosis for 15 of the 21 categories following positive PCR. Relative to pre-COVID, the diagnoses and symptoms with the greatest odds after a positive PCR test were loss of smell or taste [OR (95% CI) = 6.20 (3.18-12.09)], pulmonary fibrosis [3.50 (1.59-7.68)], and dyspnea/respiratory failure [2.14 (1.92-2.40)]. Stratification of these analyses by age, gender, race, and ethnicity showed similar results. CONCLUSION: The increased symptoms and diagnoses detected in the current study match prior analyses of PASC diagnosis and treatment-seeking patients. The current research expands upon the literature by showing that these symptoms are more frequently detected following acute COVID-19 than before COVID-19. Further, our analyses provide a broad snapshot of the population as we were able to describe PASC among all patients who tested positive for COVID-19.

SGLT2 Inhibitors, Functional Capacity, and Quality of Life in Patients With Heart Failure: A Systematic Review and Meta-Analysis.

Importance: The associations of sodium glucose cotransporter-2 inhibitors (SGLT2is) with reduction in mortality and hospitalization rates in patients with heart failure (HF) are well established. However, their association with improving functional capacity and quality of life (QOL) has been variably studied and less reported. Objective: To provide evidence on the extent to which SGLT2is are associated with improvement on objective measures of functional capacity and QOL in patients living with HF. Data Sources: The MEDLINE, EMBASE, and Cochrane databases were systematically searched for relevant articles on July 31, 2023. Study Selection: Randomized, placebo-controlled clinical trials reporting the effect of SGLT2i on functional outcomes of exercise capacity (peak oxygen consumption [peak VO2] or 6-minute walk distance [6MWD]) and/or QOL using validated questionnaires for patients with HF were included. Data Extraction and Synthesis: Data were extracted by 2 authors following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines, and a meta-analysis using the restricted maximum likelihood random-effects model was conducted. Main Outcomes and Measures: Outcomes of interest included changes in peak VO2, 6MWD, and Kansas City Cardiomyopathy Questionnaire-12 total symptom score (KCCQ-TSS), clinical summary score (KCCQ-CSS), and overall summary score (KCCQ-OSS). Results: In this meta-analysis of 17 studies, 23 523 patients (mean [range] age, 69 [60-75] years) were followed over a period ranging from 12 to 52 weeks. Four studies included peak VO2 as an outcome, 7 studies included 6MWD, and 10 studies reported KCCQ scores. Mean (SD) left ventricular ejection fraction was 43.5% (12.4%). Compared with controls, patients receiving SGLT2i treatment experienced significant increases in peak VO2 (mean difference [MD], 1.61 mL/kg/min; 95% CI, 0.59-2.63 mL/kg/min; P = .002) and 6MWD (MD, 13.09 m; 95% CI, 1.20-24.97 m; P = .03). SGLT2i use was associated with increased KCCQ-TSS (MD, 2.28 points; 95% CI, 1.74-2.81 points; P < .001), KCCQ-CSS (MD, 2.14 points; 95% CI, 1.53-2.74 points; P < .001), and KCCQ-OSS (MD, 1.90 points; 95% CI, 1.41-2.39 points; P < .001) scores. Subgroup analysis and meta-regression demonstrated almost all improvements were consistent across ejection fraction, sex, and the presence of diabetes. Conclusions and Relevance: These findings suggest that in addition to known clinical associations with mortality and hospitalization outcomes, SGLT2i use is associated with improvement in outcomes of interest to patients' everyday lives as measured by objective assessments of maximal exercise capacity and validated QOL questionnaires, regardless of sex or ejection fraction.

Development and Temporal Validation of a Machine Learning Model to Predict Clinical Deterioration.

OBJECTIVES: Early warning scores detecting clinical deterioration in pediatric inpatients have wide-ranging performance and use a limited number of clinical features. This study developed a machine learning model leveraging multiple static and dynamic clinical features from the electronic health record to predict the composite outcome of unplanned transfer to the ICU within 24 hours and inpatient mortality within 48 hours in hospitalized children. METHODS: Using a retrospective development cohort of 17 630 encounters across 10 388 patients, 2 machine learning models (light gradient boosting machine [LGBM] and random forest) were trained on 542 features and compared with our institutional Pediatric Early Warning Score (I-PEWS). RESULTS: The LGBM model significantly outperformed I-PEWS based on receiver operating characteristic curve (AUROC) for the composite outcome of ICU transfer or mortality for both internal validation and temporal validation cohorts (AUROC 0.785 95% confidence interval [0.780-0.791] vs 0.708 [0.701-0.715] for temporal validation) as well as lead-time before deterioration events (median 11 hours vs 3 hours; P = .004). However, LGBM performance as evaluated by precision recall curve was lesser in the temporal validation cohort with associated decreased positive predictive value (6% vs 29%) and increased number needed to evaluate (17 vs 3) compared with I-PEWS. CONCLUSIONS: Our electronic health record based machine learning model demonstrated improved AUROC and lead-time in predicting clinical deterioration in pediatric inpatients 24 to 48 hours in advance compared with I-PEWS. Further work is needed to optimize model positive predictive value to allow for integration into clinical practice.

The algorithm journey map: a tangible approach to implementing AI solutions in healthcare.

When integrating AI tools in healthcare settings, complex interactions between technologies and primary users are not always fully understood or visible. This deficient and ambiguous understanding hampers attempts by healthcare organizations to adopt AI/ML, and it also creates new challenges for researchers to identify opportunities for simplifying adoption and developing best practices for the use of AI-based solutions. Our study fills this gap by documenting the process of designing, building, and maintaining an AI solution called SepsisWatch at Duke University Health System. We conducted 20 interviews with the team of engineers and scientists that led the multi-year effort to build the tool, integrate it into practice, and maintain the solution. This "Algorithm Journey Map" enumerates all social and technical activities throughout the AI solution's procurement, development, integration, and full lifecycle management. In addition to mapping the "who?" and "what?" of the adoption of the AI tool, we also show several 'lessons learned' throughout the algorithm journey maps including modeling assumptions, stakeholder inclusion, and organizational structure. In doing so, we identify generalizable insights about how to recognize and navigate barriers to AI/ML adoption in healthcare settings. We expect that this effort will further the development of best practices for operationalizing and sustaining ethical principles-in algorithmic systems.

Machine Learning for Targeted Advance Care Planning in Cancer Patients: A Quality Improvement Study.

CONTEXT: Prognostication challenges contribute to delays in advance care planning (ACP) for patients with cancer near the end of life (EOL). OBJECTIVES: Examine a quality improvement mortality prediction algorithm intervention's impact on ACP documentation and EOL care. METHODS: We implemented a validated mortality risk prediction machine learning model for solid malignancy patients admitted from the emergency department (ED) to a dedicated solid malignancy unit at Duke University Hospital. Clinicians received an email when a patient was identified as high-risk. We compared ACP documentation and EOL care outcomes before and after the notification intervention. We excluded patients with intensive care unit (ICU) admission in the first 24 hours. Comparisons involved chi-square/Fisher's exact tests and Wilcoxon rank sum tests; comparisons stratified by physician specialty employ Cochran-Mantel-Haenszel tests. RESULTS: Preintervention and postintervention cohorts comprised 88 and 77 patients, respectively. Most were White, non-Hispanic/Latino, and married. ACP conversations were documented for 2.3% of hospitalizations preintervention vs. 80.5% postintervention (P<0.001), and if the attending physician notified was a palliative care specialist (4.1% vs. 84.6%) or oncologist (0% vs. 76.3%) (P<0.001). There were no differences between groups in length of stay (LOS), hospice referral, code status change, ICU admissions or LOS, 30-day readmissions, 30-day ED visits, and inpatient and 30-day deaths. CONCLUSION: Identifying patients with cancer and high mortality risk via machine learning elicited a substantial increase in documented ACP conversations but did not impact EOL care. Our intervention showed promise in changing clinician behavior. Further integration of this model in clinical practice is ongoing.

Development and preliminary testing of Health Equity Across the AI Lifecycle (HEAAL): A framework for healthcare delivery organizations to mitigate the risk of AI solutions worsening health inequities.

The use of data-driven technologies such as Artificial Intelligence (AI) and Machine Learning (ML) is growing in healthcare. However, the proliferation of healthcare AI tools has outpaced regulatory frameworks, accountability measures, and governance standards to ensure safe, effective, and equitable use. To address these gaps and tackle a common challenge faced by healthcare delivery organizations, a case-based workshop was organized, and a framework was developed to evaluate the potential impact of implementing an AI solution on health equity. The Health Equity Across the AI Lifecycle (HEAAL) is co-designed with extensive engagement of clinical, operational, technical, and regulatory leaders across healthcare delivery organizations and ecosystem partners in the US. It assesses 5 equity assessment domains-accountability, fairness, fitness for purpose, reliability and validity, and transparency-across the span of eight key decision points in the AI adoption lifecycle. It is a process-oriented framework containing 37 step-by-step procedures for evaluating an existing AI solution and 34 procedures for evaluating a new AI solution in total. Within each procedure, it identifies relevant key stakeholders and data sources used to conduct the procedure. HEAAL guides how healthcare delivery organizations may mitigate the potential risk of AI solutions worsening health inequities. It also informs how much resources and support are required to assess the potential impact of AI solutions on health inequities.

Translating ethical and quality principles for the effective, safe and fair development, deployment and use of artificial intelligence technologies in healthcare.

OBJECTIVE: The complexity and rapid pace of development of algorithmic technologies pose challenges for their regulation and oversight in healthcare settings. We sought to improve our institution's approach to evaluation and governance of algorithmic technologies used in clinical care and operations by creating an Implementation Guide that standardizes evaluation criteria so that local oversight is performed in an objective fashion. MATERIALS AND METHODS: Building on a framework that applies key ethical and quality principles (clinical value and safety, fairness and equity, usability and adoption, transparency and accountability, and regulatory compliance), we created concrete guidelines for evaluating algorithmic technologies at our institution. RESULTS: An Implementation Guide articulates evaluation criteria used during review of algorithmic technologies and details what evidence supports the implementation of ethical and quality principles for trustworthy health AI. Application of the processes described in the Implementation Guide can lead to algorithms that are safer as well as more effective, fair, and equitable upon implementation, as illustrated through 4 examples of technologies at different phases of the algorithmic lifecycle that underwent evaluation at our academic medical center. DISCUSSION: By providing clear descriptions/definitions of evaluation criteria and embedding them within standardized processes, we streamlined oversight processes and educated communities using and developing algorithmic technologies within our institution. CONCLUSIONS: We developed a scalable, adaptable framework for translating principles into evaluation criteria and specific requirements that support trustworthy implementation of algorithmic technologies in patient care and healthcare operations.