Deep Neural Network-Based Respiratory Pathology Classification Using Cough Sounds.

Intelligent systems are transforming the world, as well as our healthcare system. We propose a deep learning-based cough sound classification model that can distinguish between children with healthy versus pathological coughs such as asthma, upper respiratory tract infection (URTI), and lower respiratory tract infection (LRTI). To train a deep neural network model, we collected a new dataset of cough sounds, labelled with a clinician's diagnosis. The chosen model is a bidirectional long-short-term memory network (BiLSTM) based on Mel-Frequency Cepstral Coefficients (MFCCs) features. The resulting trained model when trained for classifying two classes of coughs-healthy or pathology (in general or belonging to a specific respiratory pathology)-reaches accuracy exceeding 84% when classifying the cough to the label provided by the physicians' diagnosis. To classify the subject's respiratory pathology condition, results of multiple cough epochs per subject were combined. The resulting prediction accuracy exceeds 91% for all three respiratory pathologies. However, when the model is trained to classify and discriminate among four classes of coughs, overall accuracy dropped: one class of pathological coughs is often misclassified as the other. However, if one considers the healthy cough classified as healthy and pathological cough classified to have some kind of pathology, then the overall accuracy of the four-class model is above 84%. A longitudinal study of MFCC feature space when comparing pathological and recovered coughs collected from the same subjects revealed the fact that pathological coughs, irrespective of the underlying conditions, occupy the same feature space making it harder to differentiate only using MFCC features.

Impact of COVID-19 on pediatric emergencies and hospitalizations in Singapore.

BACKGROUND: Coronavirus disease 2019 (COVID-19) has impacted the provision of health services in all specialties. We aim to study the impact of COVID-19 on the utilization of pediatric hospital services including emergency department (ED) attendances, hospitalizations, diagnostic categories and resource utilization in Singapore. METHODS: We performed a retrospective review of ED attendances and hospital admissions among children < 18 years old from January 1st to August 8th 2020 in a major pediatric hospital in Singapore. Data were analyzed in the following time periods: Pre-lockdown (divided by the change in Disease Outbreak Response System Condition (DORSCON) level), during-lockdown and post-lockdown. We presented the data using proportions and percentage change in mean counts per day with the corresponding 95% confidence intervals (CIs). RESULTS: We attended to 58,367 children with a mean age of 5.1 years (standard deviation, SD 4.6). The mean ED attendance decreased by 331 children/day during lockdown compared to baseline (p < 0.001), attributed largely to a drop in respiratory (% change - 87.9, 95% CI - 89.3 to - 86.3, p < 0.001) and gastrointestinal infections (% change - 72.4, 95%CI - 75.9 to - 68.4, p < 0.001). Trauma-related diagnoses decreased at a slower rate across the same periods (% change - 40.0, 95%CI - 44.3 to - 35.3, p < 0.001). We saw 226 children with child abuse, with a greater proportion of total attendance seen post-lockdown (79, 0.6%) compared to baseline (36, 0.2%) (p < 0.001). In terms of ED resource utilization, there was a decrease in the overall mean number of procedures performed per day during the lockdown compared to baseline, driven largely by a reduction in blood investigations (% change - 73.9, 95%CI - 75.9 to - 71.7, p < 0.001). CONCLUSIONS: We highlighted a significant decrease in infection-related presentations likely attributed to the lockdown and showed that the relative proportion of trauma-related attendances increased. By describing the impact of COVID-19 on health services, we report important trends that may provide guidance when planning resources for future pandemics.