Search

BACKGROUND Increased workload, complexity of conditions, and delayed supports for COVID-19 control reinforce each other to create an extra burden among Vietnamese health workers during the most severe wave of the pandemic in 2021. OBJECTIVE In this study, we assessed the psychological impact posed by COVID-19 and its associated factors on the healthcare workforce nationwide during the peak of Vietnam’s forth outbreak. METHODS A cross-sectional study was conducted in from October to November 2021 among 2,814 healthcare professionals in 61 provinces of Vietnam. An online questionnaire using PHQ-9, PSS-4 and GAD-7 scales to measure changes in working characteristics, access to healthcare and mental disorders severity was distributed randomly to a subgroup of 503 respondents. RESULTS Nearly half (49.7%) of healthcare workers experienced mild depression symptoms, 34.0% underwent moderate anxiety symptoms, and 49.3% reported high-stress levels. Respondents who were nurse, were female, had monthly income below 5 million VND and had more than 3 days of duty per week scored higher in the anxiety scales. More than half (53.9%) of respondents stated no demand for mental healthcare services, and online platform was the most popular means of accessing mental health information among healthcare workers. CONCLUSIONS Higher levels of mental distress did not come with financial rewards, implying the need for a raise in basic salary as well as compensation and encouragement schemes. To tackle hesitancy in seeking mental help, awareness raising among workers about mental health issues and promotion of mental care through eHealth and mHealth platforms should be emphasized.

To help human markers mark a large number of answers of handwritten mathematical expressions (HMEs), clustering them makes marking more efficient and reliable. Clustering HMEs, however, faces the problem of extracting both localization and classification representation of mathematical symbols for an HME image and defining the distance between two HME images. First, we propose a method based on Convolutional Neural Networks (CNN) to extract the representations for an HME. Symbols in various scales are located and classified by a combination of features from a multi-scale CNN. We use weakly supervised training combined with symbols attention to enhance localization and classification predictions. Second, we propose a multi-level spatial distance between two representations for clustering HMEs. Experiments on CROHME 2016 and CROHME 2019 dataset show the promising results of 0.99 and 0.96 in purity, respectively.