| Author: | Liu, Yanwen |
| Title: | Impact of COVID-19, influenza, and air pollution on mortality |
| Degree: | Ph.D. |
| Year: | 2025 |
| Department: | Department of Applied Mathematics |
| Pages: | xii, 114 pages : color illustrations |
| Language: | English |
| Abstract: | Associations between various influenza strains (H3N2, H1N1, and influenza B), different air pollutants (PM2.5, PM10, ozone, etc.) and mortality have been a major concern in the past decade. The worldwide outbreak of COVID-19 in early 2020 made it urgent to explore the mortality attributed to the disease. Meanwhile, the influences of influenza and air pollution on mortality keep changing during the pandemic period and the post pandemic era. This thesis will study the impact of COVID-19, influenza, and air pollution on mortality by deploying widely adopted statistical methods, such as generalized additive models, maximum likelihood estimation and generalized linear model, along with state-of-the-art data science methods such as eXtreme Gradient Boosting and neural network. Chapter 2 used death certificates dataset provided by Peru Ministry of Health and reported the heterogeneity of the median age of all-cause mortality and the daily pattern of cause-specific mortality directly and indirectly related to COVID in Peru. An assumption that most of the indirectly excess death in Peru were primarily caused by circulatory system diseases was raised. In Chapter 3, exposure history and symptom onset date for patients infected by different SARS-CoV-2 variants was collected. Maximum likelihood estimators of mean and standard deviation of the distributions of incubation periods were calculated. For the Delta variant, the mean incubation period was 6.54 days (95% CI: 5.28 – 7.68), while for the Omicron variant, it was significantly shorter at 3.43 days (95% CI: 2.47 – 3.76). Thus, a 7-day quarantine may be more effective during Omicron predominance. Moreover, a metric called daily exceedance concentration hours (DECH) was reinvestigated in Chapter 4. Generalized additive models with quasi-Poisson distribution links were fitted to calculate relative effects of DECH levels on mortality risk across the disease groups. A fairly consistent size of the association between DECH levels and mortality risk was found with a less significance during the COVID pandemic period. Futhermore, eXtreme Gradient Boosting was adopted in Chapter 5 to estimate the excess mortality attributable to air pollutants and influenza, during the pre-pandemic and pandemic period in Hong Kong. In the first two years of the COVID-19 pandemic, 8,762 (95% confidence interval, 7,503 – 9,993), and 12,496 (11,718 – 13,332) excess all-cause deaths were estimated. A notable shift in disease burden attributable to influenza and air pollutants was observed in the pandemic period, suggesting that mortality directly and indirectly caused by COVID-19 shall be considered when assessing the global and regional burden of the COVID-19 pandemic. Finally, in Chapter 6, suggestions for future research on influenza forecasting were discussed. Human Influenza hemagglutinin (HA) and Neuraminidase (NA) A(H3N2) sequences and their association with influenza spread were reviewed. Deep Learning Framework was introduced to handle both sequencing and time series data. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/14068