AI Based Examination System: A Paradigm Shift in Education Sector

COVID pandemic has brought many changes in day to day working of almost every profession. Education is one such arena where switching to online mode of teaching and assessment was received by many educators as a paradigm shift in pedagogy. Conducting classes online was not a satisfying experience for many but was recognized as the need of the hour. But when it came to online assessment of work, the bright window of light generated an array of questions. Conducting online assessments using live online proctored test, or submissions of assessment sheets has become a new-normal. Many technology platforms provide solutions for online face-to-face meetings, Artificial intelligence based proctored examinations and online objective type assessment sheets with pre-assigned rubrics for marks and evaluation. Since there was a situation where the proctor needs to be present and observe the behaviour of the student, their movements and also the application is to check the face to be recognised with the data already entered, if there is any noise or if someone else if present in the room or not.All these features had made the online examination system little efficient and effective for the universities to conduct the exam hassle free and from remote access. This paper presents a study of various online assessment tools presented by technology companies for smooth conduct of examinations. This study also presents a comparative analysis of features of these tools and their utilization in different domains. © 2022 IEEE.

Deep Learning Techniques to Identify and Classify COVID-19 Abnormalities on Chest X-ray Images

The new coronavirus disease (COVID-19) comprises the public health systems around the world. The number of infected people and deaths are escalating day-to-day, which puts enormous pressure on healthcare systems. COVID-19 symptoms include fatigue, cough, and fever. These symptoms are also diagnosed for other pneumonia, which creates complications in identifying COVID-19, especially throughout the influenza season. The rise of the COVID-19 pandemic among individuals has made it essential to improve medical image screening of this pneumonia. Rapid identification is a necessary step to stop the spread of this virus and plays a vital role in early detection. With this as a motivator, we applied deep learning techniques to diagnose the coronavirus using chest X-ray images and to implement a robust AI application to classify COVID-19 pneumonia from non-COVID-19 for the respiratory system in these images. This paper proposes different deep learning algorithms, including classification and segmentation methods. By taking advantage of convolutional neural network models, we exploited different pre-trained deep learning models such as (ResNet50, ResNet101, VGG-19, and U-Net architectures) to extract features from chest X-ray images. Four datasets of chest X-ray images have been employed to assess the performance of the proposed methods. These datasets have been split into 80% for training and 20% for validation of the architectures. The experimental results showed an overall accuracy of 99.42% for the classification and 93% for segmentation approaches. The proposed approaches can help radiologists and medical specialists to identify the insights of infected regions for the respiratory system in the early stages. © 2022 SPIE.

Follow-up arrangements in head and neck cancer clinics during the COVID-19 pandemic: results from two tertiary UK head and neck cancer centres.

INTRODUCTION: The aim of this paper is to report the pattern of follow-up that occurred for a cohort of head and neck cancer (HNC) patients across two large centres in the UK (Aintree and Leeds), as a consequence of the COVID-19 pandemic. METHODS: Patients had been treated for HNC with curative intent between April 2017 and October 2019 by 14 oral and maxillofacial (OMFS) and ear nose and throat (ENT) oncology surgeons in the Patient Concerns Inventory intervention trial. In October 2020, hospital records were reviewed, and information collected on the timing and mode (face-to-face/telephone/video) of follow-up consultations. In addition, recurrence, second primary tumours and deaths were recorded. RESULTS: At the start of 'lockdown', 212 members of the cohort were known to be alive. During the post-lockdown period (follow-up appointment data comprised 5 months in Aintree and 7 months in Leeds) 7 died and 13 were identified as palliative/recurrence/new primary/metastases ('new event'). In Aintree, the first ENT/OMFS consultations after lockdown were 51 (67%) telephone and 25 (33%) face-to-face appointments. In Leeds, 46 (78%) consultations were by telephone and 13 (22%) were face-to-face. The second ENT/OMFS consultations post lockdown included 11 (44%) telephone and 14 (56%) face-to-face in Aintree, and 21 (75%) telephone and 7 (25%) face-to-face in Leeds. CONCLUSIONS: These data suggest that clinicians favoured remote consultations. Variations in practice were observed but reached a point of a 'hybrid follow-up approach' that included both face-to-face and remote consultations. With the emergence of telemedicine, clinicians may consider a follow-up model tailored to risk stratification. The development of the mode of such a consultation model needs further evaluation.

Low Incidence of COVID-19 Infection in Patients with Acute Myeloid Leukemia Undergoing Reduced Intensity/Venetoclax Based Treatment: Initial Results of the PACE Prospective Clinical Study from the UK Trials Acceleration Program

The impact of Coronavirus disease 2019 (COVID-19) on outcomes in patients with cancer remains unclear. Acute Myeloid Leukemia (AML)/high-risk myelodysplasia (MDS) are common hematological malignancies resulting in profound immunosuppression, which is exacerbated by intensive and less-intensive chemotherapy. Importantly, venetoclax based regimens have been increasingly used during the pandemic as a strategy to reduce patient hospitalization however, there is little information concerning the impact of such regimens on COVID-19 infection rates. We therefore opened a prospective clinical study (PACE), at the start of the current pandemic in April 2020 to characterize the risk of COVID-19 infection in patients with AML/MDS-EB2 receiving intensive or non-intensive treatment, including patients treated with venetoclax-based regimens. The primary aim was to determine the incidence of COVID-19 in patients with AML /MDS-EB2 including both, prior to study entry and during treatment until 4 weeks after the last cycle of treatment. Secondary aims were to: characterize the presentation of COVID-19;define the severity and type of both non-COVID-19 and COVID-19 infections;and undertake an exploratory analysis to quantify the incidence of COVID-19 infection in patients receiving (less-intensive) venetoclax based regimens. All analysis conducted to date has been descriptive. 211/230 recruited patients had full treatment histories available, of whom 116 patients received intensive chemotherapy and 95 low intensity regimens. 48 patients received a venetoclax-based regimen. The median age of the non-intensive treatment arm was 72 years;(range 19.1-86.5) and of the intensive arm was 59 years (range 16.1-76.1). There were more cases of secondary AML and relapsed disease in the non-intensive arm as compared to the intensive arm. 25/226 evaluable patients tested positive for COVID-19 as defined by positive SARS-CoV2 PCR test, 10 with a prior diagnosis at study entry and 15 tested positive during the study. The incidence of COVID-19 infection for patients with AML/MDS-EB2 was 11.1% (90%CI: 7.8%-15.1%) (Table). A lower proportion of patients (n=6/91 6.6%) undergoing non-intensive treatment suffered COVID-19 as compared to those undergoing more intensive chemotherapy regimens (n=19/116, 16.4%). Specifically, only 3/48 (6.3%) patients undergoing a venetoclax regimen were infected with SARS-CoV2. The most common presenting symptoms of COVID-19 in this study, regardless of the intensity of chemotherapy, was fever and cough with 6/25 patients asymptomatic. The risk of death at 30 days following study entry in patients who had prior COVID-19 infection or who contracted COVID-19 during this period was 13.6%, compared to 3.9% in the overall cohort without COVID-19 infection. There was a lower incidence of non-COVID-19 related infections in patients receiving venetoclax-based regimens, n=43 infections in 24 (50.0%) of patients;with 313 infections in 94 (81%) of intensively treated patients. The overall occurrence of non-COVID-19 infection in the non-intensive arm was 87 infections in 50 (54.9%) patients. Our multi-center study provides real-world estimates for the incidence and presentation of COVID-19 infection in a cohort of patients with AML/MDS-EB2, and indicates a higher risk of death at 30 days in patients with prior COVID-19 infection prior to, or during treatment. Venetoclax based, and other non-intensive, regimens, increasingly implemented during the pandemic, to minimize patient exposure and reduce usage of hospital beds, appeared to be associated with a low incidence of COVID-19. Further follow-up will be required to understand the long-term impact of this strategy. Analysis of immune responses to COVID-19 infection and vaccination is on-going. Acknowledgments: This study was funded by Cure Leukaemia under the Trials Acceleration Program (TAP), and grants from BMS and Blood Cancer UK. [Formula presented] Disclosures: Loke: Novartis: Other: Travel;Janssen: Honoraria;Amgen: Honoraria;Pfizer: Honoraria;Daichi Sankyo: Other: Travel. K apper: Pfizer: Consultancy, Speakers Bureau;Astellas: Ended employment in the past 24 months, Speakers Bureau;Jazz: Consultancy, Speakers Bureau;Novartis: Consultancy, Research Funding, Speakers Bureau. Khan: Abbvie: Honoraria;Astellas: Honoraria;Takeda: Honoraria;Jazz: Honoraria;Gilead: Honoraria;Novartis: Honoraria. Dillon: Amgen: Other: Research support (paid to institution);Astellas: Consultancy, Other: Educational Events, Speakers Bureau;Menarini: Membership on an entity's Board of Directors or advisory committees;Novartis: Membership on an entity's Board of Directors or advisory committees, Other: Session chair (paid to institution), Speakers Bureau;Pfizer: Consultancy, Membership on an entity's Board of Directors or advisory committees, Other: educational events;Jazz: Other: Education events;Abbvie: Consultancy, Membership on an entity's Board of Directors or advisory committees, Other: Research Support, Educational Events;Shattuck Labs: Membership on an entity's Board of Directors or advisory committees. Culligan: AbbVie Ltd: Honoraria, Speakers Bureau;Celgene Ltd: Honoraria, Speakers Bureau;Gilead: Honoraria, Speakers Bureau;Jazz Pharma: Honoraria, Speakers Bureau;Takeda UK Ltd: Honoraria, Speakers Bureau. McMullin: Bristol Myers Squibb: Membership on an entity's Board of Directors or advisory committees, Other: clinical trial support, Research Funding;Celgene: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau;AbbVie: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau;Novartis: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau;AOP Orphan: Research Funding, Speakers Bureau. Murthy: Abbvie: Other: support to attend educational conferences. Craddock: Novartis Pharmaceuticals: Other: Advisory Board;Celgene/BMS: Membership on an entity's Board of Directors or advisory committees, Research Funding.

Immediate postoperative care on high dependency unit or ward following microvascular free tissue transfer: lessons learnt from a change in practice imposed during the COVID-19 pandemic.

The COVID-19 pandemic resulted in sudden changes to the established practice of using the high dependency unit (HDU) for the first night of postoperative care following microvascular free tissue transfer. Patients were managed instead on the head and neck ward. This retrospective case-note review aimed to report outcomes in consecutive patients treated before and during the pandemic, and to reflect on the implications of ward-based rather than HDU care. A total of 235 patients had free tissue transfer between 3 January 2019 and 25 February 2021: 125 before (lockdown 23 March 2020), and 110 during the pandemic (52 ward-managed and 58 HDU-managed). There were subtle case-mix differences during the pandemic, with 92% of ward-treated patients having oral cancers compared with 64% of HDU patients, and 73% of ward patients having a tracheostomy compared with 40% of HDU patients. Ward patients were less likely to receive electrolyte replacement (45% HDU vs 0% ward) and inotropes (12% HDU vs 2% ward). There were fewer returns to theatre for evacuation of a haematoma or re-anastomosis during the pandemic than there were before it. Other than fewer haematoma complications during the pandemic, the nature of complications was similar. In conclusion, the dramatic changes imposed by the pandemic have shown that the ward is a safe place for patients to be cared for immediately postoperatively, and it alleviates the bed pressures experienced in HDU. Careful case selection and clear criteria are required to identify patients who need the HDU.

COVID-19 advanced respiratory physiology (CARP) wearable respiratory monitoring: early insights

P24 Figure 1CARP trial wearable respiratory rate, respiratory support and outcome data from 3 patients with severe COVID-19[Figure omitted. See PDF]Results156 patients were screened, with 77 recruited to the CARP trial. 32 patients required non-invasive respiratory support, of which 14 were escalated to mechanical intubation. 17 patients died within trial.Bland-Altman analyses of paired RR data confirmed that wearable sensor data shows good agreement with critical care RR monitoring (Phillips Intellivue MX700), and that ward-based intermittent clinician RR measurements were imprecise.From the initial utility review of CARP physiology data visualisations, rising hourly average RR >25/min is associated with subsequent patient deterioration. Improving and stable hourly average RR of <25/min associates with stable respiratory failure and improvement to hospital discharge (figure 1).ConclusionContinuous wearable respiratory rate remote monitoring in COVID-19 inpatients is feasible. Planned machine learning and time-series analyses of the detailed physiology and clinical endpoint data will determine appropriate cut-offs and feature importance for deteriorating patient risk predictions. The CARP clinical dashboard provides an infrastructure for future implementation and evaluation of these AI insights.

What are relevant, feasible and effective approaches to promote acceptance, uptake and adherence to physical distancing measures for COVID-19 prevention and control?

Physical distancing measures have been implemented worldwide to contain the transmission of COVID-19, but how best to communicate with the public to promote acceptance, uptake and adherence to these measures is less clear. This rapid review analysed evidence regarding communication with individuals and communities within the wider structural and sociopolitical context of the pandemic to support public health decision-makers when planning and implementing physical distancing measures. Findings indicated the critical role played by public communication and information in the pandemic response. Consistent features of effective communication included clear, consistent and actionable content;attention to the timing and currency of messages;consideration of the audiences for communication within and across populations;and deliberate considerations of tailoring and equity to ensure diverse population groups are reached and existing inequalities addressed. Comprehensive practical support, including access to essential services and financial support, was also critical to promote acceptance, uptake and adherence to required measures. Findings also emphasized the importance of building and maintaining public trust in authorities and of engaging communities when planning and delivering messages related to physical distancing measures.

Mitigating the covid-19 impact on COPD care: Rapid development of remote recruitment Processes to a digital self-management service

Background Shielding requirements and interruptions to routine care resulting from the COVID-19 pandemic present substantial risks to COPD patient's outcomes. Adverse impacts including increased winter admissions, nosocomial infection risks, impaired quality of life and high service costs are anticipated. A digital remote-management service can potentially mitigate these by supporting routine COPD care and selfmanagement. In the InnovateUK funded 'DYNAMIC' project we've developed a fully integrated COPD digital support service. Patientheld smartphone progressive web application prompts daily patient reported outcome (PRO) completion with red-ambergreen self-management advice, rescue medication and linked resources including inhaler, exercise and dyspnoea management videos. Secure asynchronous patient-clinician messaging and EHR-integrated clinician dashboard with PRO data visualisation and structured clinical data enables self-management support and provides a new channel for scheduled care. Methods As COVID-19 pandemic emerged we paused the RECEIVER trial (NCT04240353) and undertook rapid codesign, governance approval and deployment of an SMS invite and web-based registration system https://support.nhscopd.scot This provides access and remote setup to the DYNAMIC COPD service for all patients in our organisation. Website tracking analytics support service evaluation and process optimisation. Feedback was obtained from 57 patients who registered an interest via website but didn't progress to setup in the service. Results 2373 high-risk COPD patients were identified from secondary care datasets. 797 patients did not have a known mobile phone number. 1576 SMS invites were sent in batches May-July 2020. 599 unique visits to support site (38%) with 195 completed applications (12.3%) to the digital service were obtained. As of August 2020, 112 patients (7.1%) have completed service setup, which includes clinician virtual review with optimisation of COPD interventions, self-management planning and COPD MDT input as required. Patient app usage in this scale-up cohort matches positive experience from the RECEIVER trial. Qualitative data highlighted requirement for increased information about the service at invitation, which has been addressed. Conclusions Novel processes for digital recruitment and remote setup have widened access to NHS Scotland COPD support service. This is a model strategy for service implementation and evaluation. Further cycles of mail-based and SMS invites and media awareness campaign will follow.

Receiver trial interim analysis: Reduction in COPD admissions with digitally supported selfmanagement

Background The Remote-management of COPD: Evaluating Implementation of Digital Innovations to Enable Routine Care (RECEIVER, NCT04240353) observational cohort trial commenced September 2019. Clinician-patient co-designed progressive web application captures patient-reported outcomes (PROs), Fitbit and home NIV data, and provides self-management resources including asynchronous messaging. Cloud-based clinician dashboard integrates these with clinical summary and event data. Inclusion criteria are diagnosis of COPD with recent severe exacerbation and/or chronic hypercapnia, and daily access to smartphone, tablet or computer. Primary endpoint is patient use of the application. Secondary endpoints include impact on clinical outcomes and deriving AI-based risk predictive models. Methods We performed a planned interim effectiveness analysis of recruitment, patient-app utilisation and clinical outcomes at week 40, reflecting a minimum of 3 months follow up per patient. Results 283 patients were screened with 83 patients (average age 65 years) enrolled in RECEIVER by March 2020. Recruitment was paused then due to COVID-19 pandemic. 25/83 patients had home NIV or CPAP. Of the 143 excluded patients, 41 had no access to technology (average age 71 yrs), 42 had not had a recent exacerbation (average age 67 yrs) and 23 declined to participate (average age 67 yrs). An average of 4.6/patient/week daily PROs were completed, with usage sustained through to week 40. Improvement in patient outcomes versus their preceding year, with reduction in annualised admissions, occupied bed days and community reviews is noted (figure 1). Improved event ratios are maintained in subgroup analyses of home NIV patients, and if follow up is censored at UK COVID-19 lockdown. Conclusion Interim analyses of the RECEIVER trial are encouraging, with sustained patient use of the application, and associated positive impact on patient outcomes. Older age of patients lacking access to digital technology is notable. To support post COVID-19 NHS recovery we have scaled-up the digital self-management service, offering this to all COPD patients in NHS GG&C. RECEIVER trial dataset will be combined with large NHS GG&C SafeHaven historical cohort and the scale-up patient data for machine-learning analyses. We aim to train, validate and operationalise prediction models for 12-month mortality, 3-month re-admission and 72-hour exacerbation risk.

Receiver trial interim analysis: Reduction in COPD admissions with digitally supported self-management

Background: The Remote-management of COPD: Evaluating Implementation of Digital Innovations to Enable Routine Care (RECEIVER, NCT04240353) observational cohort trial (Figure presented) commenced September 2019. Clinician-patient co-designed progressive web application captures patient-reported outcomes (PROs), Fitbit and home NIV data, and provides self-management resources including asynchronous messaging. Cloud-based clinician dashboard integrates these with clinical summary and event data. Inclusion criteria are diagnosis of COPD with recent severe exacerbation and/or chronic hypercapnia, and daily access to smartphone, tablet or computer. Primary endpoint is patient use of the application. Secondary endpoints include impact on clinical outcomes and deriving AI-based risk predictive models. Methods: We performed a planned interim effectiveness analysis of recruitment, patient-app utilisation and clinical outcomes at week 40, reflecting a minimum of 3 months follow up per patient. Results: 283 patients were screened with 83 patients (average age 65 years) enrolled in RECEIVER by March 2020. Recruitment was paused then due to COVID-19 pandemic. 25/83 patients had home NIV or CPAP. Of the 143 excluded patients, 41 had no access to technology (average age 71 yrs), 42 had not had a recent exacerbation (average age 67 yrs) and 23 declined to participate (average age 67 yrs). An average of 4.6/patient/week daily PROs were completed, with usage sustained through to week 40. Improvement in patient outcomes versus their preceding year, with reduction in annualised admissions, occupied bed days and community reviews is noted (figure 1). Improved event ratios are maintained in subgroup analyses of home NIV patients, and if follow up is censored at UK COVID-19 lockdown. Conclusion: Interim analyses of the RECEIVER trial are encouraging, with sustained patient use of the application, and associated positive impact on patient outcomes. Older age of patients lacking access to digital technology is notable. To support post COVID-19 NHS recovery we have scaled-up the digital self-management service, offering this to all COPD patients in NHS GG&C. RECEIVER trial dataset will be combined with large NHS GG&C SafeHaven historical cohort and the scale-up patient data for machine-learning analyses. We aim to train, validate and operationalise prediction models for 12-month mortality, 3-month re-admission and 72-hour exacerbation risk.

ASTERIX: Adaptive stratification of COVID19 to facilitate endotype-directed intervention studies

Introduction and Objectives: Severe coronavirus 19 disease (COVID 19) has rapidly emerged as a global health threat and, despite considerable advances, outcomes remain poor in many patients. Published data infers considerable heterogeneity, with 80% suffering minimal symptoms but a minority developing life-threatening disease. COVID 19 trials to-date have been necessarily broad but the emergence of established therapies (e.g. Dexamethasone) and distinct phenotypes (e.g. immune activated, prothrombotic) suggests that early stratification to licensed or trial agents might result in improved outcomes. The ASTERIX study aims to define disease endotypes, based on baseline biological signatures associated with COVID-19 pneumonia, development of respiratory failure and death, which could be targeted in future trials. Methods: >6,000 samples of blood, urine and respiratory secretions were collected and banked during the first wave of the COVID 19 pandemic in Glasgow. The cohort is organised into Tiers 0, 1 & 2 with each tier having an increasing number of samples available for downstream translational research. All tiers have the same associated comprehensive clinical data including comorbidity, ethnicity, blood results, imaging, prescription data and outcomes, including critical care support and survival. Results: Tier 0 contains 1,512 cases, Tier 1 (defined by having at least one surplus sample banked for downstream assays) contains ~1000 cases. Tier 2 (defined as having matched samples of serum, plasma and a buffy coat) contains 421 cases. Sample collation and data analysis is ongoing but preliminary review indicates a mortality rate of 29%, which is consistent with that reported in UK-wide COVID 19 series. The project team have made extensive links with collaborators and a scientific review board has been convened. The following projects are at various stages of approval and delivery: (1) Host Epigenomics (2) Host Proteomics (3) Host Metabolomics (4) miRNA Outcome Signatures (5) Host Respiratory Microbiome (6) COVID 19 Coagulopathy. Conclusions: Data and banked samples will be used to develop endotypes (biological signatures derived from statistical models) associated with progression to key clinical outcomes. This information will be used to identify high-risk cohorts that could be targeted in future studies testing suitable interventions, as directed by the content of each signature.

P34 Mitigating the COVID-19 impact on COPD care: rapid development of remote recruitment processes to a digital self-management service

BackgroundShielding requirements and interruptions to routine care resulting from the COVID-19 pandemic present substantial risks to COPD patient’s outcomes. Adverse impacts including increased winter admissions, nosocomial infection risks, impaired quality of life and high service costs are anticipated. A digital remote-management service can potentially mitigate these by supporting routine COPD care and self-management.In the InnovateUK funded ‘DYNAMIC’ project we’ve developed a fully integrated COPD digital support service. Patient-held smartphone progressive web application prompts daily patient reported outcome (PRO) completion with red-amber-green self-management advice, rescue medication and linked resources including inhaler, exercise and dyspnoea management videos. Secure asynchronous patient-clinician messaging and EHR-integrated clinician dashboard with PRO data visualisation and structured clinical data enables self-management support and provides a new channel for scheduled care.MethodsAs COVID-19 pandemic emerged we paused the RECEIVER trial (NCT04240353) and undertook rapid co-design, governance approval and deployment of an SMS invite and web-based registration system https://support.nhscopd.scot This provides access and remote setup to the DYNAMIC COPD service for all patients in our organisation. Website tracking analytics support service evaluation and process optimisation. Feedback was obtained from 57 patients who registered an interest via website but didn’t progress to setup in the service.Results2373 high-risk COPD patients were identified from secondary care datasets. 797 patients did not have a known mobile phone number. 1576 SMS invites were sent in batches May – July 2020. 599 unique visits to support site (38%) with 195 completed applications (12.3%) to the digital service were obtained. As of August 2020, 112 patients (7.1%) have completed service setup, which includes clinician virtual review with optimisation of COPD interventions, self-management planning and COPD MDT input as required. Patient app usage in this scale-up cohort matches positive experience from the RECEIVER trial. Qualitative data highlighted requirement for increased information about the service at invitation, which has been addressed.ConclusionsNovel processes for digital recruitment and remote setup have widened access to NHS Scotland COPD support service. This is a model strategy for service implementation and evaluation. Further cycles of mail-based and SMS invites and media awareness campaign will follow.