Machine Learning Model for Readmission Prediction of Patients With Heart Failure Based on Electronic Health Records: Protocol for a Quasi-Experimental Study for Impact Assessment.

BACKGROUND: Care for patients with heart failure (HF) causes a substantial load on health care systems where a prominent challenge is the elevated rate of readmissions within 30 days following initial discharge. Clinical professionals face high levels of uncertainty and subjectivity in the decision-making process on the optimal timing of discharge. Unwanted hospital stays generate costs and cause stress to patients and potentially have an impact on care outcomes. Recent studies have aimed to mitigate the uncertainty by developing and testing risk assessment tools and predictive models to identify patients at risk of readmission, often using novel methods such as machine learning (ML). OBJECTIVE: This study aims to investigate how a developed clinical decision support (CDS) tool alters the decision-making processes of health care professionals in the specific context of discharging patients with HF, and if so, in which ways. Additionally, the aim is to capture the experiences of health care practitioners as they engage with the system's outputs to analyze usability aspects and obtain insights related to future implementation. METHODS: A quasi-experimental design with randomized crossover assessment will be conducted with health care professionals on HF patients' scenarios in a region located in the South of Sweden. In total, 12 physicians and nurses will be randomized into control and test groups. The groups shall be provided with 20 scenarios of purposefully sampled patients. The clinicians will be asked to take decisions on the next action regarding a patient. The test group will be provided with the 10 scenarios containing patient data from electronic health records and an outcome from an ML-based CDS model on the risk level for readmission of the same patients. The control group will have 10 other scenarios without the CDS model output and containing only the patients' data from electronic medical records. The groups will switch roles for the next 10 scenarios. This study will collect data through interviews and observations. The key outcome measures are decision consistency, decision quality, work efficiency, perceived benefits of using the CDS model, reliability, validity, and confidence in the CDS model outcome, integrability in the routine workflow, ease of use, and intention to use. This study will be carried out in collaboration with Cambio Healthcare Systems. RESULTS: The project is part of the Center for Applied Intelligent Systems Research Health research profile, funded by the Knowledge Foundation (2021-2028). Ethical approval for this study was granted by the Swedish ethical review authority (2022-07287-02). The recruitment process of the clinicians and the patient scenario selection will start in September 2023 and last till March 2024. CONCLUSIONS: This study protocol will contribute to the development of future formative evaluation studies to test ML models with clinical professionals. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): PRR1-10.2196/52744.

The ß-Blocker Carvedilol and Related Aryloxypropanolamines Promote ERK1/2 Phosphorylation in HEK293 Cells with K A Values Distinct From Their Equilibrium Dissociation Constants as ß 2-Adrenoceptor Antagonists: Evidence for Functional Affinity.

The determination of affinity by using functional assays is important in drug discovery because it provides a more relevant estimate of the strength of interaction of a ligand to its cognate receptor than radioligand binding. However, empirical evidence for so-called, "functional affinity" is limited. Herein, we determined whether the affinity of carvedilol, a ß-adrenoceptor antagonist used to treat heart failure that also promotes extracellular signal-regulated kinases 1 and 2 (ERK1/2) phosphorylation, differed between these two pharmacological activities. Four structurally related ß-adrenoceptor antagonists (alprenolol, carazolol, pindolol, propranolol) that also activated ERK1/2 were included as comparators to enhance our understanding of how these drugs work in the clinical setting. In HEK293 cells stably expressing the human ß 2-adrenoceptor carvedilol and related aryloxypropanolamines were partial agonists of ERK1/2 phosphorylation with potencies ([A]50s) that were lower than their equilibrium dissociation constants (K Bs) as ß 2-adrenoceptor antagonists. As the [A]50 of a partial agonist is a good approximation of its K B, then these data indicated that the affinities of carvedilol and related ligands for these two activities were distinct. Moreover, there was a significant negative rank order correlation between the [A]50 of each ligand to activate ERK1/2 and their intrinsic activities (i.e., as intrinsic activity for ERK1/2 phosphorylation increased, so did affinity). Genome editing revealed that the transducer that coupled the ß 2-adrenoceptor to ERK1/2 phosphorylation in response to carvedilol and other ß 2-adrenoceptor antagonists was Gαs. Collectively, these data support the concept of "functional affinity" and indicate that the ability of the ß 2-adrenoceptor to recruit Gαs may influence the affinity of the activating ligand. SIGNIFICANCE STATEMENT: In HEK293 cells overexpressing the human ß2-adrenoceptor carvedilol and four related aryloxypropanolamines behaved as ß2-adrenoceptor antagonists and partial agonists of ERK1/2 phosphorylation with rank orders of affinity that were distinct. These data imply that carvedilol and other ß-blockers can stabilize the ß2-adrenoceptor in different affinity conformations that are revealed when functionally distinct responses are measured. This is the basis for the pharmacological concept of "functional affinity."

The Price of Explainability in Machine Learning Models for 100-Day Readmission Prediction in Heart Failure: Retrospective, Comparative, Machine Learning Study.

BACKGROUND: Sensitive and interpretable machine learning (ML) models can provide valuable assistance to clinicians in managing patients with heart failure (HF) at discharge by identifying individual factors associated with a high risk of readmission. In this cohort study, we delve into the factors driving the potential utility of classification models as decision support tools for predicting readmissions in patients with HF. OBJECTIVE: The primary objective of this study is to assess the trade-off between using deep learning (DL) and traditional ML models to identify the risk of 100-day readmissions in patients with HF. Additionally, the study aims to provide explanations for the model predictions by highlighting important features both on a global scale across the patient cohort and on a local level for individual patients. METHODS: The retrospective data for this study were obtained from the Regional Health Care Information Platform in Region Halland, Sweden. The study cohort consisted of patients diagnosed with HF who were over 40 years old and had been hospitalized at least once between 2017 and 2019. Data analysis encompassed the period from January 1, 2017, to December 31, 2019. Two ML models were developed and validated to predict 100-day readmissions, with a focus on the explainability of the model's decisions. These models were built based on decision trees and recurrent neural architecture. Model explainability was obtained using an ML explainer. The predictive performance of these models was compared against 2 risk assessment tools using multiple performance metrics. RESULTS: The retrospective data set included a total of 15,612 admissions, and within these admissions, readmission occurred in 5597 cases, representing a readmission rate of 35.85%. It is noteworthy that a traditional and explainable model, informed by clinical knowledge, exhibited performance comparable to the DL model and surpassed conventional scoring methods in predicting readmission among patients with HF. The evaluation of predictive model performance was based on commonly used metrics, with an area under the precision-recall curve of 66% for the deep model and 68% for the traditional model on the holdout data set. Importantly, the explanations provided by the traditional model offer actionable insights that have the potential to enhance care planning. CONCLUSIONS: This study found that a widely used deep prediction model did not outperform an explainable ML model when predicting readmissions among patients with HF. The results suggest that model transparency does not necessarily compromise performance, which could facilitate the clinical adoption of such models.

Interdisciplinary Human-Centered AI for Hospital Readmission Prediction of Heart Failure Patients.

The evolution of clinical decision support (CDS) tools has been improved by usage of new technologies, yet there is an increased need to develop user-friendly, evidence-based, and expert-curated CDS solutions. In this paper, we show with a use-case how interdisciplinary expertise can be combined to develop CDS tool for hospital readmission prediction of heart failure patients. We also discuss how to make the tool integrated in clinical workflow by understanding end-user needs and have clinicians-in-the-loop during the different development stages.

Temporal Context Matters: An Explainable Model for Medical Resource Utilization in Chronic Kidney Disease.

The prediction of medical resource utilization is beneficial for effective healthcare resource planning and allocation. Previous work in resource utilization prediction can be categorized into two main classes, count-based and trajectory-based. Both of these classes have some challenges, in this work we propose a hybrid approach to overcome these challenges. Our initial results promote the value of temporal context in resource utilization prediction and highlight the importance of model explainability in understanding the main important variables.

Morphological and molecular identification of third-stage larvae of Anisakis typica (Nematoda: Anisakidae) from Red Sea coral trout, Plectropomus areolatus.

Anisakidosis is a foodborne zoonotic infection induced by members of the family Anisakidae via the consumption of raw or undercooked fish such as sushi and sashimi. Identifying anisakid larval species is critical for the epidemiology and diagnosis of diseases caused by them. This study aimed at identifying Anisakis larvae collected from marine fish in Egyptian waters based on morphological characteristics and molecular analysis. Thirty marine fish coral trout, Plectropomus areolatus, were collected from Hurghada, Red Sea, Egypt, to investigate larval nematodes of the genus Anisakis. The larvae were detected encapsulated in the peritoneal cavity and muscle of the fish host. This examination revealed that anisakid larvae naturally infected 19 fish specimens with a prevalence of 63.33% and a mean intensity of 4.1 ± 0.40. Most of them (68 larvae: 71.57%) were found in the musculature. Morphological and morphometric analyses using light and scanning electron microscopy revealed a head region with a prominent boring tooth, inconspicuous lips, and a characteristic protruded cylindrical mucron. All larvae in this study possessed the same morphology as Anisakis Larval type I. Molecular analysis based on ITS region using maximum likelihood and Bayesian phylogenetic methods confirmed them as Anisakis typica. This is the first study to identify A. typica larvae from the commercial fish coral trout P. areolatus in Egyptian waters using morphological and molecular methods.

α and ß catalytic subunits of cAMP-dependent protein kinase regulate formoterol-induced inflammatory gene expression changes in human bronchial epithelial cells.

BACKGROUND AND PURPOSE: It has been proposed that genomic mechanisms contribute to adverse effects often experienced by asthmatic subjects who take regular, inhaled ß2 -adrenoceptor agonists as a monotherapy. Moreover, data from preclinical models of asthma suggest that these gene expression changes are mediated by ß-arrestin-2 rather than PKA. Herein, we tested this hypothesis by comparing the genomic effects of formoterol, a ß2 -adrenoceptor agonist, with forskolin in human primary bronchial epithelial cells (HBEC). EXPERIMENTAL APPROACH: Gene expression changes were determined by RNA-sequencing. Gene silencing and genome editing were employed to explore the roles of ß-arrestin-2 and PKA. KEY RESULTS: The formoterol-regulated transcriptome in HBEC treated concurrently with TNFα was defined by 1480 unique gene expression changes. TNFα-induced transcripts modulated by formoterol were annotated with enriched gene ontology terms related to inflammation and proliferation, notably "GO:0070374~positive regulation of ERK1 and ERK2 cascade," which is an apparent ß-arrestin-2 target. However, expression of the formoterol- and forskolin-regulated transcriptomes were highly rank-order correlated and the effects of formoterol on TNFα-induced inflammatory genes were abolished by an inhibitor of PKA. Furthermore, formoterol-induced gene expression changes in BEAS-2B bronchial epithelial cell clones deficient in ß-arrestin-2 were comparable with those expressed by their parental counterparts. Contrariwise, gene expression was partially inhibited in clones lacking the α-catalytic subunit (Cα) of PKA and abolished following the additional knockdown of the ß-catalytic subunit (Cß) paralogue. CONCLUSIONS: The effects of formoterol on inflammatory gene expression in airway epithelia are mediated by PKA and involve the cooperation of PKA-Cα and PKA-Cß.

ß 2-Adrenoceptor Agonists Promote Extracellular Signal-Regulated Kinase 1/2 Dephosphorylation in Human Airway Epithelial Cells by Canonical, cAMP-Driven Signaling Independently of ß-Arrestin 2.

Chronic use of ß 2-adrenoceptor agonists as a monotherapy in asthma is associated with a loss of disease control and an increased risk of mortality. Herein, we tested the hypothesis that ß 2-adrenoceptor agonists, including formoterol, promote biased, ß-arrestin (Arr) 2-dependent activation of the mitogen-activated protein kinases, ERK1/2, in human airway epithelial cells and, thereby, effect changes in gene expression that could contribute to their adverse clinical outcomes. Three airway epithelial cell models were used: the BEAS-2B cell line, human primary bronchial epithelial cells (HBEC) grown in submersion culture, and HBEC that were highly differentiated at an air-liquid interface. Unexpectedly, treatment of all epithelial cell models with formoterol decreased basal ERK1/2 phosphorylation. This was mediated by cAMP-dependent protein kinase and involved the inactivation of C-rapidly-activated fibrosarcoma, which attenuated downstream ERK1/2 activity, and the induction of dual-specificity phosphatase 1. Formoterol also inhibited the basal expression of early growth response-1, an ERK1/2-regulated gene that controls cell growth and repair in the airways. Neither carvedilol, a ß 2-adrenoceptor agonist biased toward ßArr2, nor formoterol promoted ERK1/2 phosphorylation in BEAS-2B cells, although ß 2-adrenoceptor desensitization was compromised in ARRB2-deficient cells. Collectively, these results contest the hypothesis that formoterol activates ERK1/2 in airway epithelia by nucleating a ßArr2 signaling complex; instead, they indicate that ß 2-adrenoceptor agonists inhibit constitutive ERK1/2 activity in a cAMP-dependent manner. These findings are the antithesis of results obtained using acutely challenged native and engineered HEK293 cells, which have been used extensively to study mechanisms of ERK1/2 activation, and highlight the cell type dependence of ß 2-adrenoceptor-mediated signaling. SIGNIFICANCE STATEMENT: It has been proposed that the adverse effects of ß 2-adrenoceptor agonist monotherapy in asthma are mediated by genomic mechanisms that occur principally in airway epithelial cells and are the result of ß-arrestin 2-dependent activation of ERK1/2. This study shows that ß 2-adrenoceptor agonists, paradoxically, reduced ERK1/2 phosphorylation in airway epithelia by disrupting upstream rat sarcoma-C-rapidly accelerated fibrosarcoma complex formation and inducing dual-specificity phosphatase 1. Moreover, these effects were cAMP-dependent protein kinase-dependent, suggesting that ß 2-adrenoceptor agonists were not biased toward ß-arrestin 2 and acted via canonical, cAMP-dependent signaling.

Prostanoid Receptors of the EP4-Subtype Mediate Gene Expression Changes in Human Airway Epithelial Cells with Potential Anti-Inflammatory Activity.

There is a clear, unmet clinical need to identify new drugs to treat individuals with asthma, chronic obstructive pulmonary disease (COPD), and idiopathic pulmonary fibrosis (IPF) in whom current medications are either inactive or suboptimal. In preclinical models, EP4-receptor agonists display efficacy, but their mechanism of action is unclear. In this study, using human bronchial epithelial cells as a therapeutically relevant drug target, we hypothesized that changes in gene expression may play an important role. Several prostanoid receptor mRNAs were detected in BEAS-2B cells, human primary bronchial epithelial cells (HBECs) grown in submersion culture and HBECs grown at an air-liquid interface with PTGER4 predominating. By using the activation of a cAMP response element reporter in BEAS-2B cells as a surrogate of gene expression, Schild analysis determined that PTGER4 mRNAs encoded functional EP4-receptors. Moreover, inhibitors of phosphodiesterase 4 (roflumilast N-oxide [RNO]) and cAMP-dependent protein kinase augmented and attenuated, respectively, reporter activation induced by 2-[3-[(1R,2S,3R)-3-hydroxy-2-[(E,3S)-3-hydroxy-5-[2-(methoxymethyl)phenyl]pent-1-enyl]-5-oxo-cyclopentyl]sulphanylpropylsulphanyl] acetic acid (ONO-AE1-329), a selective EP4-receptor agonist. ONO-AE1-329 also enhanced dexamethasone-induced activation of a glucocorticoid response element reporter in BEAS-2B cells, which was similarly potentiated by RNO. In each airway epithelial cell variant, numerous genes that may impart therapeutic benefit in asthma, COPD, and/or IPF were differentially expressed by ONO-AE1-329, and those changes were often augmented by RNO and/or dexamethasone. We submit that an EP4-receptor agonist, either alone or as a combination therapy, may be beneficial in individuals with chronic lung diseases in whom current treatment options are inadequate. SIGNIFICANCE STATEMENT: Using human bronchial epithelial cells as a therapeutically relevant drug target, we report that EP4-receptor activation promoted gene expression changes that could provide therapeutic benefit in individuals with asthma, COPD, and IPF in whom current treatment options are ineffective or suboptimal.

Impact of Phosphodiesterase 4 Inhibition on the Operational Efficacy, Response Maxima, and Kinetics of Indacaterol-Induced Gene Expression Changes in BEAS-2B Airway Epithelial Cells: A Global Transcriptomic Analysis.

The effects of phosphodiesterase (PDE) 4 inhibitors on gene expression changes in BEAS-2B human airway epithelial cells are reported and discussed in relation to the mechanism(s) of action of roflumilast in chronic obstructive pulmonary disease (COPD). Microarray-based gene expression profiling failed to identify mRNA transcripts that were differentially regulated by the PDE4 inhibitor 6-[3-(dimethylcarbamoyl)benzenesulphonyl]-4-[(3-methoxyphenyl)amino]-8-methylquinoline-3-carboxamide (GSK 256066) after 1, 2, 6, or 18 hours of exposure. However, real-time polymerase chain reaction analysis revealed that GSK 256066 was a weak stimulus, and the negative microarray results reflected low statistical power due to small sample sizes. Furthermore, GSK 256066, roflumilast, and its biologically active metabolite roflumilast N-oxide generally potentiated gene expression changes produced by the long-acting ß 2-adrenoceptor agonists (LABAs) salmeterol, indacaterol, and formoterol. Many of these genes encode proteins with antiviral, anti-inflammatory, and antibacterial activities that could contribute to the clinical efficacy of roflumilast in COPD. RNA-sequencing experiments established that the sensitivity of genes to salmeterol varied by ∼7.5-fold. Consequently, the degree to which a PDE4 inhibitor potentiated the effect of a given concentration of LABA was gene-dependent. Operational model fitting of concentration-response curve data from cells subjected to fractional, ß 2-adrenoceptor inactivation determined that PDE4 inhibition increased the potency and doubled the efficacy of LABAs. Thus, adding roflumilast to standard triple therapy, as COPD guidelines recommend, may have clinical relevance, especially in target tissues where LABAs behave as partial agonists. Collectively, these results suggest that the genomic impact of roflumilast, including its ability to augment LABA-induced gene expression changes, may contribute to its therapeutic activity in COPD.

Analysis of the Indacaterol-Regulated Transcriptome in Human Airway Epithelial Cells Implicates Gene Expression Changes in the Adverse and Therapeutic Effects of ß2-Adrenoceptor Agonists.

The contribution of gene expression changes to the adverse and therapeutic effects of ß2-adrenoceptor agonists in asthma was investigated using human airway epithelial cells as a therapeutically relevant target. Operational model-fitting established that the long-acting ß2-adrenoceptor agonists (LABA) indacaterol, salmeterol, formoterol, and picumeterol were full agonists on BEAS-2B cells transfected with a cAMP-response element reporter but differed in efficacy (indacaterol ≥ formoterol > salmeterol ≥ picumeterol). The transcriptomic signature of indacaterol in BEAS-2B cells identified 180, 368, 252, and 10 genes that were differentially expressed (>1.5- to <0.67-fold) after 1-, 2-, 6-, and 18-hour of exposure, respectively. Many upregulated genes (e.g., AREG, BDNF, CCL20, CXCL2, EDN1, IL6, IL15, IL20) encode proteins with proinflammatory activity and are annotated by several, enriched gene ontology (GO) terms, including cellular response to interleukin-1, cytokine activity, and positive regulation of neutrophil chemotaxis The general enriched GO term extracellular space was also associated with indacaterol-induced genes, and many of those, including CRISPLD2, DMBT1, GAS1, and SOCS3, have putative anti-inflammatory, antibacterial, and/or antiviral activity. Numerous indacaterol-regulated genes were also induced or repressed in BEAS-2B cells and human primary bronchial epithelial cells by the low efficacy LABA salmeterol, indicating that this genomic effect was neither unique to indacaterol nor restricted to the BEAS-2B airway epithelial cell line. Collectively, these data suggest that the consequences of inhaling a ß2-adrenoceptor agonist may be complex and involve widespread changes in gene expression. We propose that this genomic effect represents a generally unappreciated mechanism that may contribute to the adverse and therapeutic actions of ß2-adrenoceptor agonists in asthma.

GS-5759, a Bifunctional ß2-Adrenoceptor Agonist and Phosphodiesterase 4 Inhibitor for Chronic Obstructive Pulmonary Disease with a Unique Mode of Action: Effects on Gene Expression in Human Airway Epithelial Cells.

(R)-6-[(3-{[4-(5-{[2-hydroxy-2-(8-hydroxy-2-oxo-1,2-dihydroquinolin-5-yl)ethyl]amino}pent-1-yn-1-yl)phenyl] carbamoyl}phenyl)sulphonyl]-4-[(3-methoxyphenyl)amino]-8-methylquinoline-3-carboxamide trifluoroacetic acid (GS-5759) is a bifunctional ligand composed of a quinolinone-containing pharmacophore [ß2-adrenoceptor agonist orthostere (ß2A)] found in several ß2-adrenoceptor agonists, including indacaterol, linked covalently to a phosphodiesterase 4 (PDE4) inhibitor related to 6-[3-(dimethylcarbamoyl)benzenesulphonyl]-4-[(3-methoxyphenyl)amino]-8-methylquinoline-3-carboxamide (GSK 256066) by a pent-1-yn-1-ylbenzene spacer. GS-5759 had a similar affinity for PDE4B1 and the native ß2-adrenoceptor expressed on BEAS-2B human airway epithelial cells. However, compared with the monofunctional parent compound, ß2A, the KA of GS-5759 for the ß2-adrenoceptor was 35-fold lower. Schild analysis determined that the affinities of the ß-adrenoceptor antagonists, (2R,3R)-1-[(2,3-dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl) amino]-2-butanol (ICI 118551) and propranolol, were agonist-dependent, being significantly lower for GS-5759 than ß2A. Collectively, these data can be explained by "forced proximity," bivalent binding where the pharmacophore in GS-5759 responsible for PDE4 inhibition also interacts with a nonallosteric domain within the ß2-adrenoceptor that enhances the affinity of ß2A for the orthosteric site. Microarray analyses revealed that, after 2-hour exposure, GS-5759 increased the expression of >3500 genes in BEAS-2B cells that were highly rank-order correlated with gene expression changes produced by indacaterol and GSK 256066 in combination (Ind/GSK). Moreover, the line of regression began close to the origin with a slope of 0.88, indicating that the magnitude of most gene expression changes produced by Ind/GSK was quantitatively replicated by GS-5759. Thus, GS-5759 is a novel compound exhibiting dual ß2-adrenoceptor agonism and PDE4 inhibition with potential to interact on target tissues in a synergistic manner. Such polypharmacological behavior may be particularly effective in chronic obstructive pulmonary disease and other complex disorders where multiple processes interact to promote disease pathogenesis and progression.

Targeting different angiogenic pathways with combination of curcumin, leflunomide and perindopril inhibits diethylnitrosamine-induced hepatocellular carcinoma in mice.

No effective chemopreventive agent has been approved against hepatocellular carcinoma (HCC) to date. Since HCC is one of the hypervascular solid tumors, blocking angiogenesis represents an intriguing approach to HCC chemoprevention. The aim of the current study was to examine the combined effect of the anti-angiogenic agents: leflunomide; a disease modifying antirheumatic drug, perindopril; an angiotensin converting enzyme inhibitor (ACEI) and curcumin; the active principle of turmeric, on diethylnitrosamine (DEN)-induced HCC in mice. Eight weeks following DEN administration, there was a significant rise in immunohistochemical staining of CD31-positive endothelial cells and consequently hepatic microvessel density (MVD) as compared to normal liver. DEN treatment was associated with elevation in hepatic vascular endothelial growth factor (VEGF) level as compared to normal controls (P<0.05, 3842±72pg/ml and 2520.8±97pg/ml, respectively). Similarly, increased hepatic expression of hypoxia inducible growth factor-1α (HIF-1α) was observed in 100% of the DEN-treated animals compared to 0% in their normal counterparts. Treatment with leflunomide, perindopril or curcumin alone abrogated the DEN-induced increased MVD as well as the elevated expression of VEGF, while only curcumin inhibited HIF-1α hepatic expression. Combination of these agents showed further inhibitory action on neovascularization and synergistic attenuation of hepatic VEGF (1954.27±115pg/ml) when compared to each single agent. Histopathological examination revealed a more beneficial chemopreventive activity in the combination group compared to each monotherapy. In conclusion, the combination treatment of leflunomide, perindopril and curcumin targeting different angiogenic pathways, resulted in synergistic inhibition of angiogenesis and consequently more effective chemoprevention of HCC.