Analyzing the Variations in Emergency Department Boarding and Testing the Transferability of Forecasting Models across COVID-19 Pandemic Waves in Hong Kong: Hybrid CNN-LSTM approach to quantifying building-level socioecological risk (preprint)

Emergency department's (ED) boarding (defined as ED waiting time greater than four hours) has been linked to poor patient outcomes and health system performance. Yet, effective forecasting models is rare before COVID-19, lacking during the peri-COVID era. Here, a hybrid convolutional neural network (CNN)-Long short-term memory (LSTM) model was applied to public-domain data sourced from Hong Kong's Hospital Authority, Department of Health, and Housing Authority. In addition, we sought to identify the phase of the COVID-19 pandemic that most significantly perturbed our complex adaptive healthcare system, thereby revealing a stable pattern of interconnectedness among its components, using deep transfer learning methodology. Our result shows that 1) the greatest proportion of days with ED boarding was found between waves four and five; 2) the best-performing model for forecasting ED boarding was observed between waves four and five, which was based on features representing time-invariant residential buildings' built environment and sociodemographic profiles and the historical time series of ED boarding and case counts, compared to during the waves when best-performing forecasting is based on time-series features alone; and 3) when the model built from the period between waves four and five was applied to data from other waves via deep transfer learning, the transferred model enhanced the performance of indigenous models.

Comparative analysis of symptom profile and risk of death associated with infection by SARS-CoV-2 and its variants in Hong Kong (preprint)

Introduction: The recurrent multi-wave nature of COVID-19 necessitates updating its symptomatology. Before the omicron era, Hong Kong was relatively unscathed and had a low vaccine uptake rate among the old-old, giving us an opportunity to study the intrinsic severity of SARS-CoV-2 variants. A comparison of symptom patterns across variants and vaccination status in Hong Kong has yet to be undertaken. The intrinsic severity of variants and symptoms predictive of severe outcomes are also understudied as COVID-19 evolves. We therefore aim to characterize the effect of variants on symptom presentation, identify the symptoms predictive and protective of death, and quantify the effect of vaccination on symptom development. Methods: With the COVID-19 case series in Hong Kong from inception to 25 August 2022, an iterative multi-tier text-matching algorithm was developed to identify symptoms from free text. Cases were fully vaccinated if they completed two doses. Multivariate regression was used to measure associations between variants, symptom development, death and vaccination status. A least absolute shrinkage and selection operator technique was used to identify a parsimonious set of symptoms jointly associated with death. Results: Overall, 70.9% (54450/76762) of cases were symptomatic. We identified a wide spectrum of symptoms (n=102), with cough, fever, runny nose and sore throat being the most common (8.16-47.0%). Intrinsically, the wild-type and delta variant caused similar symptoms, with runny nose, sore throat, itchy throat and headache more frequent in the delta cohort; whereas symptoms were heterogeneous between the wild-type and omicron variant, with seven symptoms (fatigue, fever, chest pain, runny nose, sputum production, nausea/vomiting and sore throat) more frequent in the omicron cohort. With full vaccination, omicron was still more likely than delta to cause fever. Fever, blocked nose and shortness of breath were robustly jointly predictive of death as the virus evolved. Number of vaccine doses required for reduction in occurrence varied by symptoms. Discussion: This is the first large-scale study to evaluate the changing symptomatology by COVID-19 variants and vaccination status using free-text reporting by patients. We substantiate existing findings that omicron has a different clinical presentation compared to previous variants. Syndromic surveillance can be bettered with reduced reliance on symptom-based case identification, increased weighing on symptoms robustly predictive of mortality in outcome prediction, strengthened infection control in care homes through universal individual-based risk assessment to enable early risk stratification, adjusting the stockpile of medicine to tally with the changing symptom profiles across vaccine doses, and incorporating free-text symptom reporting by patients.

Quantifying the Risk of General Health and Early COVID-19 Spread in Residential Buildings with Deep Learning and Expert-augmented Machine Learning (preprint)

Buildings' built environment has been linked to their occupants' health. It remains unclear whether those elements that predisposed its residents to poor general health before the two SARS pandemics also put residents at risk of contracting COVID-19 during early outbreaks. Relevant research to uncover the associations is essential, but there lacks a systematic examination of the relative contributions of different elements in one's built environment and other non-environmental factors, singly or jointly. Hence, the current study developed a deep-learning approach with multiple input channels to capture the hierarchical relationships among an individual's socioecology's demographical, medical, behavioral, psychosocial, and built-environment levels. Our findings supported that 1) deep-learning models whose inputs were structured according to the hierarchy of one's socioecology outperformed plain models with one-layered input in predicting one's general health outcomes, with the model whose hierarchically structured input layers included one's built environment performed best; 2) built-environment features were more important to general health compared to features of one's sociodemographic and their health-related quality of life, behaviors, and service utilization; 3) a composite score representing built-environment features' statistical importance to general health significantly predicted building-level COVID-19 case counts; and 4) building configurations derived from the expert-augmented learning of granular built-environment features that were of high importance to the general health were also linked to building-level COVID-19 case counts of external samples. Specific built environments put residents at risk for poor general health and COVID-19 infections. Our machine-learning approach can benefit future quantitative research on sick buildings, health surveillance, and housing design.

Age-related seroprevalence trajectories of seasonal coronaviruses in children (preprint)

Four seasonal coronaviruses, including HCoV-NL63 and HCoV-229E, HCoV-OC43 and HCoV-HKU1 cause approximately 15–30% of common colds in adults. However, the frequency and timing of early infection with four seasonal coronaviruses in the infant are still not well studied. Here, we evaluated the serological response to four seasonal coronaviruses in 1886 children under 18-year-old to construct the viral infection rates. The antibody levels were also determined from the plasma samples of 485 pairs postpartum women and their newborn babies. This passive immunity waned at one year after birth and the resurgence of the IgGs were found thereafter with the increase of the age. Taken together, our results show the age-related seroprevalence trajectories of seasonal coronaviruses in children and provide useful information for deciding vaccine strategy for coronaviruses in the future.

Social Contact Patterns and Implications for Infectious Disease Transmission: A Systematic Review and Meta-Analysis of Contact Surveys (preprint)

Background Transmission of respiratory pathogens such as SARS-CoV-2 depends on patterns of contact and mixing across populations. Understanding this is crucial to predict pathogen spread and the effectiveness of control efforts. Most analyses of contact patterns to date have focussed on high-income settings. Methods Here, we conduct a systematic review and individual-participant meta-analysis of surveys carried out in low- and middle-income countries and compare patterns of contact in these settings to surveys previously carried out in high-income countries. Using individual-level data from 28,503 participants and 413,069 contacts across 27 surveys we explored how contact characteristics (number, location, duration and whether physical) vary across income settings. Results Contact rates declined with age in high- and upper-middle-income settings, but not in low-income settings, where adults aged 65+ made similar numbers of contacts as younger individuals and mixed with all age-groups. Across all settings, increasing household size was a key determinant of contact frequency and characteristics, but low-income settings were characterised by the largest, most intergenerational households. A higher proportion of contacts were made at home in low-income settings, and work/school contacts were more frequent in high-income strata. We also observed contrasting effects of gender across income-strata on the frequency, duration and type of contacts individuals made. Conclusions These differences in contact patterns between settings have material consequences for both spread of respiratory pathogens, as well as the effectiveness of different non-pharmaceutical interventions. Funding This work is primarily being funded by joint Centre funding from the UK Medical Research Council and DFID (MR/R015600/1).

Public perceptions and preventive behaviours during the early phase of the COVID-19 pandemic: a comparative study between Hong Kong and the United Kingdom (preprint)

BackgroundIn the absence of treatments and vaccines, the mitigation of COVID-19 relies on population engagement in non-pharmaceutical interventions, which is driven by their risk perception, anxiety level and knowledge. There may also be regional discrepancies in these drivers due to different historical exposure to disease outbreaks, government responses and cultures. As such, this study compared psycho-behavioral responses in two regions during the early phase of the pandemic. MethodsComparable cross-sectional surveys were administered among adults in Hong Kong (HK) and the United Kingdom (UK) during the early phase of each respective epidemic. Explanatory variables included demographics, risk perception and knowledge of COVID-19, anxiety level and preventive behaviors. Responses were weighted according to census data. Logistic regression models, including interaction terms to quantify regional differences, were used to assess the association between explanatory variables and the adoption of social-distancing measures. ResultsData of 3431 complete responses (HK:1663; UK:1768) were analysed. Perceived severity differed by region (HK: 97.5%; UK: 20.7%). A large proportion of respondents were abnormally/borderline anxious (HK:64.8%; UK:45.9%) and regarded direct contact with infected individuals as the transmission route of COVID-19 (HK:94.0-98.5%; UK:69.2-93.5%), with HK identifying additional routes. HK reported high levels of adoption of social-distancing (HK:32.4-93.7%; UK:17.6-59.0%) and mask-wearing (HK:98.8%; UK:3.1%). The impact of perceived severity and perceived ease of transmission on the adoption of social-distancing varied by region. In HK, they had no impact, whereas in the UK, those who perceived severity as "high" were more likely to adopt social-distancing (aOR:1.58-3.01), and those who perceived transmission as "easy" were prone to both general social-distancing (aOR:2.00, 95% CI:1.57, 2.55) and contact avoidance (aOR:1.80, 95% CI: 1.41, 2.30). The impact of anxiety on adopting social-distancing did not vary by region. DiscussionThese results suggest that health officials should ascertain and consider baseline levels of risk perception and knowledge in the populations, as well as prior sensitisation to infectious disease outbreaks, during the development of mitigation strategies. Risk communication should be done through suitable media channels - and trust should be maintained - while early intervention remains the cornerstone of effective outbreak response.

Are we ready when COVID-19 vaccine is available? Study on nurses' vaccine hesitancy in Hong Kong (preprint)

Introduction: Nurses are considered a trustworthy source of vaccine-related information to build public confidence in vaccination. This study estimated nurses' influenza vaccine uptake and intention to receive COVID-19 vaccine when available, and examined the corresponding psychological antecedents. Methods : A cross-sectional online survey among nurses was conducted during the main COVID-19 outbreak in Hong Kong between mid-March and late April 2020. Demographics, influenza vaccination, intention to have COVID-19 vaccine, the 5C vaccine hesitancy components (i.e., confidence, complacency, constraints, calculation, and collective responsibility), work stress and COVID-related work demands (i.e., insufficient supply of personal protective equipment, involvement in isolation rooms, and unfavorable attitudes towards workplace infection control policies) were reported. Results: The influenza vaccination coverage and the proportion intending to take COVID-19 vaccine were 49% and 63%, respectively, among 1205 eligible nurses. Influenza vaccine uptake was associated with working in public hospitals and all 5C constructs, whereas stronger COVID-19 vaccination intention was associated with younger age, more confidence, less complacency and more collective responsibility towards the vaccine. COVID-19-related demands were associated with greater work stress, and hence stronger COVID-19 vaccination intention. Conclusion: Vaccine uptake/intention was well predicted by the 5C constructs. With less work stress among nurses in the post-pandemic period, the intention to take COVID-19 vaccine will likely drop. The 5C constructs should be infused in vaccination campaigns. While a COVID-19 vaccine could be ready soon, communities are not ready to accept it. More research work is needed to boost the uptake.

Non-pharmaceutical behavioural measures for droplet-borne biological hazards prevention: Health-EDRM for COVID-19 (SARS-CoV-2) pandemic (preprint)

Introduction: Non-pharmaceutical interventions to facilitate response to the COVID-19 pandemic, a disease caused by novel coronavirus SARS-CoV-2, are urgently needed. Using the WHO health emergency and disaster risk management (health-EDRM) framework, behavioural measures for droplet-borne communicable disease, with their enabling and limiting factors at various implementation levels were evaluated. Sources of data: Keyword search was conducted in PubMed, Google Scholar, Embase, Medline, Science Direct, WHO and CDC online publication database. Using OCEBM as review criteria, 105 English-language articles, with ten bottom-up, non-pharmaceutical prevention measures, published between January 2000 and May 2020 were identified and examined. Areas of Agreement: Evidence-guided behavioural measures against COVID-19 transmission for global at-risk communities are identified. Area of Concern: Strong evidence-based systematic behavioural studies for COVID-19 prevention are lacking. Growing points: Very limited research publications are available for non-pharmaceutical interventions to facilitate pandemic response. Areas timely for research: Research with strong implementation feasibility that targets resource-poor settings with low baseline Health-EDRM capacity is urgently need.

Community responses during the early phase of the COVID-19 epidemic in Hong Kong: risk perception, information exposure and preventive measures (preprint)

Background: Community responses are important for outbreak management during the early phase when non-pharmaceutical interventions are the major preventive options. Therefore, this study aims to examine the psychological and behavioral responses of the community during the early phase of the COVID-19 epidemic in Hong Kong. Method: A cross-sectional online survey was launched within 36 hours after confirmed COVID-19 cases were first reported. Councilors of all 452 district council constituency areas were approached for survey dissemination. Respondent demographics, anxiety level, risk perception, sources to retrieve COVID-19 information, actual adoption and perceived efficacy of precautionary measures were collected. Result: Analysis from 1715 complete responses indicated high perceived susceptibility (89%) and high perceived severity (97%). Most respondents were worried about COVID-19 (97%), and had their daily routines disrupted (slightly/greatly: 98%). The anxiety level, measured by the Hospital Anxiety and Depression Scale, was borderline abnormal (9.01). Nearly all respondents were alert to the disease progression (99.5%). The most trusted information sources were doctors (84%), followed by broadcast (57%) and newspaper (54%), but they were not common information sources (doctor: 5%; broadcast: 34%; newspaper: 40%). Only 16% respondents found official websites reliable. Enhanced personal hygiene practices and travel avoidance to China were frequently adopted (>77%) and considered effective (>90%). The adoption of social-distancing measures was lower (39%-88%), and their drivers for greater adoption include: being female (adjusted odds ratio [aOR]:1.27), living in the New Territories (aOR:1.32-1.55), perceived as having good understanding of COVID-19 (aOR:1.84) and being more anxious (aOR:1.07). Discussion: Risk perception towards COVID-19 in the community was high. Most respondents are alert to the disease progression, and adopt self-protective measures. This study contributes by examining the psycho-behavioral responses of hosts, in addition to the largely studied mechanistic aspects, during the early phase of the current COVID-19 epidemic. The timely psychological and behavioral assessment of the community is useful to inform subsequent interventions and risk communication strategies as the epidemic progresses.