Bayesian Networks in the Management of Hospital Admissions: A Comparison between Explainable AI and Black Box AI during the Pandemic.

Artificial Intelligence (AI) and Machine Learning (ML) approaches that could learn from large data sources have been identified as useful tools to support clinicians in their decisional process; AI and ML implementations have had a rapid acceleration during the recent COVID-19 pandemic. However, many ML classifiers are "black box" to the final user, since their underlying reasoning process is often obscure. Additionally, the performance of such models suffers from poor generalization ability in the presence of dataset shifts. Here, we present a comparison between an explainable-by-design ("white box") model (Bayesian Network (BN)) versus a black box model (Random Forest), both studied with the aim of supporting clinicians of Policlinico San Matteo University Hospital in Pavia (Italy) during the triage of COVID-19 patients. Our aim is to evaluate whether the BN predictive performances are comparable with those of a widely used but less explainable ML model such as Random Forest and to test the generalization ability of the ML models across different waves of the pandemic.

Video clips for patient comprehension of atrial fibrillation and deep vein thrombosis in emergency care. A randomised clinical trial.

Integrating video clips in the discharge process may enhance patients' understanding and awareness of their condition. To determine the effect of video clip-integrated discharge discussion on patient comprehension of atrial fibrillation (AF) and deep vein thrombosis (DVT), and their main complications (stroke and pulmonary embolism), we designed a multicentre, pragmatic, parallel groups, randomised clinical trial, that was conducted at two Emergency Units in Italy. A convenience sample of 144 adult patients (or their caregivers) discharged home with either AF or DVT were randomised to receive standard verbal instructions (control) or video clip-integrated doctor-patient discharge discussion. Participants were guided by the discharging physician through the clip. Mean score for primary outcome (knowledge of the diagnosis and its potential complication) (range 0-18) was 5.87 (95% CI, 5.02-6.72] in the control group and 8.28 (95% CI, 7.27-9.31) in the intervention group (mean difference, -2.41; 95% CI, -3.73 to -1.09; p < 0.001). Among secondary outcomes, mean score for knowledge of the prescribed therapy (range 0-6) was 2.98 (95% CI, 2.57-3.39) in the control group and 3.20 (95% CI, 2.73-3.67) in the study group (mean difference, -0.22; 95% CI, -0.84 to 0.39). Mean score for satisfaction (range 0-12) was 7.34 (95% CI, 6.45-8.23) in the control arm and 7.97 (95% CI, 7.15-8.78) in the intervention arm (mean difference, -0.625; 95% CI -1.82 to 0.57). Initiation rate of newly prescribed anticoagulants was 80% (36/45) in the control group and 90.2% (46/51) in the intervention group. Among 109 patients reached at a median follow up of 21 (IQR 16-28) months, 5.55% (3/54) in the control arm and 1.82% (1/55) in the intervention arm had developed stroke or pulmonary embolism. In this trial, video clip-integrated doctor-patient discharge discussion, improved participants comprehension of AF and DVT and their main complications. Physicians should consider integrating these inexpensive tools during the discharge process of patients with AF or DVT.Trial Registration: ClinicalTrials.gov Identifier "NCT03734406".

A Longitudinal RCT on the Effectiveness of a Psychological Intervention for Hospital Healthcare Workers During the COVID-19 Pandemic: What We Learned to Date.

The COVID-19 pandemic has led to significant psychological distress among frontline healthcare workers (HCWs), with a particular increase in trauma-related symptoms. This study investigated the longitudinal course of trauma-associated symptoms and behaviors in HCWs and the effectiveness of a brief dialectical behavior therapy (DBT)-informed intervention in mitigating these symptoms over 12 months. The trial included 225 HCWs randomly assigned to one of three groups: no intervention (control), in-person DBT-informed intervention, or online DBT-informed intervention. Over time, a natural decrease in PTSD symptoms was observed in all groups. Contrary to expectations, no difference was found between the control and intervention groups. However, for participants with severe PTSD symptoms, the intervention significantly mitigated their distress. No differences emerged between in-person and online interventions, suggesting equal effectiveness. Females reported higher trauma-related symptoms, while no differences emerged among different professional roles. These findings underscore the importance of targeted interventions for HCWs experiencing severe symptoms and highlight the potential of online modalities. Further research is needed to optimize the deployment of mental health resources within the healthcare setting, particularly during crises.

ECG/echo indexes in the diagnostic approach to amyloid cardiomyopathy: A head-to-head comparison from the AC-TIVE study.

BACKGROUND AND AIMS: The discordance between QRS voltages on electrocardiogram (ECG) and left ventricle (LV) wall thickness (LVWT) on echocardiogram (echo) is a recognized red flag (RF) of amyloid cardiomyopathy (AC) and can be measured by specific indexes. No head-to-head comparison of different ECG/echo indexes among subjects with echocardiographic suspicion of AC has yet been undertaken. The study aimed at evaluating the performance and the incremental diagnostic value of different ECG/echo indexes in this subset of patients. METHODS: Electrocardiograms of subjects with LV hypertrophy, preserved ejection fraction and ≥ 1 echocardiographic RF of AC participating in the AC-TIVE study, an Italian prospective multicenter study, were independently analyzed by two cardiologists. Low QRS voltages and 8 different ECG/echo indexes were evaluated. Cohort specific cut-offs were computed. RESULTS: Among 170 patients, 55 (32 %) were diagnosed with AC. Combination of low QRS voltages with interventricular septum ≥ 1,6 cm was the most specific (specificity 100 %, positive predictive value 100 %) ECG/echo index, while the ratio between the sum of all QRS voltages and LVWT <7,8 was the most sensitive and accurate (sensitivity 94 %, negative predictive value 97 %, accuracy 82 %). When the latter index was added to a model using easily-accessible clinical variables, the diagnostic accuracy for AC greatly increased (AUC from 0,84 to 0,95; p = 0,007). CONCLUSIONS: Among patients with non-dilated hypertrophic ventricles with normal ejection fraction and echocardiographic RF of AC, easily-measurable ECG/echo indexes, mainly when added to few clinical variables, can help the physician orient second level investigations. External validation of the results is warranted.

FXR agonists INT-787 and OCA increase RECK and inhibit liver steatosis and inflammation in diet-induced ob/ob mouse model of NASH.

BACKGROUND AND AIMS: We have previously shown in a model of hepatic ischaemia/reperfusion injury that the farnesoid X receptor (FXR) agonist obeticholic acid (OCA) restores reversion-inducing-cysteine-rich protein with Kazal motifs (RECK), an inverse modulator of metalloproteases (MMPs) and inhibitor of the sheddases ADAM10 and ADAM17 involved in inflammation and fibrogenesis. Here, the effects of FXR agonists OCA and INT-787 on hepatic levels of RECK, MMPs, ADAM10 and ADAM17 were compared in a diet-induced ob/ob mouse model of non-alcoholic steatohepatitis (NASH). METHODS: Lep ob/ob NASH mice fed a high-fat diet (HFD) or control diet (CD) for 9 weeks (wks) were treated with OCA or INT-787 0.05% dosed via HFD admixture (30 mg/kg/day) or HFD for further 12 wks. Serum alanine transaminase (ALT) and inflammatory cytokines, liver RECK, MMP-2 and MMP-9 activity as well as ADAM10, ADAM17, collagen deposition (Sirius red), hepatic stellate cell activation (α-SMA) and pCK+ reactive biliary cells were quantified. RESULTS: Only INT-787 significantly reduced serum ALT, IL-1ß and TGF-ß. A downregulation of RECK expression and protein levels observed in HFD groups (at 9 and 21 wks) was counteracted by both OCA and INT-787. HFD induced a significant increase in liver MMP-2 and MMP-9; OCA administration reduced both MMP-2 and MMP-9 while INT-787 markedly reduced MMP-2 expression. OCA and INT-787 reduced both ADAM10 and ADAM17 expression and number of pCK+ cells. INT-787 was superior to OCA in decreasing collagen deposition and α-SMA levels. CONCLUSION: INT-787 is superior to OCA in controlling specific cell types and clinically relevant anti-inflammatory and antifibrotic molecular mechanisms in NASH.

Performance of an AI algorithm during the different phases of the COVID pandemics: what can we learn from the AI and vice versa.

Background: Artificial intelligence (AI) has proved to be of great value in diagnosing and managing Sars-Cov-2 infection. ALFABETO (ALL-FAster-BEtter-TOgether) is a tool created to support healthcare professionals in the triage, mainly in optimizing hospital admissions. Methods: The AI was trained during the pandemic's "first wave" (February-April 2020). Our aim was to assess the performance during the "third wave" of the pandemics (February-April 2021) and evaluate its evolution. The neural network proposed behavior (hospitalization vs home care) was compared with what was actually done. If there were discrepancies between ALFABETO's predictions and clinicians' decisions, the disease's progression was monitored. Clinical course was defined as "favorable/mild" if patients could be managed at home or in spoke centers and "unfavorable/severe" if patients need to be managed in a hub center. Results: ALFABETO showed accuracy of 76%, AUROC of 83%; specificity was 78% and recall 74%. ALFABETO also showed high precision (88%). 81 hospitalized patients were incorrectly predicted to be in "home care" class. Among those "home-cared" by the AI and "hospitalized" by the clinicians, 3 out of 4 misclassified patients (76.5%) showed a favorable/mild clinical course. ALFABETO's performance matched the reports in literature. Conclusions: The discrepancies mostly occurred when the AI predicted patients could stay at home but clinicians hospitalized them; these cases could be handled in spoke centers rather than hubs, and the discrepancies may aid clinicians in patient selection. The interaction between AI and human experience has the potential to improve both AI performance and our comprehension of pandemic management.

Diagnostic pathways to wild-type transthyretin amyloid cardiomyopathy: a multicentre network study.

AIM: Epidemiology of wild-type transthyretin cardiac amyloidosis (ATTRwt-CA) remains poorly defined. A better characterization of pathways leading to ATTRwt-CA diagnosis is of key importance, and potentially informative of disease course and prognosis. The aim of this study was to describe the characteristics of contemporary pathways leading to ATTRwt-CA diagnosis, and their potential association with survival. METHODS AND RESULTS: This was a retrospective study of patients diagnosed with ATTRwt-CA at 17 Italian referral centres for CA. Patients were categorized into different 'pathways' according to the medical reason that triggered the diagnosis of ATTRwt-CA (hypertrophic cardiomyopathy [HCM] pathway, heart failure [HF] pathway, incidental imaging or incidental clinical pathway). Prognosis was investigated with all-cause mortality as endpoint. Overall, 1281 ATTRwt-CA patients were included in the study. The diagnostic pathway leading to ATTRwt-CA diagnosis was HCM in 7% of patients, HF in 51%, incidental imaging in 23%, incidental clinical in 19%. Patients in the HF pathway, as compared to the others, were older and had a greater prevalence of New York Heart Association (NYHA) class III-IV and chronic kidney disease. Survival was significantly worse in the HF versus other pathways, but similar among the three others. In multivariate model, older age at diagnosis, NYHA class III-IV and some comorbidities but not the HF pathway were independently associated with worse survival. CONCLUSIONS: Half of contemporary ATTRwt-CA diagnoses occur in a HF setting. These patients had worse clinical profile and outcome than those diagnosed either due to suspected HCM or incidentally, although prognosis remained primarily related to age, NYHA functional class and comorbidities rather than the diagnostic pathway itself.

Echocardiographic findings in subjects with an amyloidogenic apolipoprotein A1 pathogenic variant.

BACKGROUND: Very small case series of patients with apolipoprotein A1 (ApoA1) amyloidosis are available. METHODS: We described the clinical and echocardiographic characteristics of individuals with the pathogenic APOA1 variant Leu75Pro (p. Leu99Pro), referred for cardiac screening. RESULTS: We enrolled 189 subjects, 54% men, median age 55 years (interquartile range 42-67), 39% with concomitant renal disease and 31% with liver disease. Median left ventricular ejection fraction was 60% (55-66). Overall, these subjects did not show overt diastolic dysfunction nor left ventricular (LV) hypertrophy. Age correlated with interventricular septal (IVS) thickness (r = 0.484), LV mass index (r = 0.459), E/e' (r = 0.501), and right ventricular free wall thickness (r = 0.594) (all p < 0.001). Some individuals displayed red flags for cardiac amyloidosis (CA), and 14% met non-invasive criteria for CA. Twenty-nine subjects died over 5.8 years (4.1-8.0), with 10 deaths for cardiovascular causes. Individuals meeting echocardiographic criteria for CA had a much higher risk of all-cause death (p = 0.009), cardiovascular death (p = 0.001), cardiovascular death or heart failure (HF) hospitalisation (p < 0.001). Subjects with both renal and liver involvement had a more prominent cardiac involvement, and shortest survival. CONCLUSIONS: Subjects with the APOA1 Leu75Pro variant displayed minor echocardiographic signs of cardiac involvement, but 14% met echocardiographic criteria for CA. Subjects with suspected CA had a worse outcome.

Reduction of BSI associated mortality after a sepsis project implementation in the ER of a tertiary referral hospital.

The emergency room (ER) is the first gateway for patients with sepsis to inpatient units, and identifying best practices and benchmarks to be applied in this setting might crucially result in better patient's outcomes. In this study, we want to evaluate the results in terms of decreased the in-hospital mortality of patients with sepsis of a Sepsis Project developed in the ER. All patients admitted to the ER of our Hospital from the 1st January, 2016 to the 31stJuly 2019 with suspect of sepsis (MEWS score ≥ of 3) and positive blood culture upon ER admission were included in this retrospective observational study. The study comprises of two periods: Period A: From the 1st Jan 2016 to the 31st Dec 2017, before the implementation of the Sepsis project. Period B: From the 1st Jan 2018 to the 31stJul 2019, after the implementation of the Sepsis project. To analyze the difference in mortality between the two periods, a univariate and multivariate logistic regression was used. The risk of in-hospital mortality was expressed as an odds ratio (OR) and a 95% confidence interval (95% CI). Overall, 722 patients admitted in ER had positive BC on admissions, 408 in period A and 314 in period B. In-hospital mortality was 18.9% in period A and 12.7% in period B (p = 0.03). At multivariable analysis, mortality was still reduced in period B compared to period A (OR 0.64, 95% CI 0.41-0.98; p = 0.045). Having an infection due to GP bacteria or polymicrobial was associated with an increased risk of death, as it was having a neoplasm or diabetes. A marked reduction in in-hospital mortality of patients with documented BSI associated with signs or symptoms of sepsis after the implementation of a sepsis project based on the application of sepsis bundles in the ER.

Perceptive SARS-CoV-2 End-To-End Ultrasound Video Classification through X3D and Key-Frames Selection.

The SARS-CoV-2 pandemic challenged health systems worldwide, thus advocating for practical, quick and highly trustworthy diagnostic instruments to help medical personnel. It features a long incubation period and a high contagion rate, causing bilateral multi-focal interstitial pneumonia, generally growing into acute respiratory distress syndrome (ARDS), causing hundreds of thousands of casualties worldwide. Guidelines for first-line diagnosis of pneumonia suggest Chest X-rays (CXR) for patients exhibiting symptoms. Potential alternatives include Computed Tomography (CT) scans and Lung UltraSound (LUS). Deep learning (DL) has been helpful in diagnosis using CT scans, LUS, and CXR, whereby the former commonly yields more precise results. CXR and CT scans present several drawbacks, including high costs. Radiation-free LUS imaging requires high expertise, and physicians thus underutilise it. LUS demonstrated a strong correlation with CT scans and reliability in pneumonia detection, even in the early stages. Here, we present an LUS video-classification approach based on contemporary DL strategies in close collaboration with Fondazione IRCCS Policlinico San Matteo's Emergency Department (ED) of Pavia. This research addressed SARS-CoV-2 patterns detection, ranked according to three severity scales by operating a trustworthy dataset comprising ultrasounds from linear and convex probes in 5400 clips from 450 hospitalised subjects. The main contributions of this study are related to the adoption of a standardised severity ranking scale to evaluate pneumonia. This evaluation relies on video summarisation through key-frame selection algorithms. Then, we designed and developed a video-classification architecture which emerged as the most promising. In contrast, the literature primarily concentrates on frame-pattern recognition. By using advanced techniques such as transfer learning and data augmentation, we were able to achieve an F1-Score of over 89% across all classes.

[Rationale and significance of the Italian Network for Cardiac Amyloidosis]. / Razionale e valore fondamentale della Rete Italiana integrata dell'Amiloidosi Cardiaca.

The perspective on amyloidosis has changed deeply over the last 10 years following major advances in diagnosis and treatment options, especially in cardiac amyloidosis. This intrinsically heterogeneous disease exposes to the risk of fragmentation of knowledge and requires the interaction among experts of different specialties and subspecialties. Suspicion of disease, timely recognition and confirmation of final diagnosis, prognostic stratification, clinical management and therapeutic strategies represent essential steps to be taken. Missing or delaying the diagnosis may have dramatic impact on patient outcome, as in the case of chemotherapy in unrecognized light-chain amyloidosis. Therefore, there is an urgent need for the foundation of an Italian Amyloidosis Network to deal with the challenges of this condition and orient clinical management at national and local levels. The present consensus document aims to provide the rationale and scopes of the Italian Amyloidosis Network, which has been conceived as an organizational framework for professionals managing patients with amyloidosis.

Electrocardiographic features and rhythm disorders in cardiac amyloidosis.

Cardiac amyloidosis (CA) is an infiltrative cardiomyopathy caused by extracellular deposition of amyloid fibrils, mainly derived from transthyretin, either wild-type or hereditary variants, or immunoglobulin light chains misfolding. It is characterized by an increased left ventricular (LV) mass and diastolic dysfunction, which can lead to heart failure with preserved ejection fraction and/or conduction disturbances. The diagnosis is based on invasive pathology demonstration of amyloid deposits, or non-invasive criteria using advanced cardiovascular imaging techniques. Nevertheless, 12-lead electrocardiogram (ECG) remains of crucial importance in the assessment of patients with CA, since they can manifest peculiar features such as low QRS voltages, in discordance with the LV hypertrophy, but also pseudo-infarction patterns, sinus node dysfunction, atrioventricular blocks, premature supraventricular and ventricular beats, which support the presence of a myocardial disease. Great awareness of these common ECG characteristics of CA is needed to increase diagnostic performance and improve patient's outcome. In the present review, we discuss the current role of the ECG in the diagnosis and management of CA, focusing on the most common ECG abnormalities and rhythm disorders.

Hyponatremia as a predictor of outcome and mortality: results from a second-level urban emergency department population.

BACKGROUND: Hyponatremia is the most common electrolyte disorder and it has been associated with increased mortality. AIMS: This study evaluated hyponatremia as a prognostic factor for severity and mortality. METHODS: We compared the prevalence of hyponatremia among patients who died during the year 2017 (from 1 January 2017 to 31 December 2017) with the prevalence of hyponatremia among subgroups of patients, i.e. outpatients, patients hospitalized for more than 2 days and patients admitted in the intensive care unit (ICU). We also described the mortality rate and the prevalence of comorbidities among hyponatremic patients, according to hyponatremia degree (slight, moderate, severe), basal characteristics, comorbidities and their outcome (discharged, hospitalized or died). RESULTS: In our population of a public hospital setting, hyponatremia was present at admission in 17% of deaths, and the comparison between hyponatremic and normonatremic patients in terms of mortality confirms the hypothesis that this disorder is in anyway strictly associated with vulnerability and with a poor prognosis. CONCLUSIONS: We conclude that hyponatremia is a predictive marker for a bad clinical course, therefore patients with this electrolyte disorder should be carefully monitored.

Mid-term follow-up chest CT findings in recovered COVID-19 patients with residual symptoms.

OBJECTIVES: More than a year has passed since the initial outbreak of SARS-CoV-2, which caused many hospitalizations worldwide due to COVID-19 pneumonia and its complications. However, there is still a lack of information detailing short- and long-term outcomes of previously hospitalized patients. The purpose of this study is to analyze the most frequent lung CT findings in recovered COVID-19 patients at mid-term follow-ups. METHODS: A total of 407 consecutive COVID-19 patients who were admitted to the Fondazione IRCCS Policlinico San Matteo, Pavia and discharged between February 27, 2020, and June 26, 2020 were recruited into this study. Out of these patients, a subset of 108 patients who presented with residual asthenia and dyspnea at discharge, altered spirometric data, positive lung ultrasound and positive chest X-ray was subsequently selected, and was scheduled to undergo a mid-term chest CT study, which was evaluated for specific lung alterations and morphological patterns. RESULTS: The most frequently observed lung CT alterations, in order of frequency, were ground-glass opacities (81%), linear opacities (74%), bronchiolectases (64.81%), and reticular opacities (63.88%). The most common morphological pattern was the non-specific interstitial pneumonia pattern (63.88%). Features consistent with pulmonary fibrosis were observed in 32 patients (29.62%). CONCLUSIONS: Our work showed that recovered COVID-19 patients who were hospitalized and who exhibited residual symptoms after discharge had a slow radiological recovery with persistent residual lung alterations. ADVANCES IN KNOWLEDGE: This slow recovery process should be kept in mind when determining the follow-up phases in order to improve the long-term management of patients affected by COVID-19.

Prognosis of Transthyretin Cardiac Amyloidosis Without Heart Failure Symptoms.

Background: Transthyretin amyloid cardiomyopathy (ATTR-CM) is increasingly recognized as a treatable cause of heart failure (HF). Advances in diagnosis and therapy have increased the number of patients diagnosed at early stages, but prognostic data on patients without HF symptoms are lacking. Moreover, it is unknown whether asymptomatic patients benefit from early initiation of transthyretin (TTR) stabilizers. Objectives: The aim of this study was to describe the natural history and prognosis of ATTR-CM in patients without HF symptoms. Methods: Clinical characteristics and outcomes of patients with ATTR-CM without HF symptoms were retrospectively collected at 6 international amyloidosis centers. Results: A total of 118 patients (78.8% men, median age 66 years [IQR: 53.8-75 years], 68 [57.6%] with variant transthyretin amyloidosis, mean left ventricular ejection fraction 60.5% ± 9.9%, mean left ventricular wall thickness 15.4 ± 3.1 mm, and 53 [45%] treated with TTR stabilizers at baseline or during follow-up) were included. During a median follow-up period of 3.7 years (IQR: 1-6 years), 38 patients developed HF symptoms (23 New York Heart Association functional class II and 14 functional class III or IV), 32 died, and 2 required cardiac transplantation. Additionally, 20 patients received pacemakers, 13 developed AF, and 1 had a stroke. Overall survival was 96.5% (95% CI: 91%-99%), 90.4% (95% CI: 82%-95%), and 82% (95% CI: 71%-89%) at 1, 3, and 5 years, respectively. Treatment with TTR stabilizers was associated with improved survival (HR: 0.31; 95% CI: 0.12-0.82; P = 0.019) and remained significant after adjusting for sex, age, ATTR-CM type, and estimated glomerular filtration rate (HR: 0.18; 95% CI: 0.06-0.55; P = 0.002). Conclusions: After a median follow-up period of 3.7 years, 1 in 3 patients with asymptomatic ATTR-CM developed HF symptoms, and nearly as many died or required cardiac transplantation. Treatment with TTR stabilizers was associated with improved prognosis.

MCD Diet Rat Model Induces Alterations in Zinc and Iron during NAFLD Progression from Steatosis to Steatohepatitis.

We evaluate the effects of the methionine-choline-deficient (MCD) diet on serum and hepatic zinc (Zn) and iron (Fe) and their relationships with matrix metalloproteinases (MMPs) and their modulators (TIMPs and RECK) as well as hepatic fatty acids using male Wistar rats fed 2-, 4- and 8-week MCD diets. Serum and hepatic Zn decrease after an 8-week MCD diet. Serum Fe increases after an 8-week MCD diet and the same occurs for hepatic Fe. An increase in hepatic MMP activity, associated with a decrease in RECK and TIMPs, is found in the MCD 8-week group. Liver Fe shows a positive correlation versus MMPs and RECK, and an inverse correlation versus TIMPs. A positive correlation is found comparing liver Zn with stearic, vaccenic and arachidonic acids, and an inverse correlation is found with linolenic and docosatetraenoic acids. An opposite trend is found between liver Fe versus these fatty acids. During NAFLD progression from steatosis to steatohepatitis, MCD rats exhibit an increase in Zn and a decrease in Fe levels both in serum and tissue associated with alterations in hepatic MMPs and their inhibitors, and fatty acids. The correlations detected between Zn and Fe versus extracellular matrix modulators and fatty acids support their potential role as therapeutic targets.