Artificial Intelligence for detecting asynchrony in non-invasive ventilation: a concordance study.

BackgroundThere has been a resurgence of Artificial Intelligence (AI) on a global scale in recent times, resulting in the development of cutting-edge AI solutions within hospitals. However, this has also led to the creation of isolated AI solutions that are not integrated into clinical IT. To tackle this issue, a clinical Artificial Intelligence (AI) platform that handles the entire development cycle of clinical AI models and is integrated into clinical IT is required. This research investigates the integration of a clinical AI platform into the clinical IT infrastructure. This is demonstrated by outlining the stages of the AI model development cycle within the clinical IT infrastructure, illustrating the interaction between different IT system landscapes within the hospital with BPMN diagrams.MethodsInitially, a thorough analysis of the requirements is conducted to refine the necessary aspects of the clinical AI platform with consideration of the individual aspects of clinical IT. Subsequently, processes representing the entire development cycle of an AI model are identified. To facilitate the architecture of the AI platform, BPMN diagrams of all the identified processes are created. Clinical use cases are used to evaluate the processes using the FEDS framework.ResultsOur BPMN process diagrams cover the entire development cycle of a clinical AI model within the clinical IT. The processes involved are Data Selection, Data Annotation, On-site Training and Testing, and Inference, with distinctions between (Semi-Automated) Batch Inference and Real-Time Inference. Three clinical use cases were assessed to evaluate the processes and demonstrate that this approach covers a wide range of clinical AI use cases.ConclusionsThe evaluations were executed successfully, which indicate the comprehensive nature of our approach. The results have shown that different clinical AI use cases are covered by the BPMN diagrams. Our clinical AI platform is ideally suited for the local development of AI models within clinical IT. This approach provides a basis for further developments, e.g., enabling the training and deployment of an AI model across multiple sites or the integration of security- and privacy-related aspects.

Noninvasive ventilation (NIV) is a widely used and well-established treatment modality for respiratory failure. In patients with increased respiratory work of breathing, accessory muscles are commonly activated along with the diaphragm. Whereas diaphragm ultrasound has been utilized to assess outcomes of mechanical ventilation, the data on intercostal muscle ultrasound remain limited. We aimed to investigate the association between intercostal muscle thickening fraction (TF) and NIV failure in critical care patients with hypercapnic respiratory failure. Critical care subjects receiving NIV for hypercapnic respiratory failure were enrolled in the study. The intercostal muscle TF was measured on admission day (day 0) and the following day (day 1). NIV failure was defined as the need for invasive mechanical ventilation or death during NIV therapy. A total of 158 subjects were enrolled, and 30 experienced NIV failure. Age, sex, and body mass index (BMI) were similar in the NIV success and failure groups. Acute Physiology And Chronic Health Evaluation II (APACHE II) and the Sequential Organ Failure Assessment (SOFA) scores were higher in the NIV failure group. In terms of causes of respiratory failure, the COPD exacerbation rate was higher in the NIV success group. TF was higher in the NIV failure group on both day 0 and day 1. The increased TF on the ICU admission day, with a cutoff value of 12%, was associated with NIV failure after adjusting for age, sex, BMI, APACHE II, and SOFA. Persistence of a higher TF value on both day 0 and day 1 was also associated with NIV failure risk. There was a positive relation between intercostal muscle TF measured by ultrasound and NIV failure, even after adjusting for APACHE II and SOFA scores.

IntroductionAcute hypercapnic respiratory failure is one of the most important cause of ICU admission over last few decades.1 The use of Non-invasive ventilation (NIV) in type 2 RF has become one of the most important developments over the past two decades.2 NIV failure has been defined as the need for endotracheal intubation (ETI) or death.3, 4The failure rate ranges from 15 to 24 % depending on various factors.5 Patients who experience NIV failure and then receive delayed intubation have higher mortality and hospital stay. Thus, the early identification of NIV failure in these patients is very important.6 HACOR score which takes into such variables had high sensitivity and specificity for the prediction of early NIV failure.7–9 The variables for the HACOR score are Heart Rate(H), Acidosis(A), Consciousness level(C), Oxygenation(O) and Respiratory rate(R), which are easily obtained at bedside.[6] Thus the aim of the study is to predict early NIV failure using HACOR score.ObjectivesTo study predictive value of HACOR SCORE in success and failure of NIV in Acute hypercapnic respiratory failure.To observe the trends of HACOR score at different intervals(initial,1-2 hours,12 hours,24 hours of NIV) among these patients and compare them among NIV success and failure patients.Materials and MethodsThis was a prospective observational single centre study conducted at Intensive Care Unit of Fortis Escort heart institute Okhla, New Delhi from May 2023 to October 2023 after obtaining approval from Institutional ethics committee.57 patients were randomly selected at our hospital maintain inclusion and exclusion criterias.Inclusion Criteria:Patients with age over 18 yearsPatients admitted to medical ICU with Acute Respiratory Failure with ABG showing pH< 7.35 and PCO2>45Patients with respiration rate >24 or presence of dyspnea at rest assessed using accessory respiratory muscles or paradoxical abdominal breathing.Exclusion Criteria:Patients who are not willing to accept NIVPatients having symptoms suggestive of unstable angina, AMI, congestive heart failure, unstable cardiac arrhythmias, severe hypoxia and hemodynamically unstable patients on very high doses of vasopressorsPatients having any contraindication for NIVResultsDatas are tabulated and analyzed by proper statistical tests. NIV failure rate was 22.8% in our study. In patients HACOR score<5 at 1—2 hrs of NIV, the failure rate is 5.26%, where as in patients with score >5 at same time is 57.89%. At 1—2 h of NIV HACOR>5, Sensitivity and Specificity being 84.6% and 82.8% respectively. In our study from ROC curve, cut off point for NIV failure seems to be HACOR>6 at 1-2 hrs of NIV and AUC, Sensitivity, Specificity, PPV, NPV are 0.937(0.86-1.00),84.6%,93.2%,78.57%,95.35% respectively. The HACOR score of NIV success and failure groups are compared at different intervals(beginning,1-2 hrs,12 hrs,24 hrs) and the result is statistically significant(p=0.001, highly significant).ConclusionHACOR score had high sensitivity and specificity for the prediction of early NIV failure in Type 2 respiratory failure. As the variables are easily obtained bedside, it is convenient to use and asses the efficacy of NIV in these patients.

OBJECTIVE: This study aims to assess the accuracy and clinical utility of artificial intelligence (AI) platforms in responding to questions related to penile fracture, a rare but urgent urological emergency. MATERIAL and METHODS: Twenty-five questions addressing key clinical aspects of penile fracture were submitted to four AI platforms: ChatGPT, Copilot, Gemini, and Perplexity. Two expert urologists evaluated each response across five domains –relevance, clarity, structure, utility, and factual accuracy– using a 5-point Likert scale. Inter-rater reliability was assessed using the intraclass correlation coefficient (ICC), and statistical comparisons were made using one-way ANOVA and Tukey’s post-hoc tests. RESULTS: Copilot and ChatGPT scored highest overall, with mean scores of 4.90 and 4.89 respectively, while Perplexity scored significantly lower (4.68; p &lt;0.001). Copilot also achieved the highest ratings in clarity and factual accuracy. Inter-rater reliability was high, and dimensional analysis confirmed the consistent superiority of Copilot and ChatGPT in clinical relevance and clarity. CONCLUSION: While AI platforms –especially Copilot and ChatGPT– show promise in generating medically relevant content about penile fracture, limitations in factual accuracy and clinical specificity remain. Caution is advised in using these tools in urgent care settings without professional oversight. Keywords: artificial intelligence, clinical accuracy, emergency medicine, large language models, penile fracture, urology

To investigate the role of lung ultrasonography (LU) in predicting noninvasive ventilation (NIV) failure and the relationship between lung ultrasonography scores (LUS) with clinical outcomes in neonatal respiratory failure (NRF). A prospective, cross-sectional study was conducted in newborns with NRF who needed NIV and were evaluated by LU. The first LUS (LUS1) was calculated at 2-6 hours and the second (LUS2) at 12-24 hours of life. The patients were divided into NIV failure and NIV non-failure groups. The relationship between LUS and clinical outcomes was evaluated. Among 157 neonates, the median (interquartile range) of gestational week and birth weight were 37 weeks (34-39), and 2890 grams (2045-3435), respectively. The reasons for NRF were transient tachypnea of the newborn (n=92, 58.6%), congenital pneumonia (n=58, 36.9%), and respiratory distress syndrome (n=7, 4.5%). The rate of NIV failure was 17.8% (n=28). Both LUS1 and LUS2 were significantly higher in neonates with NIV failure compared to neonates with NIV non-failure (P=.001). A cutoff value of LUS1 ≥ 4 predicted NIV failure with 96% sensitivity and 63% specificity. There were positive correlations between LUS and PEEP values, IMV and total MV days, carbon dioxide values, length of hospital stay, and antibiotic days (ρLUS1 , P=.843, <.001; .474, <.001; .444, <.001; .258, .001; .212, .008; .270, <.001, respectively). Lung ultrasound scores were higher in neonates with NIV failure than with NIV non-failure group, and strongly correlated with end-expiratory pressure values. Lung ultrasound scores were found to be related with some of the clinical outcomes of the NRF, and this suggested that LUS could provide information about the prognosis of NRF.

Objectives:With over 2 million cases of acute respiratory failure in the United States per year, noninvasive ventilation has become a leading treatment modality, often supplanting invasive mechanical ventilation as the initial treatment of choice. Most acute respiratory failure patients use a full face (oronasal) mask with noninvasive ventilation, which is known to impair communication, but its popularity and benefit has led many providers to accept the communication impairment. Medical staff periodically remove masks to communicate with patients, but patients are often limited to short utterances and risk lung derecruitment upon removal of positive pressure. These problems can lead to noninvasive ventilation failure, which is often linked to worse outcomes than first initiating invasive mechanical ventilation and can lead to increased hospitalization costs.Data Sources:We searched MEDLINE and Google Scholar for “speech,” “communication,” “impairment,” “failure,” “complications,” “NIPPV,” “NIV,” and “noninvasive ventilation.”Study Selection:We included articles with patients in acute respiratory failure. We excluded articles for patients using noninvasive ventilation therapy for obstructive sleep apnea.Data Synthesis:Communication impairment has been associated with increasing noninvasive ventilation anxiety (odds ratio, 1.25). Of patients using noninvasive ventilation, 48% require early discontinuation, 22% refuse noninvasive ventilation, and 9% are ultimately intubated. Improvements to communication have been shown to reduce fear and anxiety in invasive mechanical ventilation patients. Analogous communication problems exist with effective solutions in other fields, such as fighter pilot masks, that can be easily implemented to enhance noninvasive ventilation patient care, increase adherence to noninvasive ventilation treatment, and improve patient outcomes.Conclusions:Communication impairment is an underappreciated cause of noninvasive ventilation complications and failure and requires further characterization. Analogous solutions—such as throat microphones and mask-based microphones—that can be easily implemented show potential as cost-effective methods to reduce noninvasive ventilation failure.

Artificial intelligence in pediatric otorhinolaryngology: Assessing readability, understandability, and actionability of postoperative instructions.

Noninvasive ventilation (NIV) is increasingly used in patients with acute respiratory distress syndrome (ARDS). The evidence supporting NIV use in patients with ARDS remains relatively sparse. To determine whether, during NIV, the categorization of ARDS severity based on the PaO2/FiO2 Berlin criteria is useful. The LUNG SAFE (Large Observational Study to Understand the Global Impact of Severe Acute Respiratory Failure) study described the management of patients with ARDS. This substudy examines the current practice of NIV use in ARDS, the utility of the PaO2/FiO2 ratio in classifying patients receiving NIV, and the impact of NIV on outcome. Of 2,813 patients with ARDS, 436 (15.5%) were managed with NIV on Days 1 and 2 following fulfillment of diagnostic criteria. Classification of ARDS severity based on PaO2/FiO2 ratio was associated with an increase in intensity of ventilatory support, NIV failure, and intensive care unit (ICU) mortality. NIV failure occurred in 22.2% of mild, 42.3% of moderate, and 47.1% of patients with severe ARDS. Hospital mortality in patients with NIV success and failure was 16.1% and 45.4%, respectively. NIV use was independently associated with increased ICU (hazard ratio, 1.446 [95% confidence interval, 1.159-1.805]), but not hospital, mortality. In a propensity matched analysis, ICU mortality was higher in NIV than invasively ventilated patients with a PaO2/FiO2 lower than 150 mm Hg. NIV was used in 15% of patients with ARDS, irrespective of severity category. NIV seems to be associated with higher ICU mortality in patients with a PaO2/FiO2 lower than 150 mm Hg. Clinical trial registered with www.clinicaltrials.gov (NCT 02010073).

IntroductionWe assessed rates and predictive factors of non-invasive ventilation (NIV) failure in patients admitted to the intensive care unit (ICU) for non-hypercapnic acute hypoxemic respiratory failure (AHRF).MethodsThis is an observational cohort study using data prospectively collected over a three-year period in a medical ICU of a university hospital.ResultsAmong 113 patients receiving NIV for AHRF, 82 had acute respiratory distress syndrome (ARDS) and 31 had non-ARDS. Intubation rates significantly differed between ARDS and non-ARDS patients (61% versus 35%, P = 0.015) and according to clinical severity of ARDS: 31% in mild, 62% in moderate, and 84% in severe ARDS (P = 0.0016). In-ICU mortality rates were 13% in non-ARDS, and, respectively, 19%, 32% and 32% in mild, moderate and severe ARDS (P = 0.22). Among patients with moderate ARDS, NIV failure was lower among those having a PaO2/FiO2 >150 mmHg (45% vs. 74%, p = 0.04). NIV failure was associated with active cancer, shock, moderate/severe ARDS, lower Glasgow coma score and lower positive end-expiratory pressure level at NIV initiation. Among intubated patients, ICU mortality rate was 46% overall and did not differ according to the time to intubation.ConclusionsWith intubation rates below 35% in non-ARDS and mild ARDS, NIV stands as the first-line approach; NIV may be attempted in ARDS patients with a PaO2/FiO2 > 150. By contrast, 84% of severe ARDS required intubation and NIV did not appear beneficial in this subset of patients. However, the time to intubation had no influence on mortality.

The primary objective of the study was to evaluate and compare the effectiveness of four artificial intelligence (AI) platforms - ChatGPT, DeepSeek AI, Copilot, and Gemini - in providing solutions to common human resource (HR) challenges in corporate hospitals across India. A cross-sectional, questionnaire-based online study was conducted among 50 h professionals with over 10 years of experience across corporate hospitals in India. Ten real-world HR problem scenarios were designed and fed into the four AI platforms using a standardized prompt. The responses were anonymized, coded, and rated by participants on five criteria: accuracy, relevance, practicability, overall response, and likelihood of use. Data were analyzed using Epi Info (version 7.0) with Chi-square tests; P < 0.05 was considered significant. Among participants, 40% were male and 60% female; 50% had over 25 years of HR experience. ChatGPT and DeepSeek AI received the highest ratings for accuracy (85% and 84% "strongly agree"), relevance, and practicability, with 80%-81% indicating a high likelihood of future use. Copilot and Gemini showed moderate acceptance (55%-60% likelihood of use). There was no significant association between responses and participant gender or years of experience. AI platforms, particularly ChatGPT and DeepSeek AI, demonstrate strong potential to support HR professionals in solving routine challenges, improving efficiency, and decision-making. However, effective adoption requires training, data accuracy, and ethical safeguards to address issues such as bias and privacy.

To describe the outcomes of patients with bronchiectasis and acute respiratory failure (ARF) treated with noninvasive ventilation (NIV) and invasive mechanical ventilation (IMV) after a failure of conservative measures, and to identify the predictors of hospital mortality and NIV failure. Retrospective review of bronchiectatic patients on NIV (n = 31) or IMV (n = 26) for ARF over 8 years in a medical intensive care unit (ICU) experienced in NIV. At baseline, the NIV group had more patients with acute exacerbations without identified precipitating factors (87.1 vs. 34.6%, p < 0.001), higher pH (mean 7.25 vs. 7.18, p = 0.008) and PaO(2)/FiO(2) ratio (mean 249.4 vs. 173.2, p = 0.02), and a trend towards a lower APACHE II score (mean 25.3 vs. 28.4, p = 0.07) than the IMV group. There was no difference in hospital mortality between the two groups (25.8 vs. 26.9%, p > 0.05). The NIV failure rate (need for intubation or death in the ICU) was 32.3%. Using logistic regression, the APACHE II score was the only predictor of hospital mortality (OR 1.19 per point), and the PaO(2)/FiO(2) ratio was the only predictor of NIV failure (OR 1.02 per mmHg decrease). The hospital mortality of patients with bronchiectasis and ARF approximates 25% and is predicted by the APACHE II score. When selectively applied, NIV fails in one-third of the patients, and this is predicted by hypoxemia. Our findings call for randomised controlled trials to compare NIV versus IMV in such patients.

To characterize trends in noninvasive ventilation (NIV) and invasive mechanical ventilation (IMV) use over time in children with hematologic malignancy admitted to the PICU with acute respiratory failure (ARF), and to identify risk factors associated with NIV failure requiring transition to IMV. Retrospective cohort analysis using the Virtual Pediatric Systems (VPS, LLC) between January 1, 2010 and December 31, 2019. One hundred thirteen North American PICUs participating in VPS. Two thousand four hundred eighty children 0-21 years old with hematologic malignancy admitted to participating PICUs for ARF requiring respiratory support. None. There were 3013 total encounters, of which 868 (28.8%) received first-line NIV alone (NIV only), 1544 (51.2%) received first-line IMV (IMV only), and 601 (19.9%) required IMV after a failed NIV trial (NIV failure). From 2010 to 2019, the NIV only group increased from 9.6% to 43.1% and the IMV only group decreased from 80.1% to 34.2% (p < 0.001). The NIV failure group had the highest mortality compared with NIV only and IMV only (36.6% vs. 8.1%, vs. 30.5%, p < 0.001). However, risk-of-mortality (ROM) was highest in the IMV only group compared with NIV only and NIV failure (median Pediatric Risk of Mortality III ROM 8.1% vs. 2.8% vs. 5.5%, p < 0.001). NIV failure patients also had the longest median PICU length of stay compared with the other two study groups (15.2 d vs. 6.1 and 9.0 d, p < 0.001). Higher age was associated with significantly decreased odds of NIV failure, and diagnosis of non-Hodgkin lymphoma was associated with significantly increased odds of NIV failure compared with acute lymphoid leukemia. For children with hematologic malignancy admitted to the PICU with ARF, NIV has replaced IMV as the most common initial therapy. NIV failure rate remains high with high-observed mortality despite lower PICU admission ROM.

IntroductionDespite advances in respiratory distress syndrome (RDS) management over the past decade, non-invasive ventilation (NIV) failure is frequent and associated with adverse outcomes. There are insufficient data on the failure of different NIV strategies currently used in clinical practice in preterm infants.MethodsThis was a prospective, multicenter, observational study of very preterm infants [gestational age (GA) <32 weeks] admitted to the neonatal intensive care unit for RDS that required NIV from the first 30 min after birth. The primary outcome was the incidence of NIV failure, defined as the need for mechanical ventilation for <72 h of life. Secondary outcomes were risk factors associated with NIV failure and complication rates.ResultsThe study included 173 preterm infants with a median GA of 28 (IQR 27–30) weeks and a median birth weight of 1,100 (IQR 800–1,333) g. The incidence of NIV failure was 15.6%. In the multivariate analysis, lower GA (OR, 0.728; 95% CI, 0.576–0.920) independently increased the risk of NIV failure. Compared to NIV success, NIV failure was associated with higher rates of unfavorable outcomes, including pneumothorax, intraventricular hemorrhage, periventricular leukomalacia, pulmonary hemorrhage, and a combined outcome of moderate-to-severe bronchopulmonary dysplasia or death.ConclusionNIV failure occurred in 15.6% of the preterm neonates and was associated with adverse outcomes. The use of LISA and newer NIV modalities most likely accounts for the reduced failure rate. Gestational age remains the best predictor of NIV failure and is more reliable than the fraction of inspired oxygen during the first hour of life.

The effect of noninvasive ventilation (NIV) in patients with acute respiratory failure (ARF) after cardiac surgery is controversial. This study identified the feasibility of NIV and assessed the risk factors of NIV failure in patients with ARF after cardiac surgery. We retrospectively reviewed data from 112 patients with ARF requiring NIV and categorized them into the NIV failure and success groups. Patient data were extracted for further analysis, the primary outcomes were the need for endotracheal intubation and NIV-related in-hospital mortality. The risk factors for NIV failure in patients with post-extubation ARF was analyzed. The median time from extubation to NIV was 11 hours. No difference in the EuroSCORE existed between the two groups. NIV failed in 38.4% of the patients. The NIV failure group had a higher in-hospital mortality and stay at the longer intensive care unit (ICU). Most cases of NIV failure occurred within 1-48 hours of the treatment. The main causes of early NIV failure were a weak cough reflex and/or excessive secretions and hemodynamic instability. A Sequential Organ Failure Assessment (SOFA) score ≥10.5, vasoactive-inotropic score ≥6, and pneumonia were predictors of NIV failure, whereas a body mass index (BMI) ≥25.0 kg/m2 predicted NIV success. NIV was effective in the study population. Multiple organ dysfunction, pneumonia, and significant inotropic drug support before NIV were associated with NIV failure, whereas a BMI ≥25 kg/m2 was a predictor of NIV success.

Background Birth control methods (BCMs) are often underutilized or misunderstood, especially among young individuals entering their reproductive years. With the growing reliance on artificial intelligence (AI) platforms for health-related information, this study evaluates the performance of ChatGPT-4o and Google Gemini in addressing commonly asked questions about BCMs. Methods Thirty questions, derived from the American College of Obstetrics and Gynecologists (ACOG) website, were posed to both AI platforms. Questions spanned four categories: general contraception, specific contraceptive types, emergency contraception, and other topics. Responses were evaluated using a five-point rubric assessing Relevance, Completeness, and Lack of False Information (RCL). Overall scores were calculated by averaging the rubric scores. Statistical analysis, including the Wilcoxon Signed-Rank test, Friedman test, and Kruskal-Wallis test, was performed to compare metrics. Results ChatGPT-4o and Google Gemini provided high-quality responses to birth control-related queries, with overall scores averaging 4.38 ± 0.58 and 4.37 ± 0.52, respectively, both categorized as "very good" to "excellent." ChatGPT-4o demonstrated higher scores in the lack of false information, based on descriptive statistics (4.70 ± 0.60 vs. 4.47 ± 0.73), while Google Gemini outperformed in relevance, with a statistically significant difference (4.53 ± 0.57 vs. 4.30 ± 0.70, p = 0.035, large effect size). Completeness scores were comparable (p = 0.655). Statistical analyses revealed no significant differences in overall performance (p = 0.548), though Google Gemini demonstrated a potential trend of stronger performance in the "Other Topics" category. Within-model variability showed ChatGPT-4o had more pronounced differences among metrics (moderate effect size, Kendall's W = 0.357), while Google Gemini exhibited smaller variability (Kendall's W = 0.165). These findings suggest that both platforms offer reliable and complementary tools for addressing knowledge gaps in contraception, with nuanced strengths that warrant further exploration. Conclusions ChatGPT-4o and Google Gemini provided reliable and accurate responses to BCM-related queries, with slight differences in strengths. These findings underscore the potential of AI tools, in addressing public health information needs, particularly for young individuals seeking guidance on contraception. Further studies with larger datasets may elucidate nuanced differences between AI platforms.