ABSTRACT
A growing body of research shows that improving diagnostic and treatment efficiency can save lives. Artificial intelligence (AI) in healthcare is a new research topic. Human engineering and domain expertise were initially necessary to transform raw data into algorithms. One type of machine learning called deep learning creates representations from raw data, with an algorithm determining how much change should be done. That deep learning can learn from huge amounts of data is its utility. It can categorize, analyze, and forecast data to identify patterns. Weak clinical integration makes measuring current effect difficult, but simulation data reveals AI's potential to enhance screening accuracy and efficiency, minimize effort, and potentially diagnose sickness better than experts. In medical imaging, deep learning algorithms categorize, segment, and identify objects in pictures and movies. Studies on AI-based breast cancer, cardiac imaging, and melanoma screening showed promising results. Evolved deep learning algorithms such as convolutional neural networks (CNNs) effectively assess spatially invariant input. In trials assessing their diagnostic utility in object classification, CNNs were close to or at the physician level in identifying skin cancer, cardiovascular risk, and breast cancer. During the COVID-19 pandemic, AI was used for everything from vaccine/drug discovery to diagnosis, according to Abd-Alrazaq. Now, most AI systems actively combine physicians and algorithms, enhancing accuracy and efficiency. © 2023 Elsevier Inc. All rights reserved.
ABSTRACT
Introduction: The impact of Covid-19 on outcomes of In-hospital cardiac arrest IHCA remains unclear. Aims & Methodology: We, conducted a retrospective cohort study to compare the characteristics and outcome of CPR in Covid positive V/s Covid negative patients, at our institute. Data were retrieved from a review of the medical records of all patients who underwent cardiac arrest during the pandemic period. Taking Covid-19 as the exposure variable, our sample population was divided into two cohorts. Results: Eighty patients who underwent CPR were included in the study with;40 patients in each group. The mean age was 61 in the positive group and 65 in the negative group. The male to female ratio was equal in the covid positive group but the male population was higher in the negative group. (77%). The most frequent comorbidities were the same (Diabetes, hypertension, and ischaemic heart disease) in both groups. All of the patients had ARDS in Covid positive group while septic shock was the commonest diagnosis (48.9%) in the negative group. In the positive group, 74% of patients were in high dependency unit (HDU) and 26% were in ICU. In the negative group,94% of patients were in HDU and 6% in ICU. The initial rhythm was Pulseless electrical activity (PEA) in 52% of the positive group and all of the patients (100%) in the negative group. Eight percent of covid positive patients had shockable rhythm while remaining had asystole. The median duration of CPR was 15 minutes in the positive group and 17 minutes in the negative group. Although the return of spontaneous circulation (ROSC) was achieved in 14% of patients in covid positive, none of them survived to discharge. In the negative group, 30% of patients achieved ROSC while survival to discharge was 15%. The odds of mortality were 6.88 [95% confidence interval (CI)0.789-60] times higher in COVID-19-positive patients, compared to negative patients. Conclusions: Covid-19 infection is associated with poor outcomes in IHCA compared to non-covid illnesses.