Evaluating the accuracy and reliability of AI chatbots in disseminating the content of current resuscitation guidelines: a comparative analysis between the ERC 2021 guidelines and both ChatGPTs 3.5 and 4.

AIM OF THE STUDY: Artificial intelligence (AI) chatbots are established as tools for answering medical questions worldwide. Healthcare trainees are increasingly using this cutting-edge technology, although its reliability and accuracy in the context of healthcare remain uncertain. This study evaluated the suitability of Chat-GPT versions 3.5 and 4 for healthcare professionals seeking up-to-date evidence and recommendations for resuscitation by comparing the key messages of the resuscitation guidelines, which methodically set the gold standard of current evidence and recommendations, with the statements of the AI chatbots on this topic. METHODS:  This prospective comparative content analysis was conducted between the 2021 European Resuscitation Council (ERC) guidelines and the responses of two freely available ChatGPT versions (ChatGPT-3.5 and the Bing version of the ChatGPT-4) to questions about the key messages of clinically relevant ERC guideline chapters for adults. (1) The content analysis was performed bidirectionally by independent raters. The completeness and actuality of the AI output were assessed by comparing the key message with the AI-generated statements. (2) The conformity of the AI output was evaluated by comparing the statements of the two ChatGPT versions with the content of the ERC guidelines. RESULTS: In response to inquiries about the five chapters, ChatGPT-3.5 generated a total of 60 statements, whereas ChatGPT-4 produced 32 statements. ChatGPT-3.5 did not address 123 key messages, and ChatGPT-4 did not address 132 of the 172 key messages of the ERC guideline chapters. A total of 77% of the ChatGPT-3.5 statements and 84% of the ChatGPT-4 statements were fully in line with the ERC guidelines. The main reason for nonconformity was superficial and incorrect AI statements. The interrater reliability between the two raters, measured by Cohen's kappa, was greater for ChatGPT-4 (0.56 for completeness and 0.76 for conformity analysis) than for ChatGPT-3.5 (0.48 for completeness and 0.36 for conformity). CONCLUSION: We advise healthcare professionals not to rely solely on the tested AI-based chatbots to keep up to date with the latest evidence, as the relevant texts for the task were not part of the training texts of the underlying LLMs, and the lack of conceptual understanding of AI carries a high risk of spreading misconceptions. Original publications should always be considered for comprehensive understanding.

Intensive care unit cardiac arrest among very elderly critically ill patients - is cardiopulmonary resuscitation justified?

INTRODUCTION: The proportion of very elderly patients in the intensive care unit (ICU) is expected to rise. Furthermore, patients are likely more prone to suffer a cardiac arrest (CA) event within the ICU. The occurrence of intensive care unit cardiac arrest (ICU-CA) is associated with high mortality. To date, the incidence of ICU-CA and its clinical impact on outcome in the very old (≥ 90 years) patients treated is unknown. METHODS: Retrospective analysis of all consecutive critically ill patients ≥ 90 years admitted to the ICU of a tertiary care university hospital in Hamburg (Germany). All patients suffering ICU-CA were included and CA characteristics and functional outcome was assessed. Clinical course and outcome were assessed and compared between the subgroups of patients with and without ICU-CA. RESULTS: 1,108 critically ill patients aged ≥ 90 years were admitted during the study period. The median age was 92.3 (91.0-94.2) years and 67% (n = 747) were female. 2% (n = 25) of this cohort suffered ICU-CA after a median duration 0.5 (0.2-3.2) days of ICU admission. The presumed cause of ICU-CA was cardiac in 64% (n = 16). The median resuscitation time was 10 (2-15) minutes and the initial rhythm was shockable in 20% (n = 5). Return of spontaneous circulation (ROSC) could be achieved in 68% (n = 17). The cause of ICU admission was primarily medical in the total cohort (ICU-CA: 48% vs. No ICU-CA: 34%, p = 0.13), surgical - planned (ICU-CA: 32% vs. No ICU-CA: 37%, p = 0.61) and surgical - unplanned/emergency (ICU-CA: 43% vs. No ICU-CA: 28%, p = 0.34). The median Charlson Comorbidity Index (CCI) was 2 (1-3) points for patients with ICU-CA and 1 (0-2) for patients without ICU-CA (p = 0.54). Patients with ICU-CA had a higher disease severity according to SAPS II (ICU-CA: 54 vs. No ICU-CA: 36 points, p < 0.001). Patients with ICU-CA had a higher rate of mechanically ventilation (ICU-CA: 64% vs. No ICU-CA: 34%, p < 0.01) and required vasopressor therapy more often (ICU-CA: 88% vs. No ICU-CA: 41%, p < 0.001). The ICU and in-hospital mortality was 88% (n = 22) and 100% (n = 25) in patients with ICU-CA compared to 17% (n = 179) and 28% (n = 306) in patients without ICU-CA. The mortality rate for patients with ICU-CA was observed to be 88% (n = 22) in the ICU and 100% (n = 25) in-hospital. In contrast, patients without ICU-CA had an in-ICU mortality rate of 17% (n = 179) and an in-hospital mortality rate of 28% (n = 306) (both p < 0.001). CONCLUSION: The occurrence of ICU-CA in very elderly patients is rare but associated with high mortality. Providing CPR in this cohort did not lead to long-term survival at our centre. Very elderly patients admitted to the ICU likely benefit from supportive care only and should probably not be resuscitated due to poor chance of survival and ethical considerations. Providing personalized assurances that care will remain appropriate and in accordance with the patient's and family's wishes can optimise compassionate care while avoiding futile life-sustaining interventions.

Association of early changes in arterial carbon dioxide with acute brain injury in adult patients with extracorporeal membrane oxygenation: A ten-year retrospective study in a German tertiary care hospital.

PURPOSE: To assess the association between fluctuations of arterial carbon dioxide early after start of extracorporeal membrane oxygenation (ECMO) with intracranial hemorrhage (ICH) or ischemic stroke (IS). MATERIALS AND METHODS: This single-center retrospective study included patients who required ECMO for circulatory or respiratory failure between January 2011 and April 2021 and for whom a cerebral computed tomography (cCT) scan was available. Multivariable logistic regression models were fitted to evaluate the association between the relative change of arterial carbon dioxide (RelΔPaCO2) and ICH, IS or a composite of ICH, IS, and mortality. RESULTS: In 618 patients (venovenous ECMO: n = 295; venoarterial ECMO: n = 323) ICH occurred more frequently in patients with respiratory failure (19.0%) compared with patients with circulatory failure (6.8%). Conversely, the incidence of IS was higher in patients with circulatory failure (19.2%) compared with patients with respiratory failure (4.7%). While patients with ECMO for respiratory failure were more likely to have ICH (OR 3.683 [95% CI: 1.855;7.309], p < 0.001), they had a lower odds for IS (OR 0.360 [95%CI: 0.158;0.820], p = 0.015) compared with patients with circulatory failure. There was no significant association between RelΔPaCO2 and ICH or IS. CONCLUSIONS: Irrespective of the indication for ECMO, we did not find a significant association between the relative change in PaCO2 early after ECMO initiation and acute brain injury. Aside from early PaCO2 decline at cannulation, future studies should address fluctuations of PaCO2 throughout the course of ECMO support and their effect on acute brain injury.

Late Onset of Primary Hemophagocytic Lymphohistiocytosis (HLH) with a Novel Constellation of Compound Heterozygosity Involving Two Missense Variants in the PRF1 Gene.

Hemophagocytic lymphohistiocytosis (HLH) is a rare but in most cases life-threatening immune-mediated disease of the hematopoietic system frequently associated with hematologic neoplasms. Here, we report on a case in which we detected a novel constellation of two missense variants affecting the PRF1 gene, leading to de novo primary HLH. Diagnostics included a comprehensive clinical work-up and standard methods of hematopathology as well as extended molecular genomics based on polymerase chain reaction (PCR) reactions and the calculation of three-dimensional molecule reconstructions of PRF1. Subsequently, a comprehensive review of the literature was performed, which showed that this compound heterozygosity has not been previously described. The patient was a 20-year-old female. Molecular diagnostics revealed two heterozygous missense variants in the PRF1 gene (A91V and R104C) on exon 2. Apart from the finding of two inconclusive genetic variants, all clinical criteria defined by the HLH study group of Histiocyte Society were met at initial presentation. The final diagnosis was made in cooperation with the Consortium of German HLH-reference centers. Here, chemotherapy did not lead to sufficient sustained disease control. Therefore, the decision for allogenic hematopoietic stem cell transplantation (alloHSCT) was made. Hitherto, the duration of response was 6 months. Due to severe and unmanageable hepatic graft-versus-host disease (GvHD), the patient died. We report on a novel constellation of a compound heterozygosity containing two missense variants on exon 2 of the PRF1 gene. To the authors' best knowledge, this is the first presentation of a primary HLH case harboring this genomic constellation with late-onset clinical manifestation.

[Significance of natural language processing and chat-based generative language models]. / Stellenwert von Natural Language Processing und chatbasierten Generative Language Models.

BACKGROUND: Natural language processing (NLP) has experienced significant growth in recent years and shows potential for broad impacts in scientific research and clinical practice. OBJECTIVE: This study comprises an exploration of the role of NLP in scientific research and its subsequent effects on traditional publication practices, as well as an evaluation of the opportunities and challenges offered by large language models (LLM) and a reflection on necessary paradigm shifts in research culture. MATERIALS AND METHODS: Current LLMs, such as ChatGPT, and their potential applications were compared and assessed. An analysis of the literature and case studies on the integration of LLMs into scientific and clinical practice was conducted. RESULTS AND CONCLUSION: LLMs provide enhanced access to and processing capabilities of text-based information and represent a vast potential for (medical) research as well as daily clinical practice. Chat-based LLMs enable efficient completion of often time-consuming tasks, but due to their tendency for hallucinations, have a significant limitation. Current developments require critical examination and a paradigm shift to fully exploit the benefits of LLMs and minimize potential risks.

[Artificial intelligence: How will ChatGPT and other AI applications change our everyday medical practice?] / Künstliche Intelligenz: Wie werden ChatGPT und andere KI-Anwendungen unseren ärztlichen Alltag verändern?

BACKGROUND: With the free provision of the chat robot "ChatGPT" by the company OpenAI in November 2022, an application of artificial intelligence (AI) became tangible for everyone. OBJECTIVES: An explanation of the basic functionality of large language models (LLM) is given, followed by a presentation of application options of ChatGPT in medicine, and an outlook and discussion of possible dangers of AI applications. METHODS: Problem solving with ChatGPT using concrete examples. Analysis and discussion of the available scientific literature. RESULTS: There has been a significant increase in the use of AI applications in scientific work, especially in scientific writing. Wide application of LLM in writing medical documentation is conceivable. Technical functionality allows the use of AI applications as a diagnostic support system. There is a risk of spreading and entrenching inaccuracies and bias through application of LLM. Regulation of this new technology is pending. CONCLUSION: AI applications such as ChatGPT have the potential to permanently change everyday medical practice. An examination of this technology and evaluation of opportunities and risks is warranted.

Assessing the weak links - Necessity and impact of regional cardiac arrest awareness campaigns for laypersons.

Introduction: Public knowledge of out-of-hospital cardiac arrest (OHCA), and initiation of basic life support (BLS) is crucial to increase survival in OHCA. Methods: The study analysed the knowledge and willingness to perform BLS of laypersons passing an AED at a public train station. Interviewees were recruited at two time points before and after a four year-long structured regional awareness campaign, which focused on call, compress, shock in a mid-size European city (270,000 inhabitants). Complete BLS was defined as multiple responses for call for help; initiation of chest compressions; and usage of an AED, without mentioning recovery position. Minimal BLS was defined as call for help and initiation of chest compressions. Results: A total of 784 persons were interviewed, 257 at baseline and 527 post-campaign. Confronted with a fictional OHCA, at baseline 8.5% of the interviewees spontaneously mentioned actions for complete BLS and 17.9% post-campaign (p = 0.009). An even larger increase in knowledge was seen in minimal BLS (34.6% vs 60.6%, p < 0.001). Conclusion: After a regional cardiac arrest awareness campaign, we found an increase in knowledge of BLS actions in the lay public. However, our investigation revealed severe gaps in BLS knowledge, possibly resulting in weak first links of the chain of survival.

Infective endocarditis - A review of current therapy and future challenges.

Etiological, microbiological and epidemiological factors changed over time, but mortality rates remain high in infective endocarditis (IE). Healthcare-associated IE is nowadays responsible for a significant proportion of cases due to increasing numbers of cardiac devices. Cardiac implantable electronic devices, transcatheter aortic valve replacement, and percutaneous valve repair are meanwhile used, especially in old and sick patients. In suspected IE modified Duke criteria, integrating clinical results, imaging, and biomarkers are traditionally applied. Newer imaging technologies such as multi-slice computed tomography, photon-emission computed tomography, and magnetic resonance imaging might add value to conventional echocardiography in diagnosis and management of IE. Treatment consists of long-term antibiotic therapy, infectiological source control and/or cardiac surgery. Recently, antibiotic parenteral outpatient regimens and partial oral treatment strategies were shown to shorten hospital stays in patients suffering from IE. However, it remains unclear how to best select patients for partial oral therapy. This review describes new trends in diagnosing, imaging, and treating IE in a changing patient collective with particular focus on patients with implantable cardiac devices.