PREDICTION OF MORTALITY USING CLUSTERS BASED ON CLINICAL AND LAB PARAMETERS: DATA FROM 2072 PATIENTS FROM INSPIRE COHORT

BackgroundLupus is a heterogenous diseases which results in significant premature mortality. Most studies have evaluated risk factors for lupus mortality using regression models which considers the phenotype in isolation. Identifying clusters of patients on the other hand may help overcome the limitations of such analyses.ObjectivesThe objectives of this study were to describe the causes of mortality and to analyze survival across clusters based on clinical phenotype and autoantibodies in patients of the Indian SLE Inception cohort for Research (INSPIRE)MethodsOut of all patients, enrolled in the INSPIRE database till March 3st 2022, those who had <10% missing variables in the clustering variables were included in the study. The cause of mortality and duration between the recruitment into the cohort and mortality was calculated. Agglomerative unsupervised hierarchical cluster analysis was performed using 25 variables that define SLE phenotype in clinical practice. The number of clusters were fixed using the elbow and silhouette methods. Survival rates were examined using Cox proportional hazards models: unadjusted, adjusted for age at disease onset, socio-economic status, steroid pulse, CYC, MMF usage and cluster of the patients.ResultsIndian patients with lupus have significant early mortality and the majority of deaths occurs outside the hospital setting.Out of 2211 patients in the cohort, 2072 were included into the analysis. The median (IQR) age of the patients was 26 (20-33) years and 91.7% were females. There were 288 (13.1%) patients with juvenile onset lupus. The median (range) duration of follow up of the patients was 37 (6-42) months. There were 170 deaths, with only 77 deaths occurring in a health care setting. Death within 6 months of enrollment occured in in 80 (47.1%) patients. Majority (n=87) succumbed to disease activity, 23 to infections, 24 to coexisting disease activity and infection and 21 to other causes. Pneumonia was the leading cause of death (n=24). Pneumococcal infection led to death in 11 patients and SARS-COV2 infection in 7 patients. The hierarchical clustering resulted in 4 clusters and the characteristics of these clusters are represented in a heatmap (Figure-1A,B). The mean (95% confidence interval [95% CI] survival was 39.17 (38.45-39.90), 39.52 (38.71-40.34), 37.73 (36.77-38.70) and 35.80 (34.10-37.49) months (p<0.001) in clusters 1, 2, 3 and 4, respectively with an HR (95% CI) of 2.34 (1.56, 3.49) for cluster 4 with cluster 1 as reference(Figure 1C). The adjusted model showed an HR (95%CI) for cluster 4 of 2.22 (1.48, 3.22) with an HR(95%CI) of 1.78 (1.29, 2.45) for low socioeconomic status as opposed to a high socioeconomic status (Table 1).ConclusionIndian patients with lupus have significant early mortality and the majority of deaths occurs outside the hospital setting. Disease activity as determined by the traditional activity measures may not be sufficient to understand the true magnitude of organ involvement resulting in mortality. Clinically relevant clusters can help clinicians identify those at high risk for mortality with greater accuracy.Table 1.Univariate and multivariate Cox regression models predicting mortalityUnivariateMultivariateVariablesHazard ratio (95% Confidence interval)P valueHazard ratio (95% Confidence interval)P valueCluster1Reference-Reference-20.87 (0.57, 1.34)0.5320.89 (0.57, 1.38)0.59831.22 (0.81, 1.84)0.3371.15 (0.76, 1.73)0.51342.34 (1.56, 3.49)<0.0012.22(1.48, 3.22)<0.001Socioeconomic statusLower1.78 (1.29, 2.45)<0.001Pulse steroidYes1.6 (0.99, 2.58)0.051MMFYes0.71 (0.48, 1.05)0.083CYCYes1.42 (0.99, 2.02)0.052Proliferative LNYes0.99 (0.62, 1.56)0.952Date of birth age0.99 (0.98, 1.01)0.657CYC- cyclophosphamide, MMF- Mycophenolate mofetilFigure 1.A. Agglomerative clustering dendrogram depicting the formation of four clusters. B.Heatmap depicting distribution of variables used in clustering C. Kaplan-Meier curve showing the survival function across the 4 clusters[Figure omitted. See PDF]REFERENCES:NIL.Acknowledgements:NIL.Disclosure of InterestsNone eclared.

Analyzing the impact of SARS-CoV-2 variants on respiratory sound signals

The COVID-19 outbreak resulted in multiple waves of infections that have been associated with different SARS-CoV-2 variants. Studies have reported differential impact of the variants on respiratory health of patients. We explore whether acoustic signals, collected from COVID-19 subjects, show computationally distinguishable acoustic patterns suggesting a possibility to predict the underlying virus variant. We analyze the Coswara dataset which is collected from three subject pools, namely, i) healthy, ii) COVID-19 subjects recorded during the delta variant dominant period, and iii) data from COVID-19 subjects recorded during the omicron surge. Our findings suggest that multiple sound categories, such as cough, breathing, and speech, indicate significant acoustic feature differences when comparing COVID-19 subjects with omicron and delta variants. The classification areas-under-the-curve are significantly above chance for differentiating subjects infected by omicron from those infected by delta. Using a score fusion from multiple sound categories, we obtained an area-under-the-curve of 89% and 52.4% sensitivity at 95% specificity. Additionally, a hierarchical three class approach was used to classify the acoustic data into healthy and COVID-19 positive, and further COVID-19 subjects into delta and omicron variants providing high level of 3-class classification accuracy. These results suggest new ways for designing sound based COVID-19 diagnosis approaches. Copyright © 2022 ISCA.

Interpretable Acoustic Representation Learning on Breathing and Speech Signals for COVID-19 Detection

In this paper, we describe an approach for representation learning of audio signals for the task of COVID-19 detection. The raw audio samples are processed with a bank of 1-D convolutional filters that are parameterized as cosine modulated Gaussian functions. The choice of these kernels allows the interpretation of the filterbanks as smooth band-pass filters. The filtered outputs are pooled, log-compressed and used in a self-attention based relevance weighting mechanism. The relevance weighting emphasizes the key regions of the time-frequency decomposition that are important for the downstream task. The subsequent layers of the model consist of a recurrent architecture and the models are trained for a COVID-19 detection task. In our experiments on the Coswara data set, we show that the proposed model achieves significant performance improvements over the baseline system as well as other representation learning approaches. Further, the approach proposed is shown to be uniformly applicable for speech and breathing signals and for transfer learning from a larger data set. Copyright © 2022 ISCA.

Coswara: A website application enabling COVID-19 screening by analysing respiratory sound samples and health symptoms

The COVID-19 pandemic has accelerated research on design of alternative, quick and effective COVID-19 diagnosis approaches. In this paper, we describe the Coswara tool, a website application designed to enable COVID-19 detection by analysing respiratory sound samples and health symptoms. A user using this service can log into a website using any device connected to the internet, provide there current health symptom information and record few sound sampled corresponding to breathing, cough, and speech. Within a minute of analysis of this information on a cloud server the website tool will output a COVID-19 probability score to the user. As the COVID-19 pandemic continues to demand massive and scalable population level testing, we hypothesize that the proposed tool provides a potential solution towards this. Copyright © 2022 ISCA.

DiCOVA Challenge: Dataset, task, and baseline system for COVID-19 diagnosis using acoustics

The DiCOVA challenge aims at accelerating research in diagnosing COVID-19 using acoustics (DiCOVA), a topic at the intersection of speech and audio processing, respiratory health diagnosis, and machine learning. This challenge is an open call for researchers to analyze a dataset of sound recordings, collected from COVID-19 infected and non-COVID-19 individuals, for a two-class classification. These recordings were collected via crowdsourcing from multiple countries, through a website application. The challenge features two tracks, one focusing on cough sounds, and the other on using a collection of breath, sustained vowel phonation, and number counting speech recordings. In this paper, we introduce the challenge and provide a detailed description of the task, and present a baseline system for the task.

THE SECOND DICOVA CHALLENGE: DATASET AND PERFORMANCE ANALYSIS FOR DIAGNOSIS OF COVID-19 USING ACOUSTICS

The Second Diagnosis of COVID-19 using Acoustics (DiCOVA) Challenge aimed at accelerating the research in acoustics based detection of COVID-19, a topic at the intersection of acoustics, signal processing, machine learning, and healthcare. This paper presents the details of the challenge, which was an open call for researchers to analyze a dataset of audio recordings consisting of breathing, cough and speech signals. This data was collected from individuals with and without COVID-19 infection, and the task in the challenge was a two-class classification. The development set audio recordings were collected from 965 (172 COVID-19 positive) individuals, while the evaluation set contained data from 471 individuals (71 COVID-19 positive). The challenge featured four tracks, one associated with each sound category of cough, speech and breathing, and a fourth fusion track. A baseline system was also released to benchmark the participants. In this paper, we present an overview of the challenge, the rationale for the data collection and the baseline system. Further, a performance analysis for the systems submitted by the 21 participating teams in the leaderboard is also presented. © 2022 IEEE

Adapting undergraduate paediatric medical education to the challenges of COVID-19 pandemic: perspective of NUS Medicine

COVID-19 significantly impacted the teaching-learning-assessment activities in many medical schools. In this article, we discuss the impact of COVID-19 on the Yong Loo Lin School of Medicine, National University of Singapore, focusing on paediatric training and the adaptations of the system and the people. The school developed strategies to promptly disseminate information and safety measures to protect all its staff and students. By leveraging on the school’s infrastructure for technology-enabled learning, good-quality medical training and reliable assessments were able to be carried out swiftly. The paediatric curriculum was crafted based on these principles, and it provided distance-based learning with live and interactive sessions to teach core clinical skills. The faculty also tapped on standardised patients to provide consistent and life-like scenarios. Measures were implemented to minimise challenges with technology-enabled learning. Collectively, efforts from the staff, support from the leadership and students’ adaptations tremendously helped to ease the transition.

Design, development, and evaluation of rotary drum dryer for turmeric rhizomes (Curcuma longa L.)

Turmeric is a challenging crop to dry because of its heat sensitive volatile content and browning characteristics. In this study, a pilot scale biomass fired rotary drum dryer was developed for drying turmeric rhizomes. The drying experiments were conducted at 50, 60, and 70 degrees C air temperatures at an air velocity of 2 and 3 m/s and the drum speeds of 6 and 9 rpm. The turmeric rhizomes dried at a faster rate at the operating conditions of 70 degrees C air temperature, 3 m/s air velocity and 9 rpm drum rotational speed. Among the evaluated thin layer drying models, Page model comparatively gave higher R-2 values (0.998), lower sum of square error (0.001), and root mean square error (0.01) values. The effective moisture diffusivity of turmeric in the rotary dryer ranged from 0.33 to 0.5 x 10(-10) m(2) s(-1). The calculated activation energy of turmeric rhizomes was 19,338 kJ/kg mol(-1). The higher curcumin, oleoresin, starch, and color values such as lightness, redness, and yellowness value of 4.92%, 16.5%, 58.09%, and 59.87, 28.43, and 74.21, respectively, was observed at 50 degrees C, 9 rpm drum rotational speed and 3 m/s air velocity. The results indicated that the color of turmeric could be retained by drying at 50 degrees C. Practical Applications The present scenario, turmeric has been considered as an important crop against COVID-19 due to its antiviral properties and also food applications such as natural coloring agent and flavor enhancer. The unit operations involved in turmeric processing is washing, boiling, drying, polishing, and size reduction. The amount of curcumin content presents in turmeric decides its value. Drying is the time-consuming process compared to other unit operations and has an impact over the quality parameters of turmeric. Sun drying is the main traditional drying method normally practiced by the farmers for bulk drying of turmeric rhizomes. However, the sun-dried turmeric rhizomes lost its product value because of poor quality end product (dark yellowish brown color). This could be due to longer exposure of the product in sun causes change in the volatiles and color in turn degrades the market value and fetches lower price to the farmer. And also decreased drying rate, prolonged drying time, non-uniformity, and lesser energy consumption are the other factors that degrades further the quality of the product. Hence, a drying technology should be identified to overcome the problems faced by the farmers and also to dry the bulk capacities of the turmeric rhizomes at low cost. Rotary dryers are capable of processing larger capacities of various agricultural products having wide range of thermophysical and aerodynamic properties. The current research focused on developing a rotary dryer to reduce the drying time and also to retain the quality that benefits both farmers and industries by providing good quality product.

Investigating feature selection and explainability for COVID-19 diagnostics from cough sounds

In this paper, we propose an approach to automatically classify COVID-19 and non-COVID-19 cough samples based on the combination of both feature engineering and deep learning models. In the feature engineering approach, we develop a support vector machine classifier over high dimensional (6373D) space of acoustic features. In the deep learning-based approach, on the other hand, we apply a convolutional neural network trained on the log-mel spectrograms. These two methodologically diverse models are then combined by fusing the probability scores of the models. The proposed system, which ranked 9th on the 2021 Diagnosing COVID-19 using Acoustics (Di- COVA) challenge leaderboard, obtained an area under the receiver operating characteristic curve (AUC) of 0:81 on the blind test data set, which is a 10:9% absolute improvement compared to the baseline. Moreover, we analyze the explainability of the deep learning-based model when detecting COVID-19 from cough signals. Copyright © 2021 ISCA.

COVID-19 seroprevalence study in asymptomatic healthcare workers, at a tertiary healthcare centre, India

Introduction: The Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2) infection has evolved into a pandemic disease. The present knowledge is mainly based on available numerator data of confirmed positive cases only. The asymptomatic and mildly symptomatic cases are not brought into picture for testing at all, which is a major contributor to the pandemicity and hence creating bias in the documentation and understanding of the disease. The magnitude of the exposure of Healthcare Workers (HCW) and their potential for asymptomatic transmission makes it critical to know the incidence of infection in the healthcare population. Aim: The aim of the study was to evaluate the seroprevalence of Immunoglobulin G (IgG) SARS-CoV-2 among the asymptomatic HCW. Materials and Methods: This was a cross-sectional study conducted during January 2021 and February 2021 in SRM Medical College Hospital and Research Centre (SRM MCH & RC), a tertiary care hospital in Potheri, Chengalpattu district, Tamil Nadu, India. The HCW were asked to complete the standardised questionnaire including the basic information, symptoms of COVID-19 illness and utility of Personal Protective Equipment (PPE) based on World Health Organization (WHO) risk assessment and management of exposure of HCW in the context of COVID-19. They were divided into two groups, the staff who had direct patient exposure as group 1 with 82 participants and staff without direct patient exposure as group 2 with 46 participants. The serodetection of IgG SARS-CoV-2 antibodies was done using the Chemiluminescence Immunoassay (CLIA). The obtained results were statistically analysed with Statistical Package for the Social Sciences (SPSS) 20.0. A chi-square test (χ2) was performed and a p-value less than 0.05 was considered statistically significant. Results: A total of 128 HCW were studied. In group 1, there were 64.1% (n=82) of HCWs and in group 2 there were 35.9% (n=46) of HCWs. There were total of 74 (57.8%) males and 54(42.2%) females. No gender-specific differences were observed. The mean age in group 1 was 28.93 years and group 2 was 32.2 years. The staffs older than 40 years were more commonly affected. Adherence to strict PPE protocol was observed in 92.6% (76/82) in group 1 and 82.6% (38/46) in group 2. The difference between the groups were statistically significant (p=0.025). In this study, though the seroprevalence of COVID-19 infection was 9.8% (n=8) in group 1 and 13% (n=6) in group 2, it was statistically not significant. Conclusion: SARS-CoV2 Serology study helps to identify the asymptomatic (unestimated) cases. The results of the seroprevalence suggest that the strict adherence to PPE protocol helps to prevent COVID-19.

Chest radiographs in Coronavirus Disease 2019 (COVID-19)

OBJECTIVES: To recognize the radiographic patterns of coronavirus disease 2019 (COVID-19) in Malaysia. MATERIALS AND METHODS: Chest radiographs of patients confirmed with COVID-19 in Hospital Tawau, Sabah, Malaysia were retrospectively analyzed by two radiologists. The radiographic pattern, distribution among subgroups and evolution of the disease over time were determined. RESULTS: Among the 82 patients studied, 65 (79.3%) were males. Mean age of our cohorts was 37 ± 15 years. Baseline chest radiographs were abnormal in 37 patients (45.1%). Over half (52.9%) of the symptomatic patients had abnormal baseline radiograph. Among the children, patients with comorbidities, and patients 60 years of age and above, the abnormal radiographs were 14.3%, 71.4% and 69.3% respectively. Ground glass opacities were the commonest abnormal radiographic feature (35.4%), were peripherally located (35.4%) with predilection for the lower zones (29.3%). Most radiographic abnormalities were multifocal (20.7%) and frequently located in the left lung (19.5%). Radiographic recovery was observed in 15 of 18 patients (83%). Computed tomography (CT) scan demonstrated greater extent of the disease than observed in radiographs of the same patient. CONCLUSIONS: COVID-19 pneumonia presented with a specific radiographic pattern in our cohort of patients, comprising of ground glass opacities in peripheral and basilar distribution, affecting a single lung field and was observed in both symptomatic and asymptomatic patients. Chest radiograph is a useful adjunct screening tool, and in combination with clinical and epidemiological assessment may facilitate in early diagnosis of COVID-19 pneumonia.

Paediatric emergency department attendances during COVID-19 and SARS in Singapore

INTRODUCTION: We evaluated the impact of public health measures on paediatric emergency department attendances during the COVID-19 and severe acute respiratory syndrome (SARS) outbreaks in Singapore. METHODS: Between 1 January 2020 and 31 July 2020, we retrospectively reviewed paediatric emergency department attendances and admissions in a tertiary paediatric hospital in Singapore before and after a national lockdown to combat the spread of COVID-19 in Singapore. Hospital attendances and admissions were compared with data from a corresponding period in 2019 (1 January 2019 to 31 July 2019), as well as during and after the SARS outbreak (1 January 2003 to 31 December 2004). RESULTS: Compared with a corresponding non-outbreak period, emergency department attendances decreased in line with nationwide public health measures during the COVID-19 and SARS outbreaks (2020 and 2003 respectively), before increasing gradually following lifting of restrictions, albeit not to recorded levels before these outbreaks. During the COVID-19 outbreak, mean daily attendances decreased by 40%, from 458 per day in January-July 2019, to 274 per day in January-July 2020. The absolute number of hospital inpatient admissions decreased by 37% from January-July 2019 (19,629) to January-July 2020 (12,304). The proportion of emergency department attendances requiring admission remained similar: 20% in January-July 2019 and 21% in January-July 2020. CONCLUSION: Nationwide public health measures in Singapore have had an impact on paediatric emergency department attendances and hospital inpatient admissions. Data from this study could inform planning and resource allocation for emergency departments in Singapore and internationally.

Coswara - A database of breathing, cough, and voice sounds for COVID-19 diagnosis

The COVID-19 pandemic presents global challenges transcending boundaries of country, race, religion, and economy. The current gold standard method for COVID-19 detection is the reverse transcription polymerase chain reaction (RT-PCR) testing. However, this method is expensive, time-consuming, and violates social distancing. Also, as the pandemic is expected to stay for a while, there is a need for an alternate diagnosis tool which overcomes these limitations, and is deployable at a large scale. The prominent symptoms of COVID-19 include cough and breathing difficulties. We foresee that respiratory sounds, when analyzed using machine learning techniques, can provide useful insights, enabling the design of a diagnostic tool. Towards this, the paper presents an early effort in creating (and analyzing) a database, called Coswara, of respiratory sounds, namely, cough, breath, and voice. The sound samples are collected via worldwide crowdsourcing using a website application. The curated dataset is released as open access. As the pandemic is evolving, the data collection and analysis is a work in progress. We believe that insights from analysis of Coswara can be effective in enabling sound based technology solutions for point-of-care diagnosis of respiratory infection, and in the near future this can help to diagnose COVID-19. © 2020 ISCA