Nicholas Osborne

Spatial-Temporal data to better predict aeroallergen hospital admissions: use of spatial point process models

A high incidence of asthma, COPD and allergic rhinitis in the Queensland community has the risk of putting high pressure on the Queensland health system on days when exposure to environmental factors is high. These include viral infection, days with high PM2.5 and high pollen days (>50 particles per m3). In the 2022/23 financial year, there were more than 7,300 hospitalisations in Queensland with a principal diagnosis of asthma.

This project is multidisciplinary in nature and covers the Queensland Health research priorities of air quality and health and epidemiological data analysis. In collaboration with the University of Melbourne and AirHealth (CI Osborne is member of the scientific coordination committee) we will use data collected over the last 10 year to make better predictions on air quality.

Recently the University of Queensland has joined the AirHealth network. This Australia wide network of pollen monitoring provides a service including providing pollen forecast to enable patients with asthma, COPD and rhinitis to better manage their conditions via knowledge of environmental.

Using data collected in Melbourne from 2013 to 2024 we will assess the relationship between pollen levels and symptoms of allergy. Knowledge of this relationship will be used along side newly collected pollen data at the pollen monitoring station at Herston (UQ Health Sciences Building).

We have over 411,000 geolocated responses from AirHealth data app that records symptoms of those subscribed to the Australia-wide service informing. We intend to link these to daily pollen measures in Melbourne, Sydney, Brisbane and Canberra. Initial results have been published (Silver, J. D. et al. Crowd-sourced allergic rhinitis symptom data: The influence of environmental and demographic factors. Science of The Total Environment 705, 135147 (2020).2 and Silver, J. D. et al. Using crowd-sourced allergic rhinitis symptom data to improve grass pollen forecasts and predict individual symptoms. Science of The Total Environment 720, 137351 (2020)).

Potential question to probe this valuable data set include how grass pollen exposure influences symptoms and hospital admittance, how grass pollen differs from other pollens, how app users respond differently in drought and flood years and how different regions in Queensland and Australia have different responses to high pollen exposure.

We will use different spatial point process models and techniques that have been developed in the past decades to facilitate the statistical analysis of spatial point patterns. We will use deep neural networks in the statistical analysis of highly multivariate spatial point patterns and provide a new strategy for building spatio-temporal point processes using variational autoencoder generative neural network to examine the relationship between disease and health outcome in using spatial-temporal data sets.

Heat events and health burden in the tropics under the global warming

Project description

Heat events are one of the direct impacts of climate change impacting public health are increasing at an alarming rate under the global warming. Physiological response to heat events depends not only on temperature, but also on humidity, wind, and radiation, and on the clothing pattern and working condition of the person exposed to heat. Therefore, the mortality and morbidity burden of heat events vary across the tropics depending on the local and socio-demographic conditions. Most current knowledge on heat waves and heat wave warnings are derived from data in temperate zones.

We plan to model heat exchange between human body and the surrounding environment and how that is influenced by climatic factors. This will enable to understand the potential benefits of various intervention. We will examine health and climate data from India and Australia to develop heat mortality model. Further we expect to project the health burden attributable to heat events under various climate change scenarios for the two countries.

A range of questions will be asked around the different heat type in the tropics/subtropics. Smaller diurnal shifts, longer strings of days in the 98th percentile, high night-time temperatures and high humidity have been flagged as potential pathways that the heat-health nexus differs from the temperate zones. These initial studies will use extant data from Bureau of Meteorology and Health Services. This will follow on from conclusions in recent UQ paper Impact of low-intensity heat events on mortality and morbidity in regions with hot, humid summers: a scoping literature review.

Pollen and its contribution to respiratory outcomes in subtropical regions

Air quality is a strong predictor of respiratory health. In comparison to traditional sources like industry, transport, open burning, bushfires, power plants, the contribution of plant pollen is not so well studied in the sub-tropical/arid climate. Residents in the subtropical regions show higher allergic sensitivity and subfamily-specific immunoreactivity with pollen of Chloridoideae and Panicoideae grasses, compared with temperate grass pollen. Therefore, knowledge of species and amount of pollen in air column is essential to preparing more accurate predictions of episodes when respiratory disease burden would be high and focusing public health efforts.

We plan to examine what biologics are in the air column in Delhi and Brisbane as currently there is very little knowledge. Aerial dispersion of pollen gives an important snapshot of the biodiversity of the surrounding environment. New environmental DNA (eDNA) techniques allow samples of plant pollens and fungal spores to be collected and analysed using stored data (“DNA barcode” library) or other sequencing techniques allowing sapling of all genes in all organisms present in a complex sample via high-throughput molecular methods. This information, when combined with the next generation air quality models provides a variety of tools to examine the release, dispersion and transformation of bioaerosols and how it affects the environment. Further we expect to analyse the respiratory morbidity data from the two cities to understand the health risk of exposure to pollen.

We will collect pollen at different times of year. Using eDNA techniques and light microscopy we will ascertain diurnal and seasonal fluctuations in pollen genera and species, and their relationship to human health outcomes.