ASSESSMENT OF TEMPERATURE EFFECTS ON CHILDHOOD … · 2015. 5. 6. · This thesis identified the...
Transcript of ASSESSMENT OF TEMPERATURE EFFECTS ON CHILDHOOD … · 2015. 5. 6. · This thesis identified the...
ASSESSMENT OF TEMPERATURE EFFECTS ON CHILDHOOD PNEUMONIA AND
DIARRHOEA
Zhiwei Xu BMed, MMed
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
School of Public Health and Social Work
Faculty of Health
Queensland University of Technology
April 2015
i
Keywords
Children
Climate change
Cold spell
Diarrhoea
Geographic co-distribution
Heat wave
Pneumonia
Remote sensing
Temperature
ii
Abstract
Pneumonia and diarrhoea are the two leading killers of children under five years, and
climate change may impact the burden of these two paediatric diseases, even though
the strength of the association between climate change and childhood pneumonia and
diarrhoea remains largely unclear, especially in subtropical regions.
This thesis identified the risk areas of childhood pneumonia and diarrhoea in
Queensland and quantified the effects of temperature on emergency department visits
(EDVs) for childhood pneumonia and diarrhoea. Specifically, three research
questions were answered in this thesis: (I) What are the spatial and temporal patterns
of EDVs for childhood pneumonia and diarrhoea in Queensland, and is there any
geographic co-distribution of these two diseases? (II) What is the relationship
between extreme temperatures and EDVs for childhood pneumonia and diarrhoea?
(III) What is the relationship between temperature variability and EDVs for
childhood pneumonia and diarrhoea?
In a systematic review, I discussed how climatic factors may impact childhood
pneumonia and diarrhoea, and found that the symbolic parameter of climate change,
namely, increasing temperature, is reported as the most likely climatic factor
associated with childhood pneumonia and diarrhoea. I also identified six knowledge
gaps in the existing body of knowledge, two of which are filled by this thesis.
iii
As the spatiotemporal patterns of childhood pneumonia and diarrhoea in Queensland
remain unknown, I attempted to address this issue. The results of this study showed
that childhood pneumonia and diarrhoea mainly distributed in northwest of
Queensland, and Mount Isa had a high-risk cluster where childhood pneumonia and
diarrhoea co-distributed.
To assess the impact of extreme temperatures and prolonged extreme temperatures
(i.e., heat waves and cold spells) on EDVs for childhood pneumonia and diarrhoea in
Brisbane from 2001 to 2010, I conducted two time-series studies, and observed that
both high and low temperatures were associated with increases in EDVs for
childhood pneumonia and diarrhoea. During the study period, there was a decreasing
trend in the high temperature effect on childhood pneumonia, while the low
temperature effect on childhood pneumonia experienced an increasing trend. Heat
waves had significant added effects on childhood pneumonia and diarrhoea, and the
magnitude of these effects increased with intensity and duration. Added effects of
cold spells on childhood pneumonia were also detected.
Data on the effects of temperature variability on childhood pneumonia and diarrhoea
are scarce, even though a big temperature change may pose a threat to the less-
developed immune system of children and thus may trigger their underlying
respiratory or intestinal health condition. In light of this, I did the other two studies
looking at the impacts of temperature variability (defined as diurnal temperature
range (DTR), and temperature change between two neighbouring days (TCN)) on
EDVs for childhood pneumonia and diarrhoea. The results of these two studies
iv
suggested that big TCN may increase EDVs for childhood pneumonia and diarrhoea.
Great DTR was associated with the increase in EDVs for childhood diarrhoea.
Indigenous children were particularly vulnerable to the impact of temperature
variability.
In summary, this thesis adds to the large body of literature on climate variability
impact on children’s health and may have significant implications for developing
climate change adaptation and paediatric care policies. Children’s health in Mount
Isa, the city where childhood pneumonia and diarrhoea co-distributed, requires more
attention from scientific and policy communities. This study improves our
understanding on how increasing temperature may affect the burden of two important
childhood diseases, pneumonia and diarrhoea, as climate change progresses. It also
calls for attention to the possible adverse impact of large temperature variability on
children’s health.
v
Table of Contents
Keywords .................................................................................................................................................i
Abstract .................................................................................................................................................. ii
Table of Contents .................................................................................................................................... v
List of Figures ........................................................................................................................................vi
List of Tables .........................................................................................................................................ix
List of Abbreviations ............................................................................................................................ xii Statement of Original Authorship ....................................................................................................... xiii
Acknowledgements .............................................................................................................................. xiv
CHAPTER 1: INTRODUCTION ....................................................................................................... 1
CHAPTER 2: LITERATURE REVIEW ......................................................................................... 11
CHAPTER 3: RESULTS PAPER ONE ........................................................................................... 43
CHAPTER 4: RESULTS PAPER TWO .......................................................................................... 65 CHAPTER 5: RESULTS PAPER THREE .................................................................................... 103
CHAPTER 6: RESULTS PAPER FOUR ....................................................................................... 133
CHAPTER 7: RESULTS PAPER FIVE ........................................................................................ 157
CHAPTER 8: DISCUSSION AND CONCLUSIONS ................................................................... 175
APPENDICES ................................................................................................................................... 189 Appendix A Publications .................................................................................................................... 189
vi
List of Figures
Figure 1-1. Conceptual framework of the thesis…………………………. 4
Figure 2-1. The flow chart of literature selection process………………... 16
Figure 3-1. The spatial distribution of EDVs for childhood pneumonia and diarrhoea
in Queensland, from 2007 to 2011………………………………………… 50
Figure 3-2. The spatial distribution of EDVs for childhood pneumonia by age in
Queensland, from 2007 to 2011.…………………………………………... 51
Figure 3-3. The spatial distribution of EDVs for childhood pneumonia by gender in
Queensland, from 2007 to 2011…………………………………..……….. 52
Figure 3-4. The spatial distribution of EDVs for childhood diarrhoea by age in
Queensland, from 2007 to 2011…………………………………..……….. 53
Figure 3-5. The spatial distribution of EDVs for childhood diarrhoea by gender in
Queensland, from 2007 to 2011…………………………………..……….. 54
Figure 3-6. The change of EDVs for childhood pneumonia and diarrhoea in
Queensland…………………………………..…………………………...... 55
Figure 3-7. The daily distribution of EDVs for childhood pneumonia and diarrhoea
in Queensland…………………………………..………………………….. 56
Figure 3-8. The spatial clusters of EDVs for childhood pneumonia and diarrhoea in
Queensland.……………………………………..………...............................57
Figure 3-9. The spatial patterns of mean temperature and rainfall in Queensland,
from 2007 to 2011.……………………………………..…………………... 58
Figure 4-1. The areas where satellite remote sensing temperature data were
collected…………………………………..………………………………... 71
vii
Figure 4-2a. The decomposed distribution of EDVs for paediatric pneumonia in
Brisbane, from 2001 to 2010………………………………………………. 76
Figure 4-2b. The daily distributions of climate variables in Brisbane, from 2001 to
2010…………………………………..…………………………………… 77
Figure 4-2c. The daily distributions of air pollutants in Brisbane, from 2001 to
2010…………………………………..…………………………………… 78
Figure 4-3. The pairwise plot of paediatric pneumonia, mean temperature, rainfall
and relative humidity in Brisbane, from 2001 to 2010……………………. 80
Figure 4-4. The overall effect of mean temperature on paediatric pneumonia in
Brisbane, from 2001 to 2010………………………………………………. 81
Figure 4-5. The change over time in the temperature effect on childhood pneumonia;
left hand side: hot effect; right hand side: cold effect; p1= 2001-2005, p2=2002-
2006, p3=2003-2007, p4=2004-2008, p5=2005-2009, p6=2006-2010……. 85
Figure 5-1. The daily distributions of EDVs for paediatric diarrhoea and climatic
factors in Brisbane, from 2001 to 2010……………………………………. 110
Figure 5-2. The daily distribution of diarrhoea caused by different
pathogens…………………………………..……………………………… 111
Figure 5-3. The overall effect of mean temperature on paediatric diarrhoea in
Brisbane, from 2001 to 2010……………………………………………… 114
Figure 5-4. The change over time in the temperature effect on childhood diarrhoea;
left hand side: hot effect; right hand side: cold effect; p1= 2001-2005, p2=2002-
2006, p3=2003-2007, p4=2004-2008, p5=2005-2009, p6=2006-2010…… 118
Figure 6-1. The daily distributions of EDVs for paediatric pneumonia, mean
temperature, DTR and TCN in Brisbane, from 2001 to 2010…………….. 139
viii
Figure 6-2. The overall effects of DTR and TCN on paediatric pneumonia in
Brisbane, from 2001 to 2010………………………………………………. 142
Figure 6-3. Monthly average number of days with TCN < -2 °C………… 143
Figure 6-4. The lagged effect of TCN on childhood pneumonia…………. 144
Figure 6-5. The effect of TCN on the total-, age-, gender- and ethnic-specific
childhood pneumonia in Brisbane, from 2001 to 2010 (Relative risk: The risk of
EDVs for childhood pneumonia on days with temperature drop =5.7 °C relative to
days with temperature drop= 2.0 °C) …………………………………….. 145
Figure 6-6. The effects of TCN on childhood pneumonia in summer and
winter…………………………………..…………………………………. 146
Figure 6-7. The overall effects of TCN on childhood pneumonia during two
periods…………………………………..………………………………… 147
Figure 6-8. The overall effect of TCN on childhood pneumonia in Brisbane, from
2001 to 2010 (excluding 2009) …………………………………………… 148
Figure 7-1. The daily distributions of EDVs for paediatric diarrhoea, mean
temperature, DTR and TCN in Brisbane, from 2001 to 2010……………... 163
Figure 7-2. The overall effects of DTR and TCN on paediatric diarrhoea in Brisbane,
from 2001 to 2010…………………………………..……………………... 166
Figure 7-3. The monthly distribution of days when DTR > 17 °C and TCN < -2
°C.…………………………………..………… ………………………….. 167
Figure 7-4a. The effect of DTR on the total-, age-, gender- and ethnic-specific
childhood diarrhoea in Brisbane, from 2001 to 2010……………………… 168
Figure 7-4b. The effect of TCN on the total-, age-, gender- and ethnic-specific
childhood diarrhoea in Brisbane, from 2001 to 2010. …………………….. 168
ix
List of Tables
Table 2-1. Characteristics of studies about ambient temperature and childhood
pneumonia…………………………………..…………………………….. 21
Table 2-2. Characteristics of studies about ambient temperature and childhood
diarrhoea…………………………………..……………………………… 23
Table 2-3. Characteristics of studies about rainfall and childhood pneumonia
…………………………………..………………………..………………. 26
Table 2-4. Characteristics of studies about rainfall and childhood diarrhoea
…………………………………..………………………..………………. 27
Table 2-5. Characteristics of studies about relative humidity and childhood
pneumonia …………………………………..……………………………. 29
Table 2-6. Characteristics of studies about relative humidity and childhood
diarrhoea…………………………………..……………………………… 30
Table 3-1. Summary statistics for EDVs for childhood pneumonia and diarrhoea by
postcode in Queensland, Australia, during 2007-2011………..………….. 47
Table 4-1. Summary statistics for climatic variables, air pollutants and paediatric
pneumonia in Brisbane, Australia, 2001–2010…………………………… 75
Table 4-2. Spearman’s correlation between daily weather conditions, air pollutants
and paediatric pneumonia in Brisbane, Australia, from 2001–
2010…………………………………..…………………………………… 79
Table 4-3. The cumulative effect of high and low temperatures on EDVs for
paediatric pneumonia, with 99th percentile (29.6 °C) and 1st percentile (10.4 °C) of
temperature relative to reference temperature (23°C) …………………… 82
x
Table 4-4. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2001 to 2010…………………………. 84
Table 4-5a. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2001 to 2005 ………………………… 87
Table 4-5b. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2002 to 2006………………………… 88
Table 4-5c. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2003 to 2007………………………… 89
Table 4-5d. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2004 to 2008………………………… 90
Table 4-5e. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2005 to 2009………………………… 91
Table 4-5f. Paediatric pneumonia due to the added effect of heat waves and cold
spells in Brisbane, Australia, from 2006 to 2010………………………… 92
Table 5-1. Summary statistics for climatic variables and paediatric diarrhoea in
Brisbane, Australia, 2001–2010………………………………………….. 109
Table 5-2. Spearman’s correlation between daily weather conditions, air pollutants
and paediatric diarrhoea in Brisbane, Australia, from 2001–2010………. 112
Table 5-3. The cumulative effect of high and low temperatures on EDVs for
paediatric diarrhoea in Brisbane, with 99th percentile (29.6 °C) and 1st (10.4°C) of
temperature relative to reference temperature (16 °C) …………………… 115
Table 5-4. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2001 to 2010……………………………………………… 117
Table 5-5a. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2001 to 2005…………………………………..…………. 119
xi
Table 5-5b. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2002 to 2006…………………………………………….. 120
Table 5-5c. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2003 to 2007…………………………………………….. 121
Table 5-5d. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2004 to 2008…………………………………..………… 122
Table 5-5e. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2005 to 2009…………………………………..…………. 123
Table 5-5f. Paediatric diarrhoea due to the added effect of heat waves in Brisbane,
Australia, from 2006 to 2010…………………………………………….. 124
Table 6-1. Spearman’s correlation between daily weather conditions, air pollutants
and paediatric pneumonia in Brisbane, Australia, from 2001–2010…….. 140
Table 7-1. Summary statistics for climatic variables, air pollutants and paediatric
diarrhoea in Brisbane, Australia, 2001–2010……………………………. 162
Table 7-2. Spearman’s correlation between daily weather conditions, air pollutants
and paediatric diarrhoea in Brisbane, Australia, from 2001–
2010…………………………………..………………………………….. 164
xii
List of Abbreviations
ARIMA Autoregressive integrated moving average
CI Confidence interval
DLNM Distributed lag non-linear model
DOW Day of week
DTR Diurnal temperature range
EDV Emergency department visit
ICD9 International classification of disease, ninth revision
ICD10 International classification of disease, tenth revision
IPCC Intergovernmental Panel on Climate Change
MeSH Medicine’s Medical Subject Headings
NASA National Aerospace and Space Administration
NO2 Nitrogen dioxide
O3 Ozone
OR Odds ratio
PM10 Particulate matter less than 10 μm in aerodynamic diameter
RR Relative risk
SD Standard deviation
TCN Temperature change between two neighbouring days.
QUT Verified Signature
xiv
Acknowledgements
I am very grateful to many people who have contributed in various ways to the
completion of this thesis. First and foremost, I would like to thank my supervisory
team, Prof. Shilu Tong, Dr. Wenbiao Hu, Dr. Weiwei Yu and Prof. Hong Su.
I would like to express my heartfelt gratitude to Shilu, my dear principal supervisor,
for his support in my whole PhD journey. He provided me with an opportunity to do
my PhD in QUT, and encouraged me to pursue a promising research topic. He has
always supported me without any reservation despite his busy schedule, teaching me
not only how to be a good researcher but also how to be a righteous person.
Particularly, he has put a lot of efforts to the design, conceptualisation and revision
of my manuscripts. I am so blessed to be a PhD student of Shilu and I always think
that he is one of the best supervisors ever. I am sure that the invaluable things which
I’ve learned from him will always guide me in my future career.
I am deeply grateful to Wenbiao, my associate supervisor, who has also made big
contributions to my research. He was exceptionally generous with his time, teaching
me how to do spatial and temporal analysis. Without his help, most goals are
unachievable in my PhD.
I am sincerely thankful to Dr. Weiwei Yu and Prof. Hong Su for their contributions
to my research.
I would also like to acknowledge the generous financial support for the completion
of my PhD program, including a China Scholarship Council Postgraduate
Scholarship, Queensland University of Technology Fee Waiving Scholarship, and a
CSIRO Top-up Scholarship from the Climate Adaptation Flagship Collaboration
Fund.
xv
Special thanks to my dear friends and colleagues who went through the journey with
me (especially Xiaoyu Wang, Xin Qi, Xiaofang Ye, Yan Bi, Cunrui Huang, Yuming
Guo, Jiajia Wang, Lyle R. Turner, Shahera Banu, Suchithra Naish, and Sam Toloo).
My deepest expression of appreciation is to my dear family for their patience and
support. Without their love, any achievement is meaningless.
Chapter 1: Introduction 1
Chapter 1: Introduction
1.1 Background
Pneumonia and diarrhoea are the two major killers in children younger than five
years (Walker et al. 2013). It is estimated that in 2010, there were more than 1.7
billion episodes of diarrhoea and 120 million episodes of pneumonia in children
under five years (Walker et al. 2013). Although mortality due to pneumonia and
diarrhoea has been declining in the past decades in developed countries, morbidity
from pneumonia and diarrhoea in these regions still remains high (Podewils et al.
2004). For example, pneumonia and diarrhoea are the common causes of
hospitalization for Australian children (Rudan et al. 2013; Scallan et al. 2005).
Pneumonia and diarrhoea are usually preventable (Bhutta et al. 2013). Prior studies
have put emphasis on identifying some individual-level risk factors, such as under
nutrition (Black et al. 2008) and not exclusively breastfeeding (Walker et al. 2013).
Some of these factors are associated with both pneumonia and diarrhoea, which may
result in the geographic co-distribution of pneumonia and diarrhoea (Walker et al.
2013), even though the geographic co-distribution of pneumonia and diarrhoea has
not been explored in previous studies.
Existing evidence also suggests that climatic factors, such as temperature (Checkley
et al. 2000; Green et al. 2010) and rainfall (Garcia-Vidal et al. 2013; Hashizume et al.
2007), may be associated with the occurrence of pneumonia and diarrhoea,
highlighting that increasing temperature may increase the incidence of pneumonia
and diarrhoea. However, prior studies looking at the impacts of temperature on
pneumonia and diarrhoea mainly used the temperature data collected from ground
2 Chapter 1: Introduction
monitoring sites (Checkley et al. 2000; Green et al. 2010), which may cause
measurement bias as most of these monitoring sites were in or nearby urban areas but
temperature usually varies across one city (Estes et al. 2009; Laaidi et al. 2012).
Satellite remote sensing technology may largely solve this problem, given its broad
spatial coverage (Anderson et al. 2012; Estes et al. 2009; Evans et al. 2013). To the
best of my knowledge, no study has used satellite remote sensing data to look at the
effects of temperature on childhood diseases so far.
There is a widespread consensus that climate is changing (IPCC 2013), and climate
change has posed a huge threat to children’s health (Xu et al. 2012b). Particularly,
global burden of childhood pneumonia and diarrhoea may continue to rise due to the
Earth’s increasing average surface temperature (Walker et al. 2013). Further, the
frequency of unstable weather patterns (e.g., sharp increase/decrease in temperature)
will also increase (Epstein 2005), and children are particularly vulnerable to big
temperature variation (Xu et al. 2013), due to their relatively less-developed
thermoregulation capability (Xu et al. 2012a).
Understanding the spatial and temporal patterns of childhood pneumonia and
diarrhoea (especially their geographic co-distribution), and exploring the climatic
drivers behind them using accurate exposure data will shed new light on future effect
to control and prevent these two leading childhood diseases. However, to date, no
studies have assessed the spatiotemporal patterns of childhood pneumonia and
diarrhoea in Queensland. Also, no studies have identified the climatic drivers of
childhood pneumonia and diarrhoea using satellite remote sensing data. This thesis
aimed to fill these knowledge gaps. The conceptual framework for this thesis is
shown in Figure 1-1.
Chapter 1: Introduction 3
1.2 Purposes
In this thesis, I used the data on emergency department visits for childhood
pneumonia and diarrhoea in Queensland, Australia, from January 1st 2001 to
December 31st 2011 and addressed three key issues:
I. What are the spatial and temporal patterns of EDVs for childhood pneumonia
and diarrhoea in Queensland, and is there any geographic co-distribution
of these two diseases?
II. What is the relationship between extreme temperatures and EDVs for
childhood pneumonia and diarrhoea?
III. What is the relationship between temperature variations and EDVs for
childhood pneumonia and diarrhoea?
4 Chapter 1: Introduction
Figure 1-1. Conceptual framework of the thesis
Chapter 5: Extreme temperature effect on
childhood diarrhoea in Brisbane
Climate change
Temperature and weather extremes increase
Climate variability may impact the burden of childhood
pneumonia and diarrhoea
Chapter 2: Identify the
knowledge gaps
Chapter 3: Explore the spatiotemporal patterns of childhood pneumonia
and diarrhoea in Queensland
Chapter 4: Extreme temperature effect on childhood pneumonia
in Brisbane
Chapter 6: Temperature
variability effect on childhood pneumonia
in Brisbane
Chapter 7: Temperature
variability effect on childhood diarrhoea in
Brisbane
Chapter 5: Extreme temperature effect on
childhood diarrhoea in Brisbane
Chapter 1: Introduction 5
1.3 Thesis Outline
This thesis is presented in a publication style. As such, each chapter is designed to
stand alone. In Chapter two, I reviewed the existing literature regarding the
association between climatic factors and childhood pneumonia and diarrhoea, and
found that temperature is the most likely climatic driver of childhood pneumonia and
diarrhoea. As climate change continues, burden of childhood pneumonia and
diarrhoea may vary, and the variation in specific regions may be largely driven by
climatic conditions, and the proportion of vulnerable children.
In Chapter three, I explored the spatiotemporal patterns and geographic co-
distribution of childhood pneumonia and diarrhoea in Queensland. A distinct
seasonality of childhood pneumonia and diarrhoea was found. Childhood pneumonia
and diarrhoea mainly distributed in northwest of Queensland, and Mount Isa had a
high-risk cluster where childhood pneumonia and diarrhoea co-distributed.
Chapter four is a time-series study looking at the impacts of extreme temperatures
and persistent extreme temperatures (i.e., heat waves and cold spells) on EDVs for
childhood pneumonia in Brisbane. It is observed that both high and low temperatures
were associated with an increase in EDVs for childhood pneumonia. Heat waves and
cold spells had significant added effects on childhood pneumonia, and the magnitude
of these effects increased with intensity and duration. However, there were changes
over time in both the main and added effects of temperature on childhood
pneumonia.
In Chapter five, I examined the effects of extreme temperatures and heat waves on
EDVs for childhood diarrhoea in Brisbane. We found both low and high
temperatures had significant impacts on childhood diarrhoea. Heat waves had an
6 Chapter 1: Introduction
added effect on childhood diarrhoea, and this effect increased with intensity and
duration of heat waves. There was a decreasing trend of heat effect on childhood
diarrhoea in Brisbane across the study period.
Chapter six is designed to assess the impact of temperature variability (diurnal
temperature range (DTR) and temperature change between two neighbouring days
(TCN)) on EDVs for childhood pneumonia in Brisbane. An adverse impact of TCN
on EDVs for childhood pneumonia was observed, and the magnitude of this impact
increased from the first five years (2001–2005) to the second five years (2006–2010).
Children aged 5–14 years, female children and Indigenous children were particularly
vulnerable to TCN impact. However, there was no significant association between
DTR and EDVs for childhood pneumonia.
Chapter seven is a time-series study quantifying the effect of temperature variability
on EDVs for childhood diarrhoea. It was observed that high DTR and TCN were
significantly associated with an increase in EDVs for childhood diarrhoea in
Brisbane. Every year, from May to September, especially July, children were at a
high risk posed by high DTR and low TCN, and male children were particularly
vulnerable to the adverse impact of DTR and TCN on diarrhoea.
Chapter eight highlighted the key findings of this thesis, compared these findings
with prior studies, discussed the mechanisms underlying these findings and
implications behind these finding, and explored how to manage the effects of
temperature on childhood pneumonia and diarrhoea.
Chapter 1: Introduction 7
1.4 References
Anderson H, Butland B, Donkelaar AV, Brauer M, Strachan D, Clayton T, et al.
2012. Satellite-based estimates of ambient air pollution and global variations
in childhood asthma prevalence. Environ Health Perspect 120(9):1333-1339.
Bhutta ZA, Das JK, Walker N, Rizvi A, Campbell H, Rudan I, et al. 2013.
Interventions to address deaths from childhood pneumonia and diarrhoea
equitably: what works and at what cost? Lancet 381(9875):1417-1429.
Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, et al. 2008.
Maternal and child undernutrition: global and regional exposures and health
consequences. Lancet 371(9608):243-260.
Checkley W, Epstein L, Gilman R, Figueroa D, Cama R. Patz J, et al. 2000. Effect of
El Niño and ambient temperature on hospital admissions for diarrhoeal
diseases in Peruvian children. Lancet 355(9202):442-450.
Epstein PR. 2005. Climate change and human health. N Engl J Med 353(14):1433-
1436.
Estes M, Al-Hamdan M, Crosson W, Estes S, Quattrochi D, Kent S, et al. 2009. Use
of remotely sensed data to evaluate the relationship between living
environment and blood pressure. Environ Health Perspect 117(12):1832-
1838.
Evans J, van Donkelaar A, Martin RV, Burnett R, Rainham DG, Birkett NJ, et al.
2013. Estimates of global mortality attributable to particulate air pollution
using satellite imagery. Environ Res 120:33-42.
Garcia-Vidal C, Labori M, Viasus D, Simonetti A, Garcia-Somoza D, Dorca J, et al.
2013. Rainfall is a risk factor for sporadic cases of Legionella pneumophila
Pneumonia. PLoS ONE 8(4):e61036.
8 Chapter 1: Introduction
Green R, Basu R, Malig B, Broadwin R, Kim J, Ostro B. 2010. The effect of
temperature on hospital admissions in nine California counties. Int J Public
Health 55(2):113-121.
Hashizume M, Armstrong B, Hajat S, Wagatsuma Y, Faruque AS, Hayashi T, et al.
2007. Association between climate variability and hospital visits for non-
cholera diarrhoea in Bangladesh: effects and vulnerable groups. Int J
Epidemiol 36(5):1030-1037.
IPCC. 2013. Summary for policymakers. In: Climate change 2013: the physical
science basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge.
Laaidi K, Zeghnoun A, Dousset B, Bretin P, Vandentorren S, Giraudet E, et al. 2012.
The impact of heat islands on mortality in Paris during the August 2003 heat
wave. Environ Health Perspect 120(2):254-259.
Podewils LJ, Mintz ED, Nataro JP, Parashar UD. 2004. Acute, infectious diarrhea
among children in developing countries. Semin Pediatr Infect Dis 15(3):155-
168.
Rudan I, O'Brien K, Nair H, Liu L, Theodoratou E, Qazi S, et al. 2013.
Epidemiology and etiology of childhood pneumonia in 2010: estimates of
incidence, severe morbidity, mortality, underlying risk factors and causative
pathogens for 192 countries. J Glob Health 3(1):10401.
Scallan E, Majowicz SE, Hall G, Banerjee A, Bowman CL, Daly L, et al. 2005.
Prevalence of diarrhoea in the community in Australia, Canada, Ireland, and
the United States. Int J Epidemiol 34(2), 454-460.
Chapter 1: Introduction 9
Walker CLF, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. 2013. Global
burden of childhood pneumonia and diarrhoea. Lancet 381(9875):1405-1416.
Xu Z, Etzel RA, Su H, Huang C, Guo Y, Tong S. 2012a. Impact of ambient
temperature on children's health: A systematic review. Environ Res 117:120-
131.
Xu Z, Huang C, Su H, Turner L, Qiao Z, Tong S. 2013. Diurnal temperature range
and childhood asthma: a time-series study. Environ Health 12(1):12.
Xu Z, Sheffield PE, Hu W, Su H, Yu W, Qi X, et al. 2012b. Climate change and
children’s health—A call for research on what works to protect children. Int J
Environ Res Public Health 9(9):3298-3316.
Chapter 2: Literature Review 11
Chapter 2: Literature Review
Impact of climatic factors on childhood pneumonia and diarrhoea: a
review of literature
Zhiwei Xu, Wenbiao Hu, Shilu Tong
12 Chapter 2: Literature Review
Abstract
Climate change is affecting and will continue to impact children’s health. To assess the
relationship between climatic factors and childhood pneumonia and diarrhoea, a literature
search was conducted using the databases PubMed, ProQuest, ScienceDirect, Scopus and
Web of Science. Empirical papers looking at the impact of climatic factors (defined as
temperature, rainfall and relative humidity) on childhood pneumonia or diarrhoea published
up to April 1st 2014 were included. Existing literature suggests that temperature is an
important climate driver of childhood pneumonia and diarrhoea. High or low rainfall, and
low relative humidity, may also increase the occurrence of childhood diarrhoea. Little
evidence regarding the effects of rainfall and relative humidity on childhood pneumonia was
found. As climate change continues, burden of childhood pneumonia and diarrhoea may vary,
and the variation in specific regions may be due largely to the major aetiological agent,
climate zones, and the proportion of vulnerable children. Future research should focus on
assessing temperature effects on childhood pneumonia and diarrhoea using more accurate
temperature data, exploring the relationship between temperature variation and childhood
pneumonia and diarrhoea, quantifying the vulnerability of different aetiological agents of
childhood pneumonia and diarrhoea to climatic factors, elucidating the modified effects of
climate types on the climate-pneumonia and climate-diarrhoea relationships, projecting the
future burden of childhood pneumonia and diarrhoea attributable to climate change, and
developing cost-effective adaptation measures to protect children from the adverse impact of
climate change.
Keywords: climate change; children; pneumonia; diarrhoea
Chapter 2: Literature Review 13
2.1 Introduction
Pneumonia and diarrhoea are the two major killers of children (Walker et al. 2013). In 2010,
there were more than 1.7 billion diarrhoea episodes and around 120 million pneumonia
episodes in children aged under five years, worldwide, causing approximately two million
deaths in 2011 (Walker et al. 2013).
Identifying the risk factors and taking preventive measures to block children from these risk
factors are urgently needed, given the huge burden of childhood pneumonia and diarrhoea.
Existing body of knowledge has acknowledged that individual-level biological and behaviour
factors, including suboptimum breastfeeding, underweight, stunting and zinc deficiency, may
largely contribute to the occurrence of pneumonia and diarrhoea (Walker et al. 2013). Some
prior studies have reported that climatic factors, such as temperature (Checkley et al. 2000;
Green et al. 2010), rainfall (Garcia-Vidal et al. 2013; Hashizume et al. 2007) and relative
humidity (D'Souza et al. 2008; Onozuka et al. 2009), may also be associated with the
transmission of pneumonia and diarrhoea. However, regional differences and contrasting
effects of climate on pneumonia and diarrhoea due to different aetiological agents are evident
(Ebi et al. 2001; Green et al. 2010; Hashizume et al. 2007; Paynter et al. 2013; Sumi et al.
2013).
As projected by Intergovernmental Panel on Climate Change (IPCC), the global surface
average temperature may increase by 1.8 to 4.0 °C relative to the 1961–1990 level by the end
of this century (IPCC 2007b), and more intense rainy seasons are likely to occur in Africa
and Asia (IPCC 2007a), the two regions where the burden of childhood pneumonia and
diarrhoea are the greatest. Climate change may affect the burden of childhood pneumonia and
diarrhoea (Walker et al. 2013), especially in some low-income countries. It is essential to
systematically review the relationship between climatic factors and childhood pneumonia and
diarrhoea so as to find out how climate change may impact burden of these two children
14 Chapter 2: Literature Review
killers in the future. This review aimed to elucidate the effects of climate (defined as
temperature, rainfall and relative humidity) on childhood pneumonia and diarrhoea, and to
propose future research directions. Herein, children were defined as humans under 18 years
old (American Academy of Pediatrics Committee on Environmental Health 2003).
2.2 Methods
Data sources
Empirical studies regarding climate and childhood pneumonia and diarrhoea published up to
April 1st 2014 were retrieved from the electronic databases PubMed, ProQuest,
ScienceDirect, Scopus, and Web of Science. References of the identified papers were
manually checked to make sure all relevant papers were included.
Inclusion criteria
We restricted our search to peer-reviewed articles written in English. The following U.S.
National Library of Medicine’s Medical Subject Headings (MeSH terms) and keywords were
used in the primary search: “climate change”, “climate”, “temperature”, “rainfall”,
“precipitation”, “humidity”, “child”, “pneumonia”, “diarrhea”, “diarrhoea” and “rotavirus”.
Eligibility included any empirical studies which used original data and appropriate effect
estimates (e.g., regression coefficient, relative risk, odds ratio, or percentage change in
morbidity for pneumonia or diarrhoea); where temperature, rainfall and/or relative humidity
was a main exposure of interest, and where childhood pneumonia and diarrhoea were
analysed. The effect estimates (e.g., relative risks, confidence intervals) were recorded from
the papers identified. In this review, we discussed the impact of climatic factors on rotavirus
diarrhoea separately as there are too many papers focusing on this topic.
Chapter 2: Literature Review 15
2.3 Results
We identified 8275 papers in the initial search (excluding papers talking about climate
variability and rotavirus), and 14 papers were in the final review (Figure 2-1). Five papers
have looked at the impact of climate on childhood pneumonia, and nine papers have
examined the association between climate and childhood diarrhoea.
16 Chapter 2: Literature Review
Figure 2-1. The flow chart of literature selection process
Potentially relevant studies in the initial searching
(n= 8275)
7894 excluded due to irrelevant titles
Studies after reviewing the titles
(n=381)
344 did not meet inclusion criteria according to abstract
Studies retrieved for more detailed evaluation
(n=37)
26 articles excluded (5 no appropriate effect estimate; 3
no original data; 18 no specific data for children)
Studies met inclusion criteria (n=11)
3 articles added by inspecting reference lists
Studies included in final review (n=14)
Chapter 2: Literature Review 17
Temperature
Five studies, conducted in Brazil (Souza et al. 2012), China (Xu et al. 2011), Philippine
(Paynter et al. 2013), and the USA (Ebi et al. 2001; Green et al. 2010), have looked at the
impact of temperature on childhood pneumonia (Table 2-1). Ebi et al. examined the effect of
temperature on hospitalizations for viral pneumonia in six counties of California, the USA,
from 1993 to 1998, and found hospitalizations for viral pneumonia increased significantly
with the decrease of temperature in children aged under 18 years, though the percent changes
varied from 21.6% to 30.7% across different counties (Ebi et al. 2001). In the other nine
counties of California, Green et al. quantified the effect of high temperature on hospital
admissions for pneumonia from 1999 to 2005, and found hospital admissions for pneumonia
in children aged under five years increased by 5.9% per 10 oF (5.6 ºC) increase in mean
apparent temperature (Green et al. 2010). In Hangzhou, China, Xu et al. investigated the
relationship between temperature and childhood pneumonia and found Mycoplasma
pneumoniaie pneumonia rate in children under 18 years increased by 0.83% per 1ºC increase
in monthly mean air temperature (Xu et al. 2011). In Campo Grande, Brazil, Souza et al.
looked at the association between temperature and outpatient visits in children aged 5 to 14
years, and found minimum temperature was positively associated with pneumonia (relative
risk (RR): 1.12), and maximum temperature was negatively associated with pneumonia (RR:
0.92) (Souza et al. 2012). One note-worthy study conducted in Japan examined the impact of
mean temperature on reported Mycoplasma pneumonia in the total population, finding that
93.7% of the reported Mycoplasma pneumonia occurred in children under 15 years, and the
weekly number of Mycoplasma pneumonia cases increased by 16.9% for every 1% increase
in mean temperature (Onozuka et al. 2009). Recently, Paynter et al. used both time-series and
case-crossover designs to examine the effect of temperature on hospitalizations for
pneumonia in children aged under three years from 2000 to 2004 in Philippine, but did not
18 Chapter 2: Literature Review
find any significant relationship between temperature and childhood pneumonia (RR: 0.93
(95% confidence interval (CI): 0.63 to 1.35)) (Paynter et al. 2013).
Seven studies from Australia (Lam 2007), Bangladesh (Hashizume et al. 2007), Japan
(Onozuka and Hashizume 2010), Peru (Checkley et al. 2000), Sub-Saharan Africa
(Bandyopadhyay et al. 2012), Taiwan (Chou et al. 2010), and the USA (Green et al. 2010)
have examined the relationship between temperature and childhood diarrhoea (Table 2-2).
Bandyopadhyay et al. explored the effect of monthly temperature on diarrhoea cases in 14
Sub-Saharan African countries from 1992 to 2001, finding that an increase in monthly
average maximum temperature raises the prevalence of diarrhoea while an increase in
monthly minimum temperature reduces diarrhoea in children under three years of age
(Bandyopadhyay et al. 2012). Two studies used maximum temperature as temperature
indicator to examine the temperature-diarrhoea relationship (Chou et al. 2010; Lam 2007). In
Sydney, Australia, Lam investigated the association between maximum temperature and
hospital emergency department visits for gastroenteritis in children aged under six years and
found hospital emergency department visits for gastroenteritis increased by 11% per 1ºC
increase in maximum temperature (Lam 2007). In Taiwan, Chou et al. explored the impact of
maximum temperature on hospital admissions for diarrhoea in children aged under 15 years
from 1996 to 2007, and found hospital admissions for diarrhoea increased by 4% per 1 °C
increase in maximum temperature (Chou, et al. 2010). Another four studies have explored the
effect of mean temperature on diarrhoea (Checkley et al. 2000; Green et al. 2010; Hashizume
et al. 2007; Onozuka and Hashizume 2010), and three of them reported increased diarrhoeal
cases associated with increasing temperature (Checkley et al. 2000; Green et al. 2010;
Hashizume et al. 2007). While, Onozuka et al. found that there was a temperature threshold,
either above or below which, hospital admissions for infectious gastroenteritis in children
aged under 15 years increased (Onozuka and Hashizume 2010).
Chapter 2: Literature Review 19
Rainfall
So far, only two studies in Philippines (Paynter et al. 2013) and the USA (Ebi et al. 2001)
have quantified the effect of rainfall on childhood pneumonia (Table 2-3). Ebi et al. explored
the effect of rainfall on viral pneumonia in six California counties from 1983 to 1998, but
found there was no significant relationship between rainfall and hospitalizations for viral
pneumonia in children aged under 18 years (Ebi et al. 2001). In Bohol Province, Philippines,
Paynter et al. investigated the association between rainfall and hospitalizations for pneumonia
from 2000 to 2004 in children aged under three years, and also found there was no significant
relationship between them (RR:2.37 (95% CI: 0.85 to 6.60)).
Four studies, conducted in Australia (Lam 2007), Bangladesh (Hashizume et al. 2007), Brazil
(Andrade et al. 2009), and the USA (Drayna et al. 2010), have looked at the impact of rainfall
on childhood diarrhoea (Table 2-4). In Rio Grande do Norte, Brazil, Andrade et al. examined
the effect of rainfall on hospitalizations for diarrhoea in infants from 1992 to 2001, and found
the diarrhoea hospitalizations increased with the increase of rainfall (Andrade et al. 2009).
Drayna et al. found that any rainfall four days prior was significantly associated with an 11%
increase in paediatric emergency department visits for acute gastrointestinal in Wisconsin, the
USA (Drayna et al. 2010). Hashizume et al. found that in Dhaka, Bangladesh, there was a
non-linear relationship between rainfall and childhood diarrhoea. Below or above a rainfall
threshold, hospital visits for non-cholera diarrhoea in children under 15 years increased
rapidly (Hashizume et al. 2007). However, in Sydney, Australia, Lam did not find any
significant relationship between rainfall and hospital emergency department visits for
gastroenteritis in children under six years of age (RR:0.98 (P=0.09)) (Lam 2007).
Relative humidity
There are two studies looking at the effect of relative humidity on pneumonia in children
(Paynter et al. 2013; Souza et al. 2012) and both of them did not find any significant
20 Chapter 2: Literature Review
relationship between relative humidity and childhood pneumonia (Table 2-5). However,
Onozuka et al. examined the impact of relative humidity on reported Mycoplasma
pneumoniae pneumonia in not just children but also adults in Fukuoka, Japan, and found that
93.7% of the reported Mycoplasma pneumoniae pneumonia were in children under 15 years,
and the weekly number of Mycoplasma pneumoniae pneumonia cases increased by 4.1% for
every 1% increase in relative humidity (Onozuka et al. 2009).
Three studies have formally explored the impact of relative humidity on childhood diarrhoea
(D'Souza et al. 2008; Lam 2007; Onozuka and Hashizume 2010) (Table 2-6). Onozuka and
Hashizume quantified the effect of relative humidity on hospital admission for infectious
gastroenteritis in children under 15 years in Japan, and found the increase in diarrhoeal cases
per 1% drop in relative humidity was 3.9% (Onozuka and Hashizume 2010). While, Lam did
not find any significant relationship between relative humidity and hospital emergency
department visits for gastroenteritis in children aged under six years in Sydney, Australia
(Lam 2007).
Climate variability and rotavirus diarrhoea
Numerous studies have examined the impacts of temperature, rainfall and relative humidity
on rotavirus diarrhoea in children. Levy et al. reviewed the seasonality of rotavirus in the
tropics and its relationship with climate variability (Levy et al. 2009). They found rotavirus
incidence reduces by 10%, 1%, and 3%, respectively, for every 1 ºC increase in mean
temperature, 1 cm increase in mean monthly rainfall, and 1% increase in relative humidity.
Jagai et al. also quantitatively reviewed the impacts of meteorological factors on rotavirus in
South Asia, and found A 1 ºC decrease in monthly ambient temperature and a decrease of 10
mm in precipitation are associated with 1.3% and 0.3% increase above the annual level in
rotavirus infections, respectively (Jagai et al. 2012)
Chapter 2: Literature Review 21
Table 2-1. Characteristics of studies about ambient temperature and childhood pneumonia Studya Location and time Research design and
statistical analysis Main temperature
exposure variable(s) Outcome(s) Key findings Effect estimates
Ebi et al. 2001
Six California counties, the USA,
January 1983 to June 1998, children aged 0-
17 years
Time-series; Poisson generalized estimating equations
model
Sea surface temperature
Hospitalizations for viral
pneumonia in females
Hospitalizations for viral pneumonia increased
significantly with temperature decreased
Percent change in Sacramento and Yolo
Counties: 30.7 (95% CI: 29.0 to 32.4)
Percent change in San Francisco and San Mateo Counties: 24.6 (95% CI:
9.9 to41.2 ) Percent change in Los Angeles and Orange
Counties: 21.6 (95% CI: 16.4 to 27.1)
Green et al. 2010
Nine California counties, the USA, May to September,
1999 to 2005, children under five years
Case-crossover; Conditional
logistic regression,
meta-analysis
Daily mean apparent temperature
Hospital admissions for
pneumonia
Hospital admissions for pneumonia in children aged under five years
increased with the increase of mean
temperature
Percent change: 5.9% (95% CI: -1.7% to 14.1%)
Xu et al. 2011
Hangzhou, China, January 2007 to December 2009, children under 18
years
Multiple linear regression
Monthly mean air temperature
Lab-confirmed hospitalizations for pneumonia
Mycoplasma pneumonia rate increased by 0.83%
with an increase of 1ºC in monthly mean air
temperature
Percent change: 0.83%
22 Chapter 2: Literature Review
Table 2-1. Characteristics of studies about ambient temperature and childhood pneumonia (continued)
Studya Location and time Research design
and statistical analysis
Main temperature exposure
variable(s) Outcome(s) Key findings Effect estimates
Souza et al. 2012
Campo Grande, Brazil, 2004 to
2008, children aged 5-14 years
Time-series: Poisson regression
Daily maximum and minimum temperatures
Outpatient visits for pneumonia
Minimum temperature was positively associated with pneumonia, and
maximum temperature was negatively associated with
pneumonia
1). Minimum temperature:
RR:1.12; 2). Maximum
temperature: RR: 0.92
Paynter et al. 2013
Bohol Province, Philippines, 2000 to 2004, children under three years
Time-series; Poisson regression
Case-crossover; Conditional logistic
regression
Weekly mean temperature
Hospitalizations for pneumonia
No significant relationship between
temperature and hospitalizations for
pneumonia was found
RR: 0.93 (95% CI: 0.63 to 1.35)
aThese studies are ordered by the date of publication and the first author. Abbreviations: CI, confidence interval; OR, odds ratio; RR, relative risk.
Chapter 2: Literature Review 23
Table 2-2. Characteristics of studies about ambient temperature and childhood diarrhoea
Studya Location and time Research design and statistical analysis
Main temperature exposure variable(s) Outcome(s) Key findings Effect estimates
Checkley et al. 2000
Lima, Peru, January 1993 to November
1998, children under 10 years
Time-series; Generalized additive
model Mean temperature
Hospital admissions for
diarrhoea
Admissions for diarrhoea increased by 8% per 1ºC increase in mean ambient
temperature
RR:1.08
Hashizume et al. 2007
Dhaka, Bangladesh, January 1996-
December 2002, children aged under
15 years
Time-series; Poisson generalized
linear model
Daily maximum and minimum
temperature
Weekly hospital visits for non-
cholera diarrhoea
Percentage change in the number of non-cholera
diarrhoeal cases per week for 1ºC increase in
temperature at lag 0–4 weeks was only
statistically significant in children ≤ 14 years old
Percent change: 5.7%; (95% CI: 2.9% – 8.6%)
Lam 2007
Sydney, Australia, January 2001 and December 2002,
children under six years
Time-series; ARIMA
Daily maximum and minimum
temperature
Hospital emergency
department visits for gastroenteritis
Hospital emergency department visits for
gastroenteritis increased by 11% per 1ºC increase in maximum temperature
RR:1.11 (P=0.007)
24 Chapter 2: Literature Review
Table 2-2. Characteristics of studies about ambient temperature and childhood diarrhoea (continued) Studya Location and time Research design and
statistical analysis Main temperature
exposure variable(s) Outcome(s) Key findings Effect estimates
Chou et al. 2010
Taiwan, 1996-2007, children aged under
15 years
Time-series; Poisson regression
model
Monthly maximum temperature
Hospital admissions for
diarrhoea
Hospital admissions for diarrhoea among children aged 0–14 years increased by 4% per 1 °C increase in maximum temperature
RR:1.04 (P=0.012)
Green et al. 2010
Nine California counties, the USA, May to September
1999–2005, children aged under five years
Case-crossover; Conditional
logistic regression,
meta-analysis
Daily mean apparent temperature
Hospital admissions for
intestinal infectious diseases
The highest effects of ambient temperature on
intestinal infectious disease were seen in
children aged 5-18 years
Percent change: 21.3% (95% CI: 5.2% –39.8%)
Onozuka and
Hashizume 2011
Fukuoka, Japan, 2000-2008, children aged under 15 years
Time-series; Negative binomial
regression
Daily mean temperature
Hospital admission for
infectious gastroenteritis
Among children aged under 15 years, every 1
°C increase in temperature below 13 °C
was associated with a 23.2% infectious
gastroenteritis increase, while every 1 °C increase in temperature above 13
°C was associated with an 11.8% infectious
gastroenteritis decrease
1). Hot: Percent change:11.8% (95% CI:
6.6% – 17.3%)
2). Cold: Percent change: 23.2% (95% CI:16.6%–
30.2%)
Chapter 2: Literature Review 25
Table 2-2. Characteristics of studies about ambient temperature and childhood diarrhoea (continued) Studya Location and time Research design and
statistical analysis Main temperature
exposure variable(s) Outcome(s) Key findings Effect estimates
Bandyopadhyay et al. 2012
14 Sub-Saharan African countries,
1992-2001, children aged under three
years
Time-series; Random effect
model
Monthly average maximum and
minimum temperature
Diarrhoea cases from a survey
An increase in monthly average maximum
temperature raises the prevalence of diarrhoea
while an increase in monthly minimum
temperature reduces diarrhoea in children
under three years of age
1). Maximum temperature: Coefficient:
1.013 (P<0.01); 2). Minimum
temperature: Coefficient: -0.475 (P<0.01)
aThese studies are ordered by the date of publication and the first author. Abbreviations: ARIMA, autoregressive integrated moving average; CI, confidence interval; RR, relative risk.
26 Chapter 2: Literature Review
Table 2-3. Characteristics of studies about rainfall and childhood pneumonia Studya Location and time Research design and
statistical analysis Main temperature
exposure variable(s) Outcome(s) Key findings Effect estimates
Ebi et al. 2001
Six California counties, January
1983 to June 1998, children aged 0-17
years
Time-series; Poisson generalized estimating
equations model Daily rainfall
Hospitalizations for viral
pneumonia in females
There was no significant relationship between
rainfall and hospitalizations for viral
pneumonia
RR: 1.5 (95% CI: -1.8 to
4.8)
Paynter et al. 2013
Bohol Province, Philippines, July 2000
to December 2004, children aged under
three years
1). Time-series; Distributed lag
Poisson regression 2). Case-crossover; Conditional logistic
regression
Weekly rainfall Hospitalizations for pneumonia
No significant relationship between
rainy days and hospitalizations for
pneumonia was found
RR:2.37 (95% CI: 0.85 to 6.60)
aThese studies are ordered by the date of publication and the first author. Abbreviations: CI, confidence interval; RR, relative risk.
Chapter 2: Literature Review 27
Table 2-4. Characteristics of studies about rainfall and childhood diarrhoea Studya Location and time Research design and
statistical analysis Main temperature
exposure variable(s) Outcome(s) Key findings Effect estimates
Hashizume et al. 2007
Dhaka, Bangladesh, January 1996-
December 2002, children aged under
15 years
Time-series; Poisson generalized
linear model
Daily maximum and minimum
temperature
Weekly hospital visits for non-
cholera diarrhea
The number of non-
cholera diarrhoea cases per week increased both
above and below the threshold of 52mm of
average rainfall over lags 0–8 weeks.
1). Above the rainfall
threshold: Percent change: 5.1% (95% CI:
3.3%–6.8%) 2). Below the rainfall
threshold: Percent change: 3.9% (95% CI:
0.6%–7.2%)
Lam 2007
Sydney, Australia, January 2001 and December 2002,
children aged under six years
Time-series; ARIMA Daily rainfall
Hospital emergency
department visits for gastroenteritis
No significant relationship was found between rainfall and hospital emergency department visits for
gastroenteritis
RR:0.98 (P=0.09)
Andrade et al. 2009
Rio Grande do Norte, Brazil, 1992 to 2001, children under 1 year
of age
Time-series; Distributed lag model Monthly rainfall Hospitalizations
for infant diarrhea
There was a positive relationship between
rainfall and infant diarrhea
RR:1.002 (P<0.001)
Drayna et al. 2010
Wauwatosa,
Wisconsin, the USA, January 2002 to December 2007, children (age not
given)
Time-series; ARMA Daily rainfall
Emergency department visits
for acute gastrointestinal
illness
Any rainfall 4 days prior
was significantly associated with an 11%
increase in acute gastrointestinal illness
visits.
RR: 1.11
28 Chapter 2: Literature Review
aThese studies are ordered by the date of publication and the first author. Abbreviations: ARIMA, autoregressive integrated moving average; ARMA, autoregressive moving average; CI, confidence interval; RR, relative risk.
Chapter 2: Literature Review 29
Table 2-5. Characteristics of studies about relative humidity and childhood pneumonia Studya Location and time Research design and
statistical analysis Main temperature
exposure variable(s) Outcome(s) Key findings Effect estimates
Souza et al. 2012
Campo Grande,
Brazil, 2004 to 2008, children aged 5-14
years
Time-series: Poisson regression Daily humidity Outpatient visits
for pneumonia
No significant relationship between
humidity and outpatient visits for pneumonia was
found
RR: 1.00
Paynter et al. 2013
Bohol Province, Philippines, July 2000
to December 2004, children aged under
three years
Time-series; Distributed lag
Poisson regression Case-crossover;
Conditional logistic regression
Weekly relative humidity
Hospitalizations for pneumonia
No significant relationship between relative humidity and hospitalizations for
pneumonia was found
RR: 1.01 (95% CI: 0.63 to 1.35)
aThese studies are ordered by the date of publication and the first author. Abbreviations: CI, confidence interval; OR, odds ratio; RR, relative risk.
30 Chapter 2: Literature Review
Table 2-6. Characteristics of studies about relative humidity and childhood diarrhoea
Studya Location and time Research design and statistical analysis
Main temperature exposure variable(s) Outcome(s) Key findings Effect estimates
D’Souza et al. 2008
Three cities of Australia, 1993 to
2003, children aged under five years
Time-series; Negative binomial
log-linear regression Weekly relative
humidity Hospital
admissions for rotavirus diarrhoea
Higher humidity in the previous week was
associated with a decrease in rotavirus diarrhoeal
admissions in three cities
1). Canberra: RR:0.98; (95% CI:0.97 to 0.99) 2). Brisbane: RR: 0.98; (95% CI: 0.97 to 0.99)
3). Melbourne: RR: 0.99; (95% CI: 0.99 to 1.00)
Onozuka and
Hashizume 2011
Fukuoka, Japan, 2000-2008, children aged
under 15 years
Time-series; Negative binomial
regression
Daily relative humidity
Hospital admission for
infectious gastroenteritis
The increase in cases per 1% drop in relative humidity
was 3.9%
Percent change: 3.9% (95% CI:2.8 to 5.0)
Lam 2007
Sydney, Australia, January 2001 and December 2002,
children aged under six years
Time-series; ARIMA
Daily relative humidity
Hospital emergency
department visits for gastroenteritis
No significant relationship was found between
relative humidity and hospital emergency department visits for
gastroenteritis
RR:1.01 (P=0.165)
aThese studies are ordered by the date of publication and the first author. Abbreviations: CI, confidence interval; RR, relative risk.
Chapter 2: Literature Review 31
2.4 Discussion
This review systematically examined the association between climatic factors and
childhood pneumonia and diarrhoea. Majority of prior studies have found significant
impact of temperature on childhood pneumonia and diarrhoea, although the
magnitude of temperature effect varies with regions. Additionally, the difference in
climatic effects between temperate and tropical/subtropical regions was observed.
This difference can be attributable to many factors. One of the fundamental reasons
is the different relative importance of climatic factors in affecting the pathological
agents of childhood pneumonia and diarrhoea between temperate and
tropical/subtropical regions (Haynes et al. 2013; Patel et al. 2013). The popular
opinion is that viruses causing childhood pneumonia or diarrhoea (e.g., RSV and
rotavirus) had a distinct seasonal peak in regions with temperate climates but was
year-round in the tropical/subtropical areas (Patel et al. 2013; Pica and Bouvier
2014). Some data suggested that diarrhoeal diseases caused by rotavirus increased
during cool and dry seasons (Brandt et al. 1982; Haffejee 1995). At regional level,
many factors may interact and explain the seasonality of childhood infectious
diseases, including climate, transmission patterns, host behaviour and susceptibility.
The impact of these factors may also be context specific, and normally no factor
alone can fully capture the complexity of seasonality of childhood infectious
diseases.
It is not easy to do a meta-analysis quantifying the temperature impact on childhood
diarrhoea because of the different study designs, various temperature indicators and
heterogeneous types of diarrhoeal cases researchers used. No significant effect of
rainfall on childhood pneumonia was found. Existing literature suggests there is a
significant (most likely non-linear) relationship between rainfall and childhood
32 Chapter 2: Literature Review
diarrhoea. Also, little evidence regarding the association between relative humidity
and childhood pneumonia was found, while some studies imply that there might be a
reverse association between relative humidity and childhood diarrhoea (D'Souza et
al. 2008; Onozuka and Hashizume 2010).
Mechanisms underlying temperature impacts on childhood pneumonia and
diarrhoea
Temperature is the most likely climate driver of childhood pneumonia, in temperate
settings. The discrepancy in the results of studies regarding the temperature-
pneumonia relationship may be attributable to different factors, including the age of
children (Souza et al. 2012), geography, study design (Ebi et al. 2001; Green et al.
2010), the genotypes of pathogens, and the dominant pathogens etc., among which,
the different prevailing aetiological agents in the study populations may be the most
important factor. The existing body of knowledge suggests that low temperature is
associated with peaks of respiratory syncytial virus (RSV) (Yusuf et al. 2007) and
Streptococcus pneumoniae (Herrera-Lara et al. 2013; Watson et al. 2006), and high
temperature may increase the pneumonia caused by Mycoplasma pneumoniae
(Onozuka et al. 2009; Xu et al. 2011), Pneumocystis (Djawe et al. 2013), and
Legionella pneumophila (Herrera-Lara et al. 2013). Streptococcus pneumoniae,
Haemophilus influenzae type B, RSV and influenza virus are the most common
pathogens worldwide (Rudan et al. 2013), and thus it is likely that the increasing
global surface temperature may to some extent decrease burden of childhood
pneumonia, though discrepancy may exist between different regions.
Most studies looking at the impact of temperature on childhood diarrhoea reported a
linear relationship, with a constant increase for childhood diarrhoeal cases by one
unit increase or decrease in temperature (Bandyopadhyay et al. 2012; Checkley et al.
Chapter 2: Literature Review 33
2000; Chou et al. 2010; D'Souza et al. 2008; Hashizume et al. 2007; Lam 2007).
Temperature may impact the transmission of childhood diarrhoea mainly through
three pathways. Firstly, high temperature promotes the growth of bacteria
(Hashizume et al. 2007), and low temperature increases the replication and survival
of virus, e.g., rotavirus (D'Souza et al. 2008). Secondly, food chain, from food
preparation stage to production process (D'Souza et al. 2004), may be affected by
temperature. Studies in England and Wales reported that food poisoning occurred
more in high temperatures (Bentham and Langford 1995; Bentham and Langford,
2001), indicating that there might be more childhood diarrhoea caused by food
poisoning in this area as climate change continues. Thirdly, high and cold
temperatures may alter people’s hygiene behaviour. For example, in hot days, people
are more likely to have cold water without disinfection, which expose them more to
bacteria. Under climate change context, it is pivotal to explore the distribution of
aetiological agents of childhood diarrhoea globally and further to project the future
childhood diarrhoea burden variation attributable to climate change.
Mechanisms underlying rainfall impact on childhood pneumonia and diarrhoea
No study, so far, has found a significant relation between rainfall and childhood
pneumonia, even though in monsoon seasons, children are more likely to spend time
indoors, which may increase crowding and their exposure to biomass fuel smoke, as
well as decrease their sunlight exposure, possibly resulting in a higher risk of getting
pneumonia (Paynter et al. 2010). In adults, results regarding the impact of rainfall on
pneumonia are not convincing either (Murdoch and Jennings 2009; Yusuf et al.
2007). Both low and high rainfall is reported to be associated with increase in
childhood diarrhoea. The mechanisms explaining the impact of rainfall on diarrhoea
are complex. The popular opinion is: Low rainfall/drought may result in water
34 Chapter 2: Literature Review
scarcity, leading to more use of unprotected water sources and reducing hygiene
practices, and high rainfall/extreme precipitation may flush faecal contaminants from
dwellings into water supplies (Jofre et al. 2010).
Mechanisms underlying relative humidity impact on childhood pneumonia and
diarrhoea
RSV infections were reported positively associated with relative humidity in China,
Indonesia, Malaysia, Mexico, and Singapore (Chan et al. 2002; Loh et al. 2011;
Omer et al. 2008; Tang et al. 2010; Yusuf et al. 2007). A lab-based study suggested
that the survival of airborne Mycoplasma pneumoniae was a function of temperature
and relative humidity, and temperature response was mediated by relative humidity
(Wright et al. 1969). However, to date, very few studies have formally quantified the
association between relative humidity and childhood pneumonia. In terms of
humidity-diarrhoea relation, existing knowledge suggests that low relative humidity
may increase childhood diarrhoeal cases. One explanation is that low humidity
facilitates the survival and replication of rotavirus (D'Souza et al. 2008). Recently,
some researchers argued that rotavirus can be aerosolized, and low relative humidity
may facilitate the development of virus-laden dust and droplet nuclei, increasing
aerial transport (Levy et al. 2009; Shaman and Kohn 2009; Shaman et al. 2011).
Mechanism underlying climate variability impact on rotavirus diarrhoea
Strong evidence suggests that rotavirus normally peaks in dry and cold conditions.
Rotavirus can retain its infectivity for several days in aqueous environments, and
waterborne spread has been implicated in a number of rotavirus outbreaks (Ansari et
al. 1991). Some researchers argued that rotavirus spreads through air, although the
rotavirus has not been isolated from the respiratory tract (Bishop 1996; Haffejee
1995). The mechanisms underlying why rotavirus favours dry and cold environment
Chapter 2: Literature Review 35
is currently unclear. It was postulated that relative drop in humidity and rainfall
combined with drying of soils in lower temperatures might increase the aerial
transport of dried, contaminated fecal material, and might also lead to increased
formation of dust, which could provide a substrate for the virus particles (Levy et al.
2009).
2.5 Knowledge Gaps
The impact of climate on childhood pneumonia and diarrhoea has become an
important public health issue. However, there are still some knowledge gaps. First,
existing studies assessing the impact of temperature on pneumonia or diarrhoea
basically used the temperature data collected from limited ground monitors, which
may result in measurement bias because temperature across one city is spatially
variable (Zhang et al. 2011). It is urgently needed to apply some advanced
approaches (e.g., satellite remote sensing) to assessing temperature effect on
childhood pneumonia and diarrhoea. Second, seldom studies have examined the
relationship between temperature variation (diurnal temperature range (DTR) and
temperature change between two neighbouring days) and childhood pneumonia and
diarrhoea, even though a sudden temperature change may pressure on children’s
immune system (Bull 1980). Third, previous studies looking at the effect of climate
on childhood pneumonia or diarrhoea mainly used hospital admissions or outpatient
visits as the outcome variable, and the results of these studies depend largely on the
distribution of major pathogens causing pneumonia or diarrhoea in hospitalized
children. It is essential to use more accurate (e.g., lab-confirmed) data to investigate
the association between climate and a variety of specific pathogens using solid
statistical approaches. Forth, most previous studies were conducted in the settings of
36 Chapter 2: Literature Review
the same climate type (either in temperate or tropical), and few multi-city studies
have been done so far (D'Souza et al. 2004), even though it is widely accepted that
the major climate drivers of pneumonia and diarrhoea vary across different climate
regions. Future research may use the data in cities of different climates and explore
the variation of the association between climate and pneumonia or diarrhoea in
different climates, using a consistent statistical method. Fifth, as climate change
continues, burden of childhood pneumonia and diarrhoea may change accordingly.
However, no study has specifically project the future burden of childhood pneumonia
and diarrhoea under climate change scenarios. Finally, so far, little research attention
has been paid to developing climate change adaptation measures, and it is important
to develop cost-effective adaptation measures to relieve the burden of childhood
pneumonia and diarrhoea due to climate change.
2.6 Conclusions
Existing scientific literature suggests a very-likely association between temperature
and childhood pneumonia and diarrhoea. There also appears to be a non-linear
relationship between rainfall and childhood diarrhoea, with both very low or high
rainfall increasing diarrhoeal cases, and high relative humidity may decrease the
incidences of childhood diarrhoea. Limited evidence on the impacts of rainfall and
relative humidity on childhood pneumonia was offered by prior studies.
2.7 References
American Academy of Pediatrics Committee on Environmental Health. 2003. In:
Etzel RA (ed) Pediatric Environmental Health, 2nd edn. Elk Grove Village,
IL: American Academy of Pediatrics.
Chapter 2: Literature Review 37
Andrade IG, Queiroz JW, Cabral AP, Lieberman JA, Jeronimo SMB. 2009.
Improved sanitation and income are associated with decreased rates of
hospitalization for diarrhoea in Brazilian infants. Trans R Soc Trop Med Hyg
103(5):506-511.
Ansari S, Springthorpe V, Sattar S. 1991. Survival and vehicular spread of human
rotaviruses: possible relation to seasonality of outbreaks. Rev Infect Dis
13(3):448-461.
Bandyopadhyay S, Kanji S, Wang L. 2012. The impact of rainfall and temperature
variation on diarrheal prevalence in Sub-Saharan Africa. Appl Geogr 33:63-
72.
Bentham G, Langford I. 1995. Climate change and the incidence of food poisoning
in England and Wales. Int J Biometeorol 39(2):81-86.
Bentham G, Langford I. 2001. Environmental temperatures and the incidence of food
poisoning in England and Wales. Int J Biometeorol 45(1):22-26.
Bishop RF. 1996. Natural history of human rotavirus infection. In S. Chiba, M.
Estes, S. Nakata & C. Calisher (Eds.), Viral Gastroenteritis (Vol. 12, pp. 119-
128): Springer Vienna.
Brandt CD, Kim HW, Rodriguez WJ, Arrobio JO, Jeffries BC, Parrott RH. 1982.
Rotavirus gastroenteritis and weather. J Clin Microbiol 16(3):478-482.
Bull G. 1980. The weather and deaths from pneumonia. Lancet 1(8183):1405-1408.
Chan PWK, Chew FT, Tan TN, Chua KB, Hooi PS. 2002. Seasonal variation in
respiratory syncytial virus chest infection in the tropics. Pediatr Pulmonol
34(1):47-51.
38 Chapter 2: Literature Review
Checkley W, Epstein L, Gilman R, Figueroa D, Cama R, Patz J, et al. 2000. Effect of
El Niño and ambient temperature on hospital admissions for diarrhoeal
diseases in Peruvian children. Lancet 355(9202):442-450.
Chou WC, Wu JL, Wang YC, Huang H, Sung FC, Chuang CY. 2010. Modeling the
impact of climate variability on diarrhea-associated diseases in Taiwan
(1996–2007). Sci Total Environ 409(1):43-51.
D'Souza RM, Becker N, Hall G, Moodie K. 2004. Does ambient temperature affect
foodborne disease? Epidemiology 15(1):86-92.
D'Souza, RM, Hall G, Becker NG. 2008. Climatic factors associated with
hospitalizations for rotavirus diarrhoea in children under 5 years of age.
Epidemiol Infect 136(1):56-64.
Djawe K, Levin L, Swartzman A, Fong S, Roth B, Subramanian A, et al. 2013.
Environmental risk factors for pneumocystis pneumonia hospitalizations in
HIV patients. Clin Infect Dis 56(1):74-81.
Drayna P, McLellan S, Simpson P, Li S, Gorelick M. 2010. Association between
rainfall and pediatric emergency department visits for acute gastrointestinal
illness. Environ Health Perspect 118(10):1439-1443.
Ebi K, Exuzides K, Lau E, Kelsh M, Barnston A. 2001. Association of normal
weather periods and El Niño events with hospitalization for viral pneumonia
in females: California, 1983-1998. Am J Public Health 91(8):1200-1208.
Garcia-Vidal C, Labori M, Viasus D, Simonetti A, Garcia-Somoza D, Dorca J, et al.
2013. Rainfall is a risk factor for sporadic cases of legionella pneumophila
pneumonia. PLoS ONE 8(4):e61036.
Chapter 2: Literature Review 39
Green R, Basu R, Malig B, Broadwin R, Kim J, Ostro B. 2010. The effect of
temperature on hospital admissions in nine California counties. Int J Public
Health 55(2):113-121.
Haffejee I. 1995. The epidemiology of rotavirus infections: a global perspective. J
Pediatr Gastroenterol Nutr 20(3):275-286.
Hashizume M, Armstrong B, Hajat S, Wagatsuma Y, Faruque AS, Hayashi T, et al.
2007. Association between climate variability and hospital visits for non-
cholera diarrhoea in Bangladesh: effects and vulnerable groups. Int J
Epidemiol 36(5):1030-1037.
Haynes AK, Manangan AP, Iwane MK, Sturm-Ramirez K, Homaira N, Brooks WA,
et al. 2013. Respiratory syncytial virus circulation in seven countries with
Global Disease Detection Regional Centers. J Infect Dis S3:S246-254.
Herrera-Lara S, Fernández-Fabrellas E, Cervera-Juan Á, Blanquer-Olivas R. 2013.
Do seasonal changes and climate influence the etiology of community
acquired pneumonia? Archivos de Bronconeumología (English Edition),
49(4), 140-145.
IPCC. 2007a. Climate change 2007: the physical science basis. Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge Cambridge University Press.
IPCC. 2007b. Summary for policymakers. In: Climate change 2007: the physical
science basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge. Cambridge: Cambridge University Press.
Jagai JS, Sarkar R, Castronovo D, Kattula D, McEntee J, Ward H, et al. 2012.
Seasonality of rotavirus in South Asia: A meta-analysis approach assessing
40 Chapter 2: Literature Review
associations with temperature, precipitation, and vegetation Index. PLoS
ONE 7(5):e38168.
Jofre J, Blanch A, Lucena F. 2010. Water-borne infectious disease outbreaks
associated with water scarcity and rainfall events. In S. Sabater & D. Barceló
(Eds.), Water Scarcity in the Mediterranean (pp. 147-159): Springer Berlin
Heidelberg.
Lam LT. 2007. The association between climatic factors and childhood illnesses
presented to hospital emergency among young children. Int J Environ Health
Res 17(1):1-8.
Levy K, Hubbard AE, Eisenberg JN. 2009. Seasonality of rotavirus disease in the
tropics: a systematic review and meta-analysis. Int J Epidemiol 38(6):1487-
1496.
Loh TP, Lai FY, Tan ES, Thoon KC, Tee NWS, Cutter J, et al. 2011. Correlations
between clinical illness, respiratory virus infections and climate factors in a
tropical paediatric population. Epidemiol Infect 139(12):1884-1894.
Murdoch DR, Jennings LC. 2009. Association of respiratory virus activity and
environmental factors with the incidence of invasive pneumococcal disease. J
Infect 58(1):37-46.
Omer SB, Sutanto A, Sarwo H, Linehan M, Djelantik, IGG, Mercer D, et al. 2008.
Climatic, temporal, and geographic characteristics of respiratory syncytial
virus disease in a tropical island population. Epidemiol Infect 136(10):1319-
1327.
Onozuka D, Hashizume M. 2010. Effects of weather variability on infectious
gastroenteritis. Epidemiol Infect 138(2):236-243.
Chapter 2: Literature Review 41
Onozuka D, Hashizume M, Hagihara A. 2009. Impact of weather factors on
Mycoplasma pneumoniae pneumonia. Thorax 64(6), 507-511.
Patel MM, Pitzer VE, Alonso WJ, Vera D, Lopman B, Tate J, et al. 2013. Global
seasonality of rotavirus disease. Pediatr Infect Dis J 32(4):e134-147.
Paynter S, Ware RS, Weinstein P, Williams G, Sly PD. 2010. Childhood pneumonia:
a neglected, climate-sensitive disease? Lancet 376(9755), 1804-1805.
Paynter S, Weinstein P, Ware RS, Lucero MG, Tallo V, Nohynek H, et al. 2013.
Sunshine, rainfall, humidity and child pneumonia in the tropics: time-series
analyses. Epidemiol Infect 141(6):1328-1336.
Pica N, Bouvier NM. 2014. Ambient temperature and respiratory virus infection.
Pediatr Infect Dis J 33(3):311-313.
Rudan I, O'Brien K, Nair H, Liu L, Theodoratou E, Qazi S, et al. 2013.
Epidemiology and etiology of childhood pneumonia in 2010: estimates of
incidence, severe morbidity, mortality, underlying risk factors and causative
pathogens for 192 countries. J Glob Health 3(1):10401.
Shaman J, Goldstein E, Lipsitch M. 2011. Absolute humidity and pandemic versus
epidemic influenza. Am J Epidemiol 173(2):127-135.
Shaman J, Kohn M. 2009. Absolute humidity modulates influenza survival,
transmission, and seasonality. Proc Natl Acad Sci U S A 106(9):3243-3248.
Souza AD, Fernandes W, Pavão H, Lastoria G, Edo AA. 2012. Potential impacts of
climate variability on respiratory morbidity in children, infants, and adults. J
Bras Pneumol 38(6):708-715.
Sumi a, Rajendran K, Ramamurthy T, Krishnan T, Nair GB, Harigane K, et al. 2013.
Effect of temperature, relative humidity and rainfall on rotavirus infections in
Kolkata, India. Epidemiol Infect 141(8):1652-1661.
42 Chapter 2: Literature Review
Tang JW, Lai FYL, Wong F, Hon KLE. 2010. Incidence of common respiratory viral
infections related to climate factors in hospitalized children in Hong Kong.
Epidemiol Infect 138(2):226-235.
Walker CLF, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. 2013. Global
burden of childhood pneumonia and diarrhoea. Lancet 381(9875):1405-1416.
Watson M, Gilmour R, Menzies R, Ferson M, McIntyre P, Network NS. 2006. The
association of respiratory viruses, temperature, and other climatic parameters
with the incidence of invasive pneumococcal disease in Sydney, Australia.
Clin Infect Dis 42(2): 211-215.
Wright DN, Bailey GD, Goldberg LJ. 1969. Effect of temperature on survival of
airborne Mycoplasma pneumoniae. J Bacteriol 99(2):491-495.
Xu YC, Zhu LJ, Xu D, Tao XF, Li SX, Tang LF, et al. 2011. Epidemiological
characteristics and meteorological factors of childhood Mycoplasma
pneumoniae pneumonia in Hangzhou. World J Pediatr 7(3):240-244.
Yusuf S, Piedimonte G, Auais A, Demmler G, Krishnan S, Van Caeseele P, et al.
2007. The relationship of meteorological conditions to the epidemic activity
of respiratory syncytial virus. Epidemiol Infect 135(7):1077-1090.
Zhang K, Oswald EM, Brown DG, Brines SJ, Gronlund CJ, White-Newsome JL, et
al. 2011. Geostatistical exploration of spatial variation of summertime
temperatures in the Detroit metropolitan region. Environ Res 111(8):1046-
1053.
Chapter 2: Results paper one 43
Chapter 3: Results paper one
The spatiotemporal patterns and geographic co-
distribution of childhood pneumonia and diarrhoea in
Queensland, Australia
Xu Z, Hu W, Tong S (2014). The geographic co-distribution and socio-ecological drivers of
childhood pneumonia and diarrhoea in Queensland, Australia. Epidemiology and Infection,
143(5):1096-104.
44 Chapter 3: Results paper one
Abstract
This study aimed to explore the spatiotemporal patterns and geographic co-distribution of
childhood pneumonia and diarrhoea in Queensland. A seasonal decomposition analysis was
conducted to assess the long-term trend and seasonality of childhood pneumonia and
diarrhoea. A spatial analysis was conducted to explore the spatial patterns of childhood
pneumonia and diarrhoea. A cluster analysis was also used to identify the high-risk clusters
and geographic co-distribution of childhood pneumonia and diarrhoea. The results suggest a
distinct seasonality of childhood pneumonia and diarrhoea. Childhood pneumonia and
diarrhoea mainly distributed in northwest of Queensland. Mount Isa was the high-risk cluster
where childhood pneumonia and diarrhoea co-distributed. Future pneumonia and diarrhoea
prevention and control measures in Queensland should focus more on Mount Isa.
Chapter 2: Results paper one 45
3.1 Introduction
Pneumonia and diarrhoea are the leading causes of child mortality (Liu et al. 2012). In 2011,
two million children died before reaching their fifth birthday because of pneumonia and
diarrhoea worldwide (Walker et al. 2013). While the incidences of mortality due to
pneumonia and diarrhoea have been declining in some industrialized countries, they are still
an important source of morbidity in these regions (Podewils et al. 2004). In Western Pacific
region, the total episodes of pneumonia and diarrhoea in children under five years of age
were 256.3 million and 12.2 million in 2010, respectively (Walker et al. 2013).
Pneumonia and diarrhoea are largely preventable, and hence it is essential to identify the risk
factors and take targeted preventive measures (Bhutta et al. 2013). Existing studies have
confirmed some individual-level biological factors for pneumonia and diarrhoea, such as
underweight, stunting and zinc deficiency (Walker et al. 2013). Some of these poverty-related
risk factors, such as suboptimum breastfeeding, under-nutrition and zinc deficiency, are
shared by pneumonia and diarrhoea, and these overlapping risk factors may result in the
geographic co-distribution of pneumonia and diarrhoea (Walker et al. 2013).
Australia shoulders a considerable burden of childhood pneumonia and diarrhoea (Rudan et
al. 2013; Scallan et al. 2005). It is urgently needed to reveal the spatiotemporal patterns of
childhood pneumonia and diarrhoea in Australia. This study explored the spatiotemporal
patterns and geographic co-distribution of childhood pneumonia and diarrhoea in
Queensland, Australia.
3.2 Methods
Data collection
Queensland is located in the northeast of Australia. Its mean temperature of summer is 25 °C
and mean temperature of winter is 15 °C. There is a significant variation in mean annual
46 Chapter 3: Results paper one
rainfall across Queensland, varying from less than 150 mm/year in the southwest to 4000
mm/year in the northern coast (Australian Bureau of Statistics 2010). Data on emergency
department visits (EDVs) by postcode from January 1st 2007 through 31st December 2011 in
Queensland were obtained from Queensland Health. These data were collected from
emergency departments of hospitals and rural emergency departments of most Queensland
public facilities through the Emergency Department Information System (EDIS) (Toloo, et
al., 2012).The anonymised EDV data were classified according to the International
Classification of Disease, 10th version (ICD–code10). In this study, we included EDVs with
the principle cause coded as pneumonia (ICD–10 codes: J12–J18) and diarrhoeal disease of
any cause (ICD–10 codes: A00–A03, A04, A05, A06.0–A06.3, A06.9, A07.0–A07.2, A07.9,
A08–A09) among children aged 0–14 years. Ethical approval was obtained from the Human
Research Ethics Committee of Queensland University of Technology (Australia) prior to the
data being collected (number: 1000001168). Patient information was de-identified and thus
no written informed consent was obtained.
Statistical analysis
We plotted the decomposed daily distributions of EDVs for childhood pneumonia and
diarrhoea using a time-series approach. The change of EDVs for childhood pneumonia and
diarrhoea from 2008-2009 to 2010-2011 was calculated using the following equation:
2010 2011 2008 2009( ) /i i iMc EDV EDV population− −= −
Where Mc represents the morbidity change, 2010 2011iEDV − represents the EDVs for childhood
pneumonia (diarrhoea) for postal area i during 2010-2011, 2008 2009iEDV − represents the EDVs
for childhood pneumonia (diarrhoea) for postal area i during 2008-2009, and
ipopulation refers to the population for postal area i.
Chapter 2: Results paper one 47
Time-series analysis was conducted using the R statistical environment, version 2.15.3.
Visual maps were created using ArcGIS version 9.3 (ESRI Inc., Redlands, CA, USA), and
spatial cluster analysis was conducted using SatScan version 9.1.
3.3 Results
Summary statistics
Table 3-1 presents the summary statistics of EDVs for childhood pneumonia and diarrhoea
by postcode in Queensland. The average counts of childhood pneumonia and diarrhoea were
43.7 and 135.8, respectively.
Table 3-1. Summary statistics for EDVs for childhood pneumonia and diarrhoea by postcode
in Queensland, Australia, during 2007-2011
Variables Mean SD Min Max
Pneumonia (cases) 43.7 79.5 0 739
Diarrhoea (cases) 135.8 247.7 0 1750
48 Chapter 3: Results paper one
Spatial pattern
Figure 3-1 shows the spatial distribution of rates of EDVs for childhood pneumonia and
diarrhoea, illustrating that EDVs for both pneumonia were the highest in central west,
northwest and far north of Queensland, and the EDVs for childhood diarrhoea were the
highest in the northwest of Queensland (Mount Isa). Figure 3-2 and Figure 3-3 show the
spatial distribution of rates of EDVs for childhood pneumonia by age and gender. Figure 3-4
and Figure 3-5 show the spatial distribution of rates of EDVs for childhood diarrhoea by age
and gender. No significant differences between two age groups and genders were observed in
terms of both pneumonia and diarrhoea spatial patterns. Figure 3-6 illustrates the change in
EDVs for childhood pneumonia and diarrhoea from years 2008-2009 to 2010-2011,
indicating that EDVs for pneumonia and diarrhoea changed from northwest or southeast of
Queensland in the past couple of years.
Temporal pattern
Figure 3-7 shows the decomposed daily distributions of EDVs for childhood pneumonia and
diarrhoea, showing a distinct seasonal trend for the two diseases, especially for pneumonia.
This figure indicates that EDVs for childhood pneumonia in Queensland were more likely to
occur in cold season. The particularly great number of EDVs for childhood pneumonia in
2009 is because of the 2009 pandemic H1N1 influenza.
Geographical co-distribution
The cluster results in Figure 3-8 reveal that EDVs for childhood pneumonia and diarrhoea in
Queensland were co-distributed in Mount Isa.
Spatial patterns of climatic factors
As the coming chapters will talk about the effects of climatic factors on childhood pneumonia
and diarrhoea, we presented the spatial patterns of mean temperature and rainfall in
Queensland in Figure 3-9. The results show that mean temperature in northwest of
Chapter 2: Results paper one 49
Queensland was higher than other places. No significant spatial pattern of rainfall was
detected.
50 Chapter 3: Results paper one
Figure 3-1. The spatial distribution of EDVs for childhood pneumonia and diarrhoea in Queensland, from 2007 to 2011
Pneumonia is on the left side, and diarrhoea is on the right side.
Chapter 2: Results paper one 51
Figure 3-2. The spatial distribution of EDVs for childhood pneumonia by age in Queensland, from 2007 to 2011
52 Chapter 3: Results paper one
Figure 3-3. The spatial distribution of EDVs for childhood pneumonia by gender in Queensland, from 2007 to 2011
Chapter 2: Results paper one 53
Figure 3-4. The spatial distribution of EDVs for childhood diarrhoea by age in Queensland, from 2007 to 2011
54 Chapter 3: Results paper one
Figure 3-5. The spatial distribution of EDVs for childhood diarrhoea by gender in Queensland, from 2007 to 2011
Chapter 2: Results paper one 55
Figure 3-6. The change of EDVs for childhood pneumonia and diarrhoea in Queensland, from 2008-2009 to 2010-2011
Pneumonia is on the left side, and diarrhoea is on the right side.
56 Chapter 3: Results paper one
Figure 3-7. The daily distribution of EDVs for childhood pneumonia and diarrhoea in Queensland, from 2007 to 2011
Pneumonia is on the left side, and diarrhoea is on the right side.
Chapter 2: Results paper one 57
Figure 3-8. The spatial clusters of EDVs for childhood pneumonia and diarrhoea in Queensland, from 2007 to 2011
Pneumonia is on the left side, and diarrhoea is on the right side.
58 Chapter 3: Results paper one
Figure 3-9. The spatial patterns of mean temperature and rainfall in Queensland, from 2007 to 2011
Chapter 3: Results paper one 59
3.4 Discussion
This study has yielded several notable findings. There was a strong seasonal trend in
EDVs for childhood pneumonia, with more cases occurring in cold season. Children
suffering pneumonia and diarrhoea who visited emergency departments in
Queensland from 2007 to 2011 were mainly from central west, northwest and far
north of Queensland. According to the cluster analysis results, Mount Isa was the
high risk area for both childhood pneumonia and diarrhoea. Interestingly, in recent
years, Mount Isa has been experiencing a substantial decrease in EDVs for childhood
pneumonia and diarrhoea, and EDVs for childhood pneumonia and diarrhoea were
moving from west to southeast of Queensland.
Several reasons, such as nutritional factors (Vitamin A or Zinc deficiency) (Black et
al.), poverty (Fonseca et al. 1996), and Indigenous status (Janu et al. 2014), may
contribute to the high EDVs for childhood pneumonia and diarrhoea in central west,
northwest and far north of Queensland. Mount Isa city, a major lead, zinc and copper
producer, is the largest emitter of sulphur dioxide, lead and some other metals in
Australia (National Pollutant Inventory 2010). It has been convincingly documented
that the blood lead level of children in Mount Isa, especially for those aged 1–4
years, is way higher than children from other regions of Australia (Queensland
Health, 2009), and the consequent life-long negative health and intellectual impacts
of lead exposure on children have also been extensively reported (Lanphear et al.
2005; Tong et al. 1998). In this study, we found that pneumonia and diarrhoea in
children were co-distributed in Mount Isa, highlighting that there might be some
common risk factors in this area. Exposure to air pollutants (e.g., sulphur dioxide)
emitted by mining could increase hospital admissions for childhood pneumonia
(Barnett et al. 2005). Mining also had a significant adverse effect on semi-arid
60 Chapter 3: Results paper one
freshwater aquatic system in Mount Isa (Taylor et al. 2009). The densities of
bacterial indicators in remnant pools throughout Leichhardt River have exceeded the
acceptable guidelines, which might expose children to greater risk of diarrhoea. In
this study, we also found the risk areas for childhood pneumonia and diarrhoea
changed from northwest to southeast of Queensland, and the EDVs for childhood
pneumonia and diarrhoea in Mount Isa have been decreasing sharply (though still
high) in recent years, indicating that protective measures may have been taken to
prevent children from continuously adverse impacts of mining. Previous studies have
also proposed that the high morbidity of childhood pneumonia and diarrhoea in
Mount Isa may be attributable to prematurity and intrauterine growth restriction,
under-nutrition, exposure to cigarette smoke, aspiration of infected nasopharyngeal
secretions and social factors of overcrowded living conditions and poor hygiene
(Janu et al. 2014). Many of the children in Mount Isa are members of common
extended families, raising the question of the role of inherited polymorphisms in
genes involved in immunity or inflammation (Thiel et al. 2009).
To the best of our knowledge, this is the first study to explore the geographic co-
distribution of childhood pneumonia and diarrhoea. The results from this study,
especially the high risk areas of pneumonia and diarrhoea we identified, may have
important implications for future control and prevention for childhood pneumonia
and diarrhoea in Queensland. A major limitation of this study is that the disease data
we collected from emergency departments may underestimate the actual infected
population because only children with severe symptoms would go to emergency
departments for treatment. Data on EDVs by postcode is not ideal in doing spatial
analysis, and we are collecting data on EDVs by Statistical Local Areas (SLA) or
Local Government Areas (LGA) for our future studies.
Chapter 3: Results paper one 61
3.5 Conclusions
Childhood pneumonia and diarrhoea were predominantly distributed in northwest of
Queensland, and Mount Isa was the region where these two childhood diseases co-
distributed. In recent years, the high risk areas of these two childhood diseases have
been changing from northwest to southeast of Queensland.
3.6 References
Australian Bureau of Statistics. 2010. Census Data. Available
http://www.abs.gov.au/CDataOnline Retrieved September 24th, 2013
Barnett AG, Williams GM, Schwartz J, Neller AH, Best TL, Petroeschevsky AL, et
al. 2005. Air pollution and child respiratory health. Am J Respir Crit Care
Med 171(11):1272-1278.
Bhutta ZA, Das JK, Walker N, Rizvi A, Campbell H, Rudan I, et al. 2013.
Interventions to address deaths from childhood pneumonia and diarrhoea
equitably: what works and at what cost? Lancet 381(9875):1417-1429.
Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M et al. 2008.
Maternal and child undernutrition: global and regional exposures and health
consequences. Lancet 371(9608):243-260.
Fonseca W, Kirkwood BR, Victora CG, Fuchs SR, Flores JA, Misago C. 1996. Risk
factors for childhood pneumonia among the urban poor in Fortaleza, Brazil: a
case-control study. Bull World Health Organ 74(2): 199-208.
Janu EK, Annabattula BI, Kumariah S, Zajaczkowska M, Whitehall JS, Edwards MJ
et al. 2014. Paediatric hospitalisations for lower respiratory tract infections in
Mount Isa. Med J Aust 200(10): 591-594.
62 Chapter 3: Results paper one
Lanphear B, Hornung R, Khoury J, Yolton K, Baghurst P, Bellinger D, et al. 2005.
Low-level environmental lead exposure and children's intellectual function:
an international pooled analysis. Environ Health Perspect 113(7):894-899.
Liu L, Johnson HL, Cousens S, Perin J, Scott S, Lawn JE, et al. 2012. Global,
regional, and national causes of child mortality: an updated systematic
analysis for 2010 with time trends since 2000. Lancet 379(9832):2151-2161.
Podwils LJ, Mintz ED, Nataro JP, Parashar UD. 2004. Acute, infectious diarrhea
among children in developing countries. Semin Pediatr Infect Dis 15(3):155-
168.
National Pollutant Inventory (2010).
http://www.npi.gov.au/npidata/action/load/individual-facility-
detail/criteria/year/2009/browse-type/Company/regbusiness-
name/MOUNT%2BISA%2BMINES%2BLTD/jurisdiction-
facility/Q020MIM001. Retrieved October 1st 2013
Podewils LJ, Mintz ED, Nataro JP, Parashar UD. 2004. Acute, infectious diarrhea
among children in developing countries. Semin Pediatr Infect Dis 15(3):155-
168.
Queensland Health. 2009. Mount Isa community lead screening program 2006–07: a
report into the results of a blood-lead screening program of 1–4 year old
children in Mount Isa, Queensland.
http://www.health.qld.gov.au/ph/documents/tphn/mtisa_leadrpt.asp
Retrieved October 1st 2013
Rudan I, O'Brien K, Nair H, Liu L, Theodoratou E, Qazi S, et al. 2013.
Epidemiology and etiology of childhood pneumonia in 2010: estimates of
Chapter 3: Results paper one 63
incidence, severe morbidity, mortality, underlying risk factors and causative
pathogens for 192 countries. J Glob Health 3(1):10401.
Scallan E, Majowicz SE, Hall G, Banerjee A, Bowman CL, Daly L, et al. 2005.
Prevalence of diarrhoea in the community in Australia, Canada, Ireland, and
the United States. Int J Epidemiol 34(2):454-460.
Taylor MP, Mackay A, Kuypers T, Hudson-Edwards K. 2009. Mining and urban
impacts on semi-arid freshwater aquatic systems: the example of Mount Isa,
Queensland. J Environ Monit 11(5):977-986.
Thiel S, Kolev M, Degn S, Steffensen R, Hansen AG, Ruseva M, et al. 2009.
Polymorphisms in mannan-binding lectin (MBL)-associated serine protease 2
affect stability, binding to MBL, and enzymatic activity. J Immunol
182(5):2939-2947.
Toloo S, Rego J, FitzGerald G, Aitken P, Ting J, Quinn J, et al. 2012. Emergency
Health Services (EHS): Demand and Service Delivery Models. Queensland
University of Technology.
Tong S, Baghurst P, Sawyer MG, Burns J, McMichael AJ. 1998. Declining blood
lead levels and changes in cognitive function during childhood: The port pirie
cohort study. JAMA 280(22):1915-1919.
Walker CLF, Perin J, Katz J, Tielsch J, Black R. 2013. Diarrhea as a risk factor for
acute lower respiratory tract infections among young children in low income
settings. J Glob Health 3(1):10402.
Walker CLF, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. 2013. Global
burden of childhood pneumonia and diarrhoea. Lancet 381(9875):1405-1416.
Chapter 4: Results paper two 65
Chapter 4: Results paper two
Impact of temperature on childhood pneumonia estimated
from satellite remote sensing
Xu Z, Liu Y, Ma Z, Li S, Hu W, Tong S (2014). Impact of temperature on childhood
pneumonia estimated from satellite remote sensing. Environmental Research, 132:334-341.
66 Chapter 4: Results paper two
Abstract
The effect of temperature on childhood pneumonia in subtropical regions is largely unknown
so far. This study examined the impact of temperature on childhood pneumonia in Brisbane,
Australia. A quasi-Poisson generalized linear model combined with a distributed lag non-
linear model was used to quantify the main effect of temperature on emergency department
visits (EDVs) for childhood pneumonia in Brisbane from 2001 to 2010. The model residuals
were checked to identify added effects due to heat waves or cold spells. Both high and low
temperatures were associated with an increase in EDVs for childhood pneumonia. Children
aged 2–5 years (excluding 5 years), and female children were particularly vulnerable to the
impacts of heat and cold. Indigenous children were sensitive to heat. Heat waves and cold
spells had significant added effects on childhood pneumonia, and the magnitude of these
effects increased with intensity and duration. There were changes over time in both the main
and added effects of temperature on childhood pneumonia. Children, especially those female
and Indigenous, should be particularly protected from extreme temperatures. Future
development of early warning systems should take the change over time in the impact of
temperature on children’s health into account.
Chapter 4: Results paper two 67
4.1 Introduction
Climate change has been widely recognized as the biggest health threat in the 21st century
(McMichael 2013), and its possible impact on infectious disease has attracted public health
attention (Altizer et al. 2013). Children, particularly their respiratory system (Sheffield and
Landrigan 2011), are vulnerable to the adverse impact of climate change (McKie 2013).
Pneumonia, the leading killer of children, has been reported responsible for 1.3 million deaths
in children aged under five years in 2011 (Walker et al. 2013), and the global burden of
childhood pneumonia may continue to rise due to the Earth’s increasing average surface
temperature (Walker et al. 2013), though the true scale of the association between
temperature and childhood pneumonia is largely unknown.
Persistent extreme temperatures (i.e., heat waves and cold spells) occur across the globe and
heat waves are projected to become more frequent and intense in the future (Meehl and
Tebaldi 2004), posing a huge challenge to children’s well-being (Xu et al. 2014). Existing
literature indicates that the effects of persistent extreme temperatures on human health can be
attributable to the independent effects of daily ambient temperature (main effect) and of
persistent periods of heat and cold (added effect) (Anderson and Bell 2009; Hajat et al. 2006).
During periods of persistent extreme temperatures, children are more likely to stay indoors,
which may increase crowding and their exposure to biomass fuel smoke from cooking,
possibly resulting in a higher risk of getting pneumonia. However, to our best knowledge,
few data are available on the effects of heat waves or cold spells on childhood pneumonia,
and no study has examined whether heat waves or cold spells have an added effect on
childhood pneumonia.
Epidemiological studies examining the effect of temperature on health tend to use
temperature from one ground-monitoring site or the average from several ground-monitoring
sites, which might result in measurement bias or exposure misclassification, especially for
68 Chapter 4: Results paper two
those areas without extensive monitoring sites, because temperature across one city is
spatially variable (Zhang et al. 2011), and temperatures in urban areas are normally higher
than those in rural areas because of the urban heat island (Laaidi et al. 2012). Satellite remote
sensing data can substantially supplement ground monitoring networks to quantify the effect
of exposure to environmental hazards on health (Wang et al. 2013). The fundamental bias
satellite remote sensing data reduces is exposure error reduction due to better coverage and
higher spatial resolution. If using weather station data, researchers would probably need to
draw a buffer and assign everybody's temperature exposure to the readings at this central
station. Satellite data, on the other hand, are gridded at a pretty high spatial resolution so that
the exposure estimates can be more accurate. In addition, land surface temperature is
different from air temperature in that it considers the impact of direct solar radiation and the
surface long-wave radiation, so someone will feel hotter under the sun than in the shade even
though the difference in air temperature between under the sun and in the shade is smaller,
and thus it can be strongly related to heat-related morbidity and mortality. Although satellite
remote sensing data have been successfully used to link the relationship between air pollution
and acute health outcomes (Evans et al. 2013; Wang et al. 2013), it has been scarcely applied
to assess the impact of temperature on human health (Estes et al. 2009).
This study used the data on satellite remote sensing temperature and emergency department
visits (EDVs) for childhood pneumonia in Brisbane, Australia, from 2001 to 2010 and aimed
to minimize the measurement bias and answer three research questions: i) What is the
relationship between temperature and EDVs for childhood pneumonia? ii) Is there any added
effect due to heat waves and cold spells? iii) Whether there is any significant change over
time in the effect of temperature on childhood pneumonia across the study period?
4.2 Methods
Chapter 4: Results paper two 69
Data collection
Health data
Brisbane is the capital city of Queensland, Australia. It has a subtropical climate and rarely
experiences very cold temperatures. The daily EDV data from January 1st 2001 to December
31st 2010 classified according to the International Classification of Diseases, 9th version
and10th version (ICD 9 and 10), were obtained from Queensland Health. These data were
originally collected from emergency departments of hospitals and rural emergency
departments of most Queensland public facilities. Specifically, the Brisbane data we used
were from Caboolture Hospital, Ipswich Hospital, Logan Hospital, Mater Children’s Public
Hospital, Princess Alexandra Hospital, Queen Elizabeth II Jubilee Hospital, Redcliffe
Hospital, Redland Hospital, Royal Brisbane and Women’s Hospital, Royal Children’s
Hospital and the Prince Charles Hospital. These hospitals cover the Greater Brisbane region
well. Those coded as pneumonia (ICD 9 codes: 480–486; ICD 10 codes: J12–J18) in children
aged 0–14 years were selected.
Ground-monitoring data
Daily weather data, including rainfall and relative humidity, were supplied by the Australian
Bureau of Meteorology. Data on air pollutants, including daily average particular matter ≤
10µm (PM10) (µg/m3), daily average nitrogen dioxide (NO2) (µg/m3) and daily average ozone
(O3) (ppb), were obtained from the Queensland Department of Environment and Heritage
Protection (former Queensland Environmental Protection Agency).
Satellite remote sensing temperature data
Land surface temperature (LST) is the mean radiative skin temperature of an area of land
resulting from the energy balance between solar heating and land-atmosphere cooling. LST is
more closely related to the physiological activities of leaves, soil moisture, and near-surface
meteorology. Therefore, it has stronger spatial heterogeneity imposed by landscape variations
70 Chapter 4: Results paper two
than air temperature. The Moderate Resolution Imaging Spectroradiometer (MODIS)
instruments were launched into low Earth polar orbits aboard the Naitonal Aerospace and
Space Administration (NASA)'s Terra and Aqua satellites in 1999 and 2002,
respectively(Anderson et al. 2005; Kaufman et al. 1998). They cross the equator at around
10:30 a.m. and 1:30 p.m. local time, respectively. MODIS LST was retrieved based on a
split-window algorithm that corrects for atmospheric effects based on the differential
absorption in MODIS's two adjacent infrared bands (bands 31 and 32) (Wan and Dozier
1996). Version 5 MODIS LST data have been extensively validated globally, showing that
the accuracy of the MODIS LST product is better than 1 K in most cases (Wan 2008; Wan et
al. 2002). For the current study, Level 3 MODIS Land Surface Temperature data (MOD11B1
for Terra from 2001 to 2010 and MYD11B1 for Aqua from 2002 to 2010) at 6 km spatial
resolution were downloaded from NASA's Level 1 and Atmospheric Archive and
Distribution System (http://ladsweb.nascom.nasa.gov) (Figure 4-1). Each data file contains
both a day time (~10:30 am for Terra, ~1:30 pm for Aqua) and a night time (~10:30 pm for
Terra, and ~1:30 am for Aqua) LST measurement. These LST values retrieved from the two
satellites were averaged to get the daily mean temperature (satellite remote sensing
temperature). Similar approaches have been applied to merge the LST data from Terra and
Aqua in the United States (Crosson et al. 2012).
Chapter 4: Results paper two 71
Figure 4-1. The areas where satellite remote sensing temperature data were collected
72 Chapter 4: Results paper two
Data analysis
Heat waves and Cold Spells
There is no consistent definition for heat waves or cold spells. We combined temperature
duration and intensity to define heat waves and cold spells: (1) the 1st and 5th percentiles of
daily mean temperature were defined as the cold threshold, and the 95th and 99th percentiles
of the daily mean temperature as the heat threshold; and (2) a minimum of 2–4 consecutive
days with temperatures below the cold threshold or above the heat threshold were required.
Stage I: Estimating the main temperature effects
A quasi-Poisson generalized linear regression model combined with a distributed lag non-
linear model (DLNM) was used to quantify the effect of temperature on EDVs for childhood
pneumonia (Xu et al. 2013b). A natural cubic spline with four degrees of freedom (df) was
used to capture a potentially non-linear temperature effect. A lag of 21 days was used to
quantify the lagged effect of temperature (Xu et al. 2013a). Rainfall and relative humidity
were controlled for by using a natural cubic spline with four df. NO2, PM10 and O3 were
controlled for using a linear function. Seasonal patterns and long-term trends were controlled
by using a natural cubic spline with six df per year of data. Day of week was controlled as a
categorical variable. Influenza epidemics and public holiday were also controlled for in the
model.
Yt ~ Poisson(μt)
Log (μt) = α + β1Tt,l + ns(RHt, 4) + ns(Rainfallt, 4) + β2PM10t + β3O3t + β4NO2t
+ns(Timet,6)+ β5Day of Weekt + β6Influenzat + β7Holidayt
Where t is the day of the observation; Yt is the observed daily EDVs for childhood pneumonia
on day t; α is the model intercept; Tt,l is a matrix obtained by applying the DLNM to
temperature; β1 is vector of coefficients for Tt,l, and l is the lag days; ns(RHt, 4) is a natural
cubic spline with four degree of freedom for relative humidity; ns(Rainfallt, 4) is a natural
Chapter 4: Results paper two 73
cubic spline with four degree of freedom for rainfall; PM10t, O3t, and NO2t are the
concentrations of PM10, O3, and NO2 on day t; ns(Timet,6) is a natural cubic spline with six
degrees of freedom for seasonality and long-term trend; Day of Weekt is the categorical
variable as EDVs varied with week days and weekends; Influenzat is the number of lab-
confirmed influenza cases on day t; Holidayt is a binary variable which is “1” if day t was a
holiday.
We checked the temperature–pneumonia plot and chose the temperature corresponding to the
lowest risk as the reference temperature. We quantified the relative risk of EDVs for
childhood pneumonia associated with high temperature (29.6°C, 99th centile of mean
temperature) relative to the reference temperature (chosen to be 23.0°C). Similarly, we
calculated the relative risk of EDVs for childhood pneumonia associated with low
temperature (9.8°C, 1st centile of mean temperature) relative to the reference temperature
(23.0°C). We specifically examined the association between temperature and EDVs for
childhood pneumonia for every five years (2001–2005, 2002–2006, 2003–2007, 2004–2008,
2005–2009 and 2006–2010) to test whether there was any change over time in this
association.
Stage II Examining the added effects of heat waves and cold spells
We used the residuals of stage I as the dependent variable of stage II model to quantify the
possible added effect of heat waves and cold spells, meaning that the main effect of
temperature has been removed in stage II (Xu et al. 2013a). We assumed a maximum lag of
21 days for examining the lagged effects of heat waves and cold spells. EDVs for childhood
pneumonia on days of heat waves and cold spells were compared with those non-extreme
temperature days.
Log (μt) = Log (μt1) +β1Ct,l + β2Ht,l t=1,2,….,n
74 Chapter 4: Results paper two
Where Log (μt1) is the estimated EDVs for childhood pneumonia counts on day t from the
stage-I model; Ct,l is a matrix applying DLNM to cold spells; and Ht,l is a matrix applying
DLNM to heat waves.
All data analysis was conducted using R (V 2.15), and “dlnm” package was used to fit the
regression model. Sensitivity analysis was conducted by adjusting the dfs for temperature and
time to assess the robustness of model choices.
4.3 Results
Summary statistics
Table 4-1 shows the summary statistics of daily climatic variables, air pollutants, influenza
and EDVs for childhood pneumonia. The mean value of satellite remote sensing temperature
was 19.8 °C. There were 17,238 EDVs for childhood pneumonia, with a daily mean of 4.7
cases. Figure 4-2 plots the EDVs for childhood pneumonia (decomposed), weather variables,
and air pollutants, showing a strong seasonal pattern of EDVs for childhood pneumonia,
satellite remote sensing temperature, O3 and NO2.
Chapter 4: Results paper two 75
Table 4-1. Summary statistics for climatic variables, air pollutants and paediatric pneumonia in Brisbane, Australia, 2001–2010
Variables Mean SD Min
Percentile
Max
25 50 75
RS mean temperature (°C)* 19.8 4.6 6.5 16.2 19.9 23.3 32.6
Relative humidity (%) 65.0 15.0 13.0 56.0 65.0 75.0 100.0
Rainfall (mm) 2.2 8.3 0 0 0 0.4 149.0
O3 (ppb) 13.4 4.6 1.7 10.2 12.8 16.0 34.2
PM10 (µg/m3) 16.1 18.4 3.9 11.5 14.3 17.8 960.0
NO2 (µg/m3) 7.2 4.3 0 4.0 6.3 9.8 25.3
Influenza 1.6 2.1 0 0.6 1.3 2.0 28
Pneumonia 4.7 4.1 0 2 4 6 53
*RS: remote sensing
76 Chapter 4: Results paper two
010
2030
4050
data
-20
24
6
seas
onal
34
56
78
trend
-10
010
2030
402002 2004 2006 2008 2010
rem
aind
er
time
Figure 4-2a. The decomposed distribution of EDVs for paediatric pneumonia in Brisbane,
from 2001 to 2010
Chapter 4: Results paper two 77
1015
2025
3035
rsm
eant
emp
050
100
150
rain
fall
2040
6080
100
2002 2004 2006 2008 2010
hum
idity
Time
Figure 4-2b. The daily distributions of climate variables in Brisbane, from 2001 to 2010
78 Chapter 4: Results paper two
020
040
060
080
0
pm10
510
1520
2530
35
o30
510
1520
25
2002 2004 2006 2008 2010
no2
Time
Figure 4-2c. The daily distributions of air pollutants in Brisbane, from 2001 to 2010
Chapter 4: Results paper two 79
Table 4-2 indicates the Spearman correlation between weather variables, air pollutants and
EDVs for childhood pneumonia. Childhood pneumonia was negatively correlated with
temperature and relative humidity, but positively correlated with air pollutants. Figure 4-3
reveals the scatter plots of pneumonia and weather variables.
Table 4-2. Spearman’s correlation between daily weather variables, air pollutants and paediatric pneumonia in Brisbane, Australia, from 2001–2010
RS mean temperature†
Relative humidity Rainfall PM10 O3 NO2 Pneumonia
RS mean temperature† 1.00
Relative humidity -0.25* 1.00
Rainfall -0.03 0.38* 1.00
PM10 0.30* -0.29* -0.31* 1.00
O3 0.13* -0.28* -0.07* 0.31* 1.00
NO2 -0.55* 0.13* -0.14* 0.01 -0.10* 1.00
Pneumonia -0.32* 0.03 0.06* 0.07* 0.10* 0.28* 1.00
* P<0.01; † RS, remote sensing
80 Chapter 4: Results paper two
pneumo
10 15 20 25 30 35 20 40 60 80 100
010
2030
4050
1015
2025
3035
rsmean
rainfall
050
100
150
0 10 20 30 40 50
2040
6080
100
0 50 100 150
humidity
Figure 4-3. The pairwise plot of paediatric pneumonia, mean temperature, rainfall and
relative humidity in Brisbane, from 2001 to 2010
Chapter 4: Results paper two 81
Effect of temperature on EDVs for childhood pneumonia
The overall effect of temperature on EDVs for childhood pneumonia is presented in Figure 4-
4. EDVs for childhood pneumonia increased in both low and high temperatures. The impacts
of temperature on age-, gender- and ethnicity-specific EDVs for childhood pneumonia are in
Table 4-3. Children aged 2–5 years and female children appeared particularly vulnerable to
the temperature effect. Interestingly, Indigenous children were more sensitive to the heat
effect, and non-Indigenous children were more vulnerable to the cold effect.
10 15 20 25 30 35
12
34
56
78
Remote sensing temperatu
RR
(ped
iatri
c pn
eum
onia
)
Figure 4-4. The overall effect of mean temperature on paediatric pneumonia in Brisbane,
from 2001 to 2010
82 Chapter 4: Results paper two
Table 4-3. The cumulative effect of high and low temperatures on EDVs for paediatric pneumonia, with 99th percentile (29.6 °C) and 1st percentile (10.4 °C) of temperature relative to reference temperature (23°C)
Diseases Heat effect (Relative risk (95% CI)) Cold effect (Relative risk (95% CI))
Lag 0–1 Lag 0–13 Lag 0–21 Lag 0–1 Lag 0–13 Lag 0–21
All ages 1.07(0.92,1.26) 1.33(0.93,1.91) 1.72(1.07,2.76)* 0.89(0.77,1.03) 1.30(0.91,1.86) 2.76(1.71,4.47)*
(0,1) 0.95(0.62,1.44) 1.02(0.42,2.49) 0.99(0.31,3.18) 1.10(0.77,1.56) 1.08(0.43,2.69) 1.87(0.55,6.32)
[1,2) 0.94(0.69,1.29) 1.17(0.57,2.40) 1.47(0.58,3.72) 0.92(0.71,1.21) 1.33(0.72,2.81) 1.91(0.76,4.76)
[2,5) 1.08(0.85,1.37) 1.27(0.74,2.18) 2.04(1.01,4.11)* 0.78(0.63,1.01) 1.37(0.80,2.36) 3.03(1.47,6.26)
[5,14] 1.18(0.89,1.57) 1.42(0.72,2.78) 1.14(0.46,2.81) 1.00(0.76,1.31) 1.46(0.73,2.88) 1.76(0.83,3.71)
Male 0.96(0.76,1.21) 1.11(0.66,1.86) 1.35(0.69,2.64) 0.88(0.72,1.07) 1.11(0.67,1.82) 2.18(1.11,4.25)*
Female 1.17(0.96,1.42) 1.50(0.95,2.36) 2.02(1.11,3.66)* 0.90(0.75,1.08) 1.48(0.94,2.35) 3.36(1.81,6.24)*
Indigenous 1.13(0.58,2.21) 1.92(0.74,4.98) 4.20(1.03,8.16)* 0.80(0.45,1.41) 0.79(0.20,3.55) 1.46(0.23,9.25)
Non-indigenous 1.07(0.91,1.26) 1.25(0.87,1.80) 1.57(0.97,2.54) 0.90(0.78,1.04) 1.36(0.94,1.96) 2.89(1.77,4.71)*
*P-value<0.05
Chapter 4: Results paper two 83
The added effects of heat waves and cold spells
The daily excess EDVs for childhood pneumonia on heat wave days and cold spell days are
in Table 4-4. Using the heat wave definitions of two, three or four days with the temperature
over the 95th centile, we did not find any significant added effect of heat waves on EDVs for
childhood pneumonia. While, using the temperature over the 99th centile as the temperature
cut-off, we found there were significant added effects of heat waves on EDVs for childhood
pneumonia, and the EDVs due to added effect increased from three to seven when the heat
wave duration increased from two to three consecutive days. A significant increase in EDVs
for childhood pneumonia during cold spells was found while using the definition of four days
with the temperature over the 95th centile.
84 Chapter 4: Results paper two
Table 4-4. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2001 to 2010
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 58 0 (-1,1) ≤5th 52 1 (-1,2)
≥99th 6 3 (1,5) ≤1st 2 -1 (-6,4)
≥3
≥95th 31 0 (-1,2) ≤5th 22 1 (-1,3)
≥99th 2 7 (1,13) ≤1st - - -
≥4
≥95th 15 0 (-1,2) ≤5th 12 3 (1,5)
≥99th - - - ≤1st - - -
Chapter 4: Results paper two 85
Change over time in the effect of temperature on childhood pneumonia
The change of temperature effect on EDVs for childhood pneumonia over time can be seen in
Figure 4-5. The effect of high temperature on EDVs for childhood pneumonia experienced a
decreasing trend, while low temperature impact on EDVs for childhood pneumonia
experienced an increasing trend.
Figure 4-5. The change over time in the temperature effect on childhood pneumonia
Left hand side: hot effect; right hand side: cold effect; p1= 2001-2005, p2=2002-2006, p3=2003-2007, p4=2004-2008, p5=2005-2009, p6=2006-2010.
86 Chapter 4: Results paper two
Table 4-5 presents the added effect of heat waves and cold spells on EDVs for childhood
pneumonia during the six periods (2001–2005, 2002–2006, 2003–2007, 2004–2008, 2005–
2009 and 2006–2010), revealing that the added effect of heat waves and cold spells varied
greatly over time. The statistically significant added effect of cold spells on EDVs for
childhood pneumonia occurred in the last two periods (2005–2009 and 2006–2010).
Chapter 4: Results paper two 87
Table 4-5a. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2001 to 2005
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 26 1 (-1,2) ≤5th 22 0 (-1,1)
≥99th 4 2 (-3,6) ≤1st 1 -2 (-12,8)
≥3
≥95th 11 1 (-2,3) ≤5th 12 0 (-1,1)
≥99th 1 0 (-10,10) ≤1st - - -
≥4
≥95th 4 2 (-4,8) ≤5th 9 0 (-2,1)
≥99th - - - ≤1st - - -
88 Chapter 4: Results paper two
Table 4-5b. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2002 to 2006
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 22 1 (-1,2) ≤5th 23 0 (-1,1)
≥99th 2 3* (1,5)* ≤1st 1 -4 (-14,6)
≥3
≥95th 9 1 (-1,3) ≤5th 12 0 (-2,1)
≥99th - - - ≤1st - - -
≥4
≥95th 4 2 (-2,5) ≤5th 9 -1 (-2,1)
≥99th - - - ≤1st - - -
*P<0.05
Chapter 4: Results paper two 89
Table 4-5c. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2003 to 2007
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 22 1 (-1,2) ≤5th 25 0 (-1,1)
≥99th 2 3* (1,5)* ≤1st 1 -3 (-12,7)
≥3
≥95th 9 1 (-1,3) ≤5th 10 0 (-2,1)
≥99th - - - ≤1st - - -
≥4
≥95th 4 1 (-3,5) ≤5th 6 0 (-2,1)
≥99th - - - ≤1st - - -
*P<0.05
90 Chapter 4: Results paper two
Table 4-5d. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2004 to 2008
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 24 1 (-1,2) ≤5th 34 0 (-1,1)
≥99th 3 -1 (-6,4) ≤1st 1 -2 (-12,8)
≥3
≥95th 10 1 (-1,3) ≤5th 15 0 (-1,2)
≥99th 1 5* (1,10)* ≤1st - - -
≥4
≥95th 4 2 (-1,5) ≤5th 8 0 (-2,2)
≥99th - - - ≤1st - - -
*P<0.05
Chapter 4: Results paper two 91
Table 4-5e. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2005 to 2009
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 29 0 (-1,1) ≤5th 34 1 (0,2)
≥99th - - - ≤1st 1 0 (-12,11)
≥3
≥95th 14 0 (-1,1) ≤5th 14 2 (0,4)
≥99th - - - ≤1st - - -
≥4
≥95th 9 0 (-2,2) ≤5th 7 3* (1,5)*
≥99th - - - ≤1st - - -
*P<0.05
92 Chapter 4: Results paper two
Table 4-5f. Paediatric pneumonia due to the added effect of heat waves and cold spells in Brisbane, Australia, from 2006 to 2010
Heat Waves Cold Spells
No. of Consecutive
Days Percentile Days Pneumonia 95% CI Percentile Days Pneumonia 95% CI
≥2
≥95th 33 0 (-1,1) ≤5th 27 2 (0,3)
≥99th 2 -4 (-12,4) ≤1st 1 3 (-9,14)
≥3
≥95th 19 0 (-1,2) ≤5th 7 0 (-2,2)
≥99th 1 -2 (-10,4) ≤1st - - -
≥4
≥95th 12 0 (-2,2) ≤5th 2 15* (9,21)*
≥99th - - - ≤1st - - -
*P<0.05
Chapter 4: Results paper two 93
4.4 Discussion
This is the first study using satellite remote-sensing data to quantify the temperature-
pneumonia relationship and it has yielded several novel findings: i) Both low and
high temperatures were associated with an increase in childhood pneumonia; ii)
Children aged 2–5 years and female children were more vulnerable to temperature
effects on pneumonia, compared with children in other age groups and male children,
respectively. Indigenous children were more sensitive to the heat effect, compared
with non-Indigenous children; iii) Both heat waves and cold spells had added effects
on childhood pneumonia, and the magnitude of the added effects increased with
intensity and duration; iv) There was a decreasing trend in the high temperature
effect on childhood pneumonia, while the low temperature effect on childhood
pneumonia experienced an increasing trend. Meanwhile, the impact of heat waves
and cold spells on childhood pneumonia varied over time.
Previous studies looking at the impact of temperature on either mortality or
morbidity mainly rely on the data obtained from ground monitors (Basu and Samet
2002; Basu 2009; Ye et al. 2012), which may be not representative of the whole
population exposure (Kloog et al. 2013). Satellite remote sensing technology has
provided an unprecedented chance to increase the accuracy and precision of
environmental variable measurements (Goetz et al. 2000). As climate change
progresses, the global surface average temperature will increase, and cold-related
adverse impact on human well being may decrease accordingly (Xu et al. 2012). It is
pivotal to explore whether the decreasing cold-related impact can offset the
increasing heat-related impact, as climate change continues. Using the satellite
remote sensing data, we found the magnitude of the main effects of heat and cold
temperatures on childhood pneumonia was similar, suggesting that the increase in
94 Chapter 4: Results paper two
heat-related pneumonia would be compensated by a reduction in cold-related
pneumonia, and hence EDVs for childhood pneumonia in Brisbane attributable to the
main effect of temperature may not increase sharply in the near future. However, as
globe warms, the high risk season may differ, and the pattern may change.
Surface temperatures in urban areas are usually higher than rural regions, which may
have an exacerbating effect during heat waves (Johnson et al. 2009). Using the
temperature in the urban areas to examine the effect of heat waves on morbidity or
mortality in populations living in both urban and rural locations may also cause
measurement bias (Zeger et al. 2000). We used satellite remote sensing temperature
to avoid this problem and found that there were significant added effects of heat
waves and cold spells on childhood pneumonia, which increased with intensity and
duration. In the future, more frequent, intense, and longer-lasting persistent extreme
temperatures will occur as climate change continues (Meehl and Tebaldi 2004),
especially in Australia (IPCC 2013), and therefore the burden of childhood
pneumonia due to heat waves and cold spells might increase accordingly, which
requires the government to develop effective strategies incorporating other child
protective health measures to mitigate and adapt to adverse impact of heat waves and
cold spells (Xu et al. 2014).
The significant effect of temperature on childhood pneumonia we observed in this
study is not in accord with some previous studies. For example, Paynter et al. have
looked at the relationship between temperature and clinical pneumonia cases in
children <3 years in Bohol Province, Philippines, but did not find significant
association between temperature and childhood pneumonia (Paynter et al. 2013).
Temperature effect on childhood pneumonia can largely be due to its impact on the
aetiological pathogens. Existing science suggests that low temperature is associated
95
Chapter 4: Results paper two 95
with peaks of respiratory syncytial virus (RSV) (Yusuf et al., 2007), and
Streptococcus pneumoniae (Herrera-Lara et al. 2013; Watson et al. 2006), and high
temperature may increase the replication and survival of Mycoplasma pneumoniae
(Onozuka et al. 2009; Xu et al. 2011), Pneumocystis (Djawe et al. 2013) and
Legionella pneumophila (Herrera-Lara et al. 2013). The data we collected did not
include the information of lab-confirmed pneumonia pathogens, and thus we could
not separately analyse the relations between temperature and different aetiological
pathogens of childhood pneumonia.
Indigenous children have been found particularly vulnerable to high temperature in
this study, echoing to the findings that high temperatures substantially increased the
hospitalization risk for Indigenous Australians (Guo et al. 2013). Indigenous children
do not have adequate access to heat adaptation infrastructures, and they experience
more poverty than non-Indigenous children, which may render their great
vulnerability to heat (Ford 2012). The poor household infrastructure also adds their
risk of being exposed to extreme heat and cold (Bailie et al. 2010). Pneumonia in
children aged 2–5 years and female children were sensitive to both heat and cold,
which may be due to their anthropometry, body composition and social behaviour
(e.g., daily activity).
In this study, we also examined the changes in both main and added effects of
temperature on childhood pneumonia over time, and found that heat appeared to have
a decreasing impact on childhood pneumonia across a ten year study period, but cold
impact experienced an increasing trend, implying that children in Brisbane may have
gradually adapted to the heat effect while are still quite sensitive to cold effect. The
increasing use of air conditioning in Brisbane may contribute to the fact that heat
impact on childhood pneumonia declined in the past decade (Ostro et al. 2010), and
96 Chapter 4: Results paper two
the increasing cold effect on childhood pneumonia across the study period may be
due to the fact that Brisbane children rarely experience cold days, and they may not
take precautionary initiatives before cold days come. The change over time in main
effect of temperature on childhood pneumonia has an important implication for early
warning systems for extreme temperatures, because the existing heat alert systems or
early warning systems (Díaz et al. 2006; Nicholls et al. 2008) typically are based on
the average risks of temperature over multiple years but have not taken the temporal
variation of temperature impact into account (Xu et al. 2014). The impacts of heat
waves and cold spells on childhood pneumonia also experienced great changes over
time in this study. Effect of heat waves occurred in the first couple of periods (2002–
2006, 2003–2007 and 2004–2008), and effect of cold spells happened in the last two
periods (2005–2009 and 2006–2010), indicating that parents and caregivers of
children, especially those with the history of pneumonia, should take precautionary
measures, particularly during cold spells in the future.
This study has several strengths. As the first study using the satellite remote sensing
technology to measure children’s temperature exposure, it greatly minimizes
measurement error. We assessed both the main and added effects of temperature on
childhood pneumonia, and found the change over time in the main and added effects,
which gives important implications for future childhood pneumonia prevention. The
similar magnitude of cold and heat main effects on childhood pneumonia we
observed in this study indicates that temperature-related burden of childhood
pneumonia in Brisbane may not change dramatically due to increasing temperature in
the future. Several limitations of this study should also be acknowledged. First, the
spatial resolution of 6km*6km may be too crude to reflect the exposure at individual
level. However, we will be trying to get higher spatial resolution (e.g., 3km*3km or
97
Chapter 4: Results paper two 97
1km*1km) data and will use it in the future. Second, only one city is included in this
study, and thus it should be cautious to generalize our findings to other regions with
different climates. Third, people are mostly exposed to air temperature rather than
land surface temperature. However, we found that the correlation between air
temperature and surface temperature in Brisbane was very high (r>0.98), and thus we
think it is appropriate to use surface temperature as a temperature indicator to assess
its effects on childhood pneumonia.
4.5 Conclusions
Both high and low temperatures increased the risk of childhood pneumonia. As
climate change continues, persistent extreme temperatures increase, and children
with pneumonia history, especially those who are 2–5 years, female and Indigenous,
are at particular risk. Parents and caregivers should take precautionary measures to
protect children from being attacked by future frequent, intense and long-lasting
extreme temperatures. Policy makers should be aware of the temporal change in
temperature effect on children’s health while developing early warning systems.
4.6 References
Altizer S, Ostfeld RS, Johnson PTJ, Kutz S, Harvell CD. 2013. Climate change and
infectious diseases: From evidence to a predictive framework. Science
341(6145): 514-519.
Anderson B, Bell M. 2009. Weather-related mortality: how heat, cold, and heat
waves affect mortality in the United States. Epidemiology 20(2):205-213.
98 Chapter 4: Results paper two
Anderson TL, Charlson RJ, Bellouin N, Boucher O, Chin M, Christopher SA, et al.
2005. An “A-Train” strategy for quantifying direct climate forcing by
anthropogenic aerosols. Bull Amer Meteor Soc 86(12):1795-1809.
Bailie R, Stevens M, McDonald E, Brewster D, Guthridge S. 2010. Exploring cross-
sectional associations between common childhood illness, housing and social
conditions in remote Australian Aboriginal communities. BMC Public Health
10(1):147.
Basu R. 2009. High ambient temperature and mortality: a review of epidemiologic
studies from 2001 to 2008. Environ Health 8(1):40.
Basu R, Samet JM. 2002. Relation between elevated ambient temperature and
mortality: A review of the epidemiologic evidence. Epidemiol Rev
24(2):190-202.
Crosson WL, Al-Hamdan MZ, Hemmings SNJ, Wade GM. 2012. A daily merged
MODIS Aqua–Terra land surface temperature data set for the conterminous
United States. Remote Sens Environ 119:315-324.
Díaz J, Linares C, Tobías A. 2006. A critical comment on heat wave response plans.
Eur J Public Health 16(6):600.
Djawe K, Levin L, Swartzman A, Fong S, Roth B, Subramanian A, et al. 2013.
Environmental risk factors for pneumocystis pneumonia hospitalizations in
HIV patients. Clin Infect Dis 56(1):74-81.
Estes M, Al-Hamdan M, Crosson W, Estes S, Quattrochi D, Kent S, et al. 2009. Use
of remotely sensed data to evaluate the relationship between living
environment and blood pressure. Environ Health Perspect 117(12):1832-
1838.
99
Chapter 4: Results paper two 99
Evans J, van Donkelaar A, Martin RV, Burnett R, Rainham DG, Birkett NJ, et al.
2013. Estimates of global mortality attributable to particulate air pollution
using satellite imagery. Environ Res 120:33-42.
Ford JD. 2012. Indigenous health and climate change. Am J Public Health
102(7):1260-1266.
Goetz S, Prince S, Small J. 2000. Advances in satellite remote sensing of
environmental variables for epidemiological applications. Adv Parasitol
47:289-307.
Guo Y, Wang Z, Li S, Tong S, Barnett A. 2013. Temperature sensitivity in
indigenous Australians. Epidemiology 24(3):471-472.
Hajat S, Armstrong B, Baccini M, Biggeri A, Bisanti L, Russo A, et al. 2006. Impact
of high temperatures on mortality: Is there an added heat wave effect?
Epidemiology 17(6): 632-638.
Herrera-Lara S, Fernández-Fabrellas E, Cervera-Juan Á, Blanquer-Olivas R. 2013.
Do seasonal changes and climate influence the etiology of community
acquired pneumonia? Arch Bronconeumol 49(4):140-145.
IPCC. 2013. Summary for policymakers. In: Climate change 2013: the physical
science basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge.
Johnson D, Wilson J, Luber G. 2009. Socioeconomic indicators of heat-related health
risk supplemented with remotely sensed data. Int J Health Geogr 8(1):57.
Kaufman YJ, Herring DD, Ranson KJ, Collatz GJ. 1998. Earth Observing System
AM1 mission to Earth. Geoscience and Remote Sensing, IEEE Transactions
on 36(4): 1045-1055.
100 Chapter 4: Results paper two
Kloog I, Ridgway B, Koutrakis P, Coull B, Schwartz J. 2013. Long- and short-term
exposure to PM2.5 and mortality: using novel exposure models.
Epidemiology 24(4):555-561.
Laaidi K, Zeghnoun A, Dousset B, Bretin P, Vandentorren S, Giraudet E, et al. 2012.
The impact of heat islands on mortality in Paris during the August 2003 heat
wave. Environ Health Perspect 120(2):254-259.
McKie R. 2013. Children will suffer most as climate change increases in coming
decades, say scientists. BMJ 347:f5799.
McMichael AJ. 2013. Globalization, climate change, and human health. N Engl J
Med 368(14):1335-1343.
Meehl GA, Tebaldi C. 2004. More intense, more frequent, and longer lasting heat
waves in the 21st century. Science, 305(5686):994-997.
Nicholls N, Skinner C, Loughnan M, Tapper N. 2008. A simple heat alert system for
Melbourne, Australia. Int J Biometeorol 52(5):375-384.
Onozuka D, Hashizume M, Hagihara A. 2009. Impact of weather factors on
Mycoplasma pneumoniae pneumonia. Thorax 64(6):507-511.
Ostro B, Rauch S, Green R, Malig B, Basu R. 2010. The effects of temperature and
use of air conditioning on hospitalizations. Am J Epidemiol 172(9):1053-
1061.
Paynter S, Weinstein P, Ware RS, Lucero MG, Tallo V, Nohynek H, et al. 2013.
Sunshine, rainfall, humidity and child pneumonia in the tropics: time-series
analyses. Epidemiol Infect 141(06):1328-1336.
Sheffield P, Landrigan P. 2011. Global climate change and children's health: threats
and strategies for prevention. Environ Health Perspect 119(3):291-298.
101
Chapter 4: Results paper two 101
Walker CLF, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. 2013. Global
burden of childhood pneumonia and diarrhoea. Lancet 381(9875):1405-1416.
Wan Z. 2008. New refinements and validation of the MODIS Land-Surface
Temperature/Emissivity products. Remote Sens Environ 112(1):59-74.
Wan Z, Zhang Y, Zhang Q, Li Z. 2002. Validation of the land-surface temperature
products retrieved from Terra Moderate Resolution Imaging
Spectroradiometer data. Remote Sens Environ 83(1–2):163-180.
Wang Z, Liu Y, Hu M, Pan X, Shi J, Chen F, et al. 2013. Acute health impacts of
airborne particles estimated from satellite remote sensing. Environ Int
51:150-159.
Watson M, Gilmour R, Menzies R, Ferson M, McIntyre P, Network NS. 2006. The
association of respiratory viruses, temperature, and other climatic parameters
with the incidence of invasive pneumococcal disease in Sydney, Australia.
Clin Infect Dis 42(2): 211-215.
Xu YC, Zhu LJ, Xu D, Tao XF, Li SX, Tang LF, et al. (2011). Epidemiological
characteristics and meteorological factors of childhood Mycoplasma
pneumoniae pneumonia in Hangzhou. World J Pediatr 7(3):240-244.
Xu Z, Huang C, Hu W, Turner LR, Su H, Tong S. 2013a. Extreme temperatures and
emergency department admissions for childhood asthma in Brisbane,
Australia. Occup Environ Med 70(10):730-735.
Xu Z, Huang C, Su H, Turner L, Qiao Z, Tong S. 2013b. Diurnal temperature range
and childhood asthma: a time-series study. Environ Health 12(1):12.
Xu Z, Sheffield Pe, Su H, Wang X, Bi Y, Tong, S. 2014. The impact of heat waves
on children’s health: a systematic review. Int J Biometeorol 58(2):239-47
102 Chapter 4: Results paper two
Xu Z, Sheffield PE, Hu W, Su H, Yu W, Qi X, et al. 2012. Climate change and
children’s health—A call for research on what works to protect children. Int J
Environ Res Public Health 9(9):3298-3316.
Ye X, Wolff R, Yu W, Vaneckova P, Pan X, Tong S. 2012. Ambient temperature
and morbidity: a review of epidemiological evidence. Environ Health
Perspect 120(1):19-28.
Yusuf S, Piedimonte G, Auais A, Demmler G, Krishnan S, Van Caeseele P, et al.
2007. The relationship of meteorological conditions to the epidemic activity
of respiratory syncytial virus. Epidemiol Infect 135(07):1077-1090.
Zeger S, Thomas D, Dominici F, Samet J, Schwartz J, Dockery D, et al. 2000.
Exposure measurement error in time-series studies of air pollution: concepts
and consequences. Environ Health Perspect 108(5):419-426.
Zhang K, Oswald EM, Brown DG, Brines SJ, Gronlund CJ, White-Newsome JL, et
al. 2011. Geostatistical exploration of spatial variation of summertime
temperatures in the Detroit metropolitan region. Environ Res 111(8):1046-
1053.
Wan Z, Dozier J. 1996. A generalized split-window algorithm for retrieving land-
surface temperature from space. Geoscience and Remote Sensing, IEEE
Transactions on 34(4), 892-905.
Chapter 5: Results paper three 103
Chapter 5: Results paper three
Assessment of the temperature effect on childhood
diarrhoea using satellite imagery
Xu Z, Liu Y, Ma Z, Toloo GS, Hu W, Tong S (2014). Assessment of the temperature effect
on childhood diarrhoea using satellite imagery. Scientific Reports, 4:5389.
104 Chapter 5: Results paper three
Abstract
A quasi-Poisson generalized linear model combined with a distributed lag non-linear model
was used to quantify the main effect of temperature on emergency department visits (EDVs)
for childhood diarrhoea in Brisbane from 2001 to 2010. Residual of the model was checked
to examine whether there was an added effect due to heat waves. The change over time in
temperature-diarrhoea relation was also assessed. Both low and high temperatures had
significant impact on childhood diarrhoea. Heat waves had an added effect on childhood
diarrhoea, and this effect increased with intensity and duration of heat waves. There was a
decreasing trend in the main effect of heat on childhood diarrhoea in Brisbane across the
study period. Brisbane children appeared to have gradually adapted to mild heat, but they are
still very sensitive to persistent extreme heat. Development of future heat alert systems
should take the change in temperature-diarrhoea relation over time into account.
105
Chapter 5: Results paper three 105
5.1 Introduction
Climate change has impacted and will increasingly influence human health, especially in the
context of rapid globalization (McMichael 2013). Children are particularly vulnerable to
climate change impact (Sheffield and Landrigan, 2011). They may experience greater risk of
infectious diseases (e.g., diarrhoea) as global surface average temperature increases
(Checkley et al. 2000).
Prior studies have well documented that heat waves may increase morbidity and mortality
(Gasparrini and Armstrong 2011; Ma et al. 2013). Some researchers have claimed that the
impact of heat waves on human health may be due to both the main effect of daily high
temperature and the added effect of persistent periods of heat (Anderson and Bell 2009;
Gasparrini and Armstrong 2011; Hajat et al. 2006). As climate change continues, there will
be more frequent, more intense and longer-lasting heat waves (Meehl and Tebaldi 2004).
Food chain, from food preparation stage to production process, may be affected by persistent
high temperatures, possibly resulting in more food-borne diseases (D'Souza et al. 2004).
Some studies have reported that food poisoning (Graham et al. 1995; Bentham and Langford
2001) and electrolyte imbalance (Knowlton et al. 2009) are more likely to occur during
periods of persistent hot temperatures. However, studies on the effect of heat waves on
childhood diarrhoea are scarce.
Prior studies looking at the impact of temperature on diarrhoeal diseases mainly used time-
series approach and obtained temperature data averaged from a network of sites (Checkley et
al. 2000; Hashizume et al. 2007), and the several monitoring sites are normally in or nearby
the urban areas (Xu et al. 2013b). This may render measurement bias because temperature
usually varies spatially across one city (Zhang et al. 2011) due to urban heat island (Laaidi et
al. 2012). Satellite-based monitoring data can largely solve this problem, given its broad
spatial coverage. Estes et al. have applied the remote sensing technology to examining the
106 Chapter 5: Results paper three
effect of temperature on blood pressure (Estes et al. 2009). However, to date, no study has
used satellite remote sensing data to examine the relationship between temperature and
childhood diarrhoea.
This study used the data on satellite remote sensing temperature and attempted to address
three research issues: i) What is the relationship between temperature and emergency
department visits (EDVs) for childhood diarrhoea in Brisbane, Australia? ii) Is there any
added effect attributable to heat waves? iii) Is there any change over time in the effect of
temperature on childhood diarrhoea during the study period?
5.2 Methods
Data collection
Public hospital emergency departments are a significant and high-profile component of
Australia’s health care system (FitzGerald et al. 2012). EDVs data, which were classified
according to International Classification of Diseases, 9th and 10th versions (ICD-9 and ICD-
10), were supplied by Queensland Health. The details of the Brisbane data (selected hospitals
and covered regions, etc.) have been clarified in Chapter 4. We selected the following codes
for diarrhoea in children aged 0–14 years: ICD-9 codes: 001–003, 004, 005, 006.0–006.2,
007.0–007.5, 008–009; ICD–10 codes: A00–A03, A04, A05, A06.0–A06.3, A06.9, A07.0–
A07.2, A07.9, A08–A09. Existing evidence suggests that there is significant difference in the
seasonal variations of infection caused by various pathogens (Chui et al. 2011; Naumova et
al., 2007). Viral (008–009, and A08–A09), bacterial (001–003, 004, 005, and A00–A03, A04,
A05) and parasitic infections (006.0–006.2, 007.0–007.5 and A06.0–A06.3, A06.9, A07.0–
A07.2, A07.9) were separately analysed. Ethical approval was obtained from the Human
Research Ethics Committee of Queensland University of Technology (Australia) prior to the
data being collected. Patient information was de-identified and thus no written informed
107
Chapter 5: Results paper three 107
consent was obtained. Data on rainfall and relative humidity were obtained from the
Australian Bureau of Meteorology. The data were collected from eight monitor stations
throughout Brisbane, and then averaged. Details on the collection of land surface temperature
have been discussed in the Methods section of Chapter four of this thesis.
Data analysis
To assess the main effect of temperature on EDVs for childhood diarrhoea, we used a quasi-
Poisson generalized linear model combined with a distributed lag non-linear model (DLNM)
(Xu et al. 2013). We used a “natural cubic spline–natural cubic spline” DLNM to examine
the temperature effect using four degrees of freedom (df) for both temperature and lag
dimensions, and selected a lag of 10 days to capture the possible lagged effect (Xu et al.
2013). Rainfall, relative humidity, long-term trend, seasonality, day of week and public
holiday were controlled for using the same approaches in Chapter four.
After all the other parameters were confirmed, we checked the temperature–diarrhoea plot
and chose the reference temperature by visual inspection. We calculated the relative risk of
EDVs for childhood diarrhoea associated with high temperature (29.6°C, 99th percentile of
mean temperature) and low temperature (10.4°C, 1st percentile of mean temperature) relative
to the reference temperature (chosen to be 16.0°C). To detect the change over time in the
association between temperature and diarrhoea, we specifically quantified the effect of
temperature on EDVs for childhood diarrhoea for a sliding window of five years (2001–2005,
2002–2006, 2003–2007, 2004–2008, 2005–2009 and 2006–2010). Further, the same
approach in Chapter four was used to detect whether there was any added effect of heat
waves on childhood diarrhoea. All data analysis was conducted using R environment
(Version 2.15). The sensitivity analysis was conducted by adjusting df for temperature and
time.
108 Chapter 5: Results paper three
5.3 Results
Summary statistics
There were a total of 58166 EDVs for childhood diarrhoea during the study period. Table 5-1
presents the summary statistics of daily weather variables and EDVs for childhood diarrhoea
in the total children population and each subgroup. The mean value of satellite remote
sensing temperature was 19.8°C. The average values of relative humidity and rainfall were
65.0% and 2.2 mm, respectively. The mean value of daily EDVs for childhood diarrhoea was
15.9 (range=10–91), with the predominant pathogen being virus (mean=15.6). There were
very few EDVs for bacterial (mean=0.3) and parasitic (mean=0.04) diarrhoea every day.
Figure 5-1 shows the daily distributions of decomposed EDVs for childhood diarrhoea and
weather variables, illustrating a strong seasonal trend for diarrhoea and satellite remote
sensing temperature. The daily distributions of EDVs for viral, bacterial and parasitic
diarrhoea were presented in Figure 5-2.
109
Chapter 5: Results paper three 109
Table 5-1. Summary statistics for climatic variables and paediatric diarrhoea in Brisbane, Australia, 2001–2010
Variables Mean SD Min Percentile
Max 25 50 75
RS mean temperature (°C)* 19.8 4.6 6.5 16.2 19.9 23.3 32.6
Relative humidity (%) 65.0 15.0 13.0 56.0 65.0 75.0 100.0
Rainfall (mm) 2.2 8.3 0 0 0 0.4 149.0
Diarrhoea 15.9 8.8 0 10 14 20 91
Viral diarrhoea 15.6 8.8 0 10 14 20 91
Bacterial diarrhoea 0.3 0.7 0 0 0 0 7
Parasitic diarrhoea 0.04 0.2 0 0 0 0 3
(0-1) 4.0 2.7 0 2 4 6 18
[1-2) 4.0 3.3 0 2 3 5 25
[2,5) 5.0 3.5 0 2 4 6 37
[5,14] 2.9 1.9 0 1 2 4 15
Male 7.3 4.6 0 4 7 9 47
Female 8.6 5.2 0 5 8 11 44
Indigenous 0.7 1.0 0 0 1 2 15
Non-indigenous 15.2 8.8 0 9 14 20 91
* RS: remote sensing
110 Chapter 5: Results paper three
Figure 5-1. The daily distributions of EDVs for paediatric diarrhoea and climatic factors in Brisbane, from 2001 to 2010
The left side is the temporal distribution of diarrhoea, and the right side is the temporal distributions of remote sensing temperature, rainfall and
relative humidity.
111
Chapter 5: Results paper three 111
Figure 5-2. The daily distribution of diarrhoea caused by different pathogens
From the top to the bottom: total, virus, bacteria and parasite.
112 Chapter 5: Results paper three
The Spearman correlations between climate variables and EDVs for childhood diarrhoea are
presented in Table 5-2. EDVs for childhood diarrhoea were positively correlated with
satellite remote sensing temperature (r=0.04, P<0.01), and negatively correlated with relative
humidity (r=-0.11, P<0.01).
Table 5-2. Spearman’s correlation between daily weather conditions, air pollutants and paediatric diarrhoea in Brisbane, Australia, from 2001–2010
RS mean temperature†
Relative humidity Rainfall Diarrhoea Viral
diarrhoea Bacterial diarrhoea
Parasitic diarrhoea
RS mean temperature† 1.00
Relative humidity -0.25* 1.00
Rainfall -0.03 0.38* 1.00
Diarrhoea 0.04* -0.11* 0.01 1.00
Viral diarrhoea 0.02 -0.13* -0.01 0.99* 1.00
Bacterial diarrhoea 0.02 -0.04* -0.01 -0.03 -0.11* 1.00
Parasitic diarrhoea 0.01 -0.01 -0.02 0.05* 0.03 -0.03 1.00
* P<0.01; †RS, remote sensing
113
Chapter 5: Results paper three 113
The effect of temperature on EDVs for childhood diarrhoea
Figure 5-3 reveals that both low and high temperatures were associated with increase in
EDVs for childhood diarrhoea. Table 5-3 quantitatively depicts the effects of temperature on
EDVs for childhood diarrhoea by pathogen, age, gender and Indigenous status. Due to the
very limited number of parasitic diarrhoea, only the results for viral and bacterial diarrhoea
were presented. No significant relationship between temperature and bacterial diarrhoea was
found. The relative risk (RR) of diarrhoea during hot days in children aged 1–2 years (not
including 2 years) was (RR: 1.17; 95% Confidence interval (CI): 1.10–1.25) greater than
children of other age groups, and the RR of diarrhoea during cold days in children aged 2–5
years (RR: 1.10; 95% CI: 1.03–1.18) was greater than other age groups. The effects of
extreme temperatures on male children and Indigenous children appeared to be higher than
female children and non-Indigenous children, respectively. Heat effect on EDVs for
childhood diarrhoea was acute, mainly occurring on the current day of exposure, and cold
effect happened after several days of exposure.
114 Chapter 5: Results paper three
Figure 5-3. The overall effect of mean temperature on paediatric diarrhoea in Brisbane, from 2001 to 2010
115
Chapter 5: Results paper three 115
Table 5-3. The cumulative effect of high and low temperatures on EDVs for paediatric diarrhoea in Brisbane, with 99th percentile (29.6 °C) and 1st (10.4°C) of temperature relative to reference temperature (16 °C)
Diseases Heat effect (Relative risk (95% CI)) Cold effect (Relative risk (95% CI))
Lag 0–1 Lag 0–7 Lag 0–10 Lag 0–1 Lag 0–7 Lag 0–10
Total 1.08(1.04,1.13)* 0.99(0.96,1.02) 1.01(0.96,1.06) 1.04(0.99,1.09) 1.02(0.98.1.05) 1.05(1.01,1.10)*
Viral diarrhoea 1.10(1.06,1.13)* 1.01(0.97,1.05) 0.99(0.93,1.05) 1.02(0.99,1.06) 1.01(0.98,1.05) 1.06(1.01,1.12)*
Bacterial diarrhoea 1.01(0.83,1.23) 0.85(0.68,1.07) 1.34(0.95,1.91) 1.12(0.92,1.37) 1.05(0.84,1.28) 0.87(0.61,1.24)
(0-1) 1.06(1.01,1.13)* 0.99(0.95,1.03) 0.98(00.90,1.06) 1.06(0.99,1.13) 1.01(0.97,1.05) 0.99(0.92,1.07)
[1-2) 1.17(1.10,1.25)* 0.98(0.94,1.03) 1.07(0.98,1.17) 1.03(0.97,1.09) 1.01(0.97.1.05) 1.06(0.98,1.04)
[2,5) 1.07(1.01,1.13)* 1.00(0.96,1.04) 1.01(0.93,1.10) 1.03(0.97,1.09) 1.02(0.98,1.06) 1.10(1.03,1.18)*
[5,14] 1.03(0.96,1.11) 0.99(0.94,1.04) 0.97(0.88,1.06) 1.07(0.98,1.17) 1.02(0.97,1.07) 1.09(1.01,1.18)*
Male 1.10(1.05,1.15)* 0.98(0.95,1.02) 1.01(0.95,1.08) 1.04(0.99,1.09) 1.02(0.98,1.05) 1.06(1.01,1.12)*
Female 1.07(1.02,1.12)* 0.99(0.96,1.03) 1.00(0.94,1.07) 1.03(0.99,1.08) 1.02(0.99,1.05) 1.05(0.99,1.11)
Indigenous 1.10(1.01,1.28)* 1.05(0.95,1.16) 1.12(0.92,1.36) 1.15(0.98,1.33) 1.01(0.92,1.11) 1.18(1.01,1.39)*
Non-indigenous 1.08(1.04,1.12)* 0.99(0.96,1.01) 1.00(0.95,1.06) 1.03(0.98,1.09) 1.02(0.99,1.04) 1.03(1.01,1.05)* *P-value<0.05
116 Chapter 5: Results paper three
The added effect of heat waves
Table 5-4 shows the daily excess EDVs for childhood pneumonia on heat wave days
compared with non-heat wave days. We found no apparent added effect of heat waves on
EDVs for childhood diarrhoea while using the heat wave definitions of two or more
consecutive days with the temperature over the 95th percentile. However, we found
significant added effects of heat waves on EDVs for childhood diarrhoea at the temperature
threshold over the 99th percentile. Further, with heat wave days increasing from two to three
consecutive days, the number of EDVs for childhood diarrhoea due to added effect of heat
waves rose from three to seven.
117
Chapter 5: Results paper three 117
Table 5-4. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2001 to 2010
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 58 0 (-1,1)
≥99th 6 3* (1,5)
≥3
≥95th 31 0 (-1,1)
≥99th 2 7* (2,13)
≥4
≥95th 15 0 (-1,2)
≥99th - - -
*P<0.05
118 Chapter 5: Results paper three
Change over time in the effect of temperature on childhood diarrhoea
Figure 5-4 illustrates the change in the temperature effect of EDVs for childhood diarrhoea
over time. The heat effects increased slightly from Period 1 (2001-2005) to Period 2 (2002-
2006) and showed a decreasing trend thenceforward. No significant change over time in the
cold effect on EDVs for childhood diarrhoea was found.
Table 5-5 shows the added effects of heat waves on EDVs for childhood diarrhoea in the six
periods (2001–2005, 2002–2006, 2003–2007, 2004–2008, 2005–2009 and 2006–2010).
Statistically significant added effect of heat waves on EDVs for childhood diarrhoea was
observed only in the last period (2006–2010).
Figure 5-4. The change over time in the temperature effect on childhood diarrhoea
Left hand side: hot effect; right hand side: cold effect; p1= 2001-2005, p2=2002-2006,
p3=2003-2007, p4=2004-2008, p5=2005-2009, p6=2006-2010.
119
Chapter 5: Results paper three 119
Table 5-5a. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2001 to 2005
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 26 1 (0,1)
≥99th 4 1 (-2,5)
≥3
≥95th 11 1 (-1,2)
≥99th 1 0 (-7,7)
≥4
≥95th 4 0 (-1,2)
≥99th - - -
120 Chapter 5: Results paper three
Table 5-5b. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2002 to 2006
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 22 0 (-1,1)
≥99th 2 1 (-2,4)
≥3
≥95th 9 0 (-1,2)
≥99th - - -
≥4
≥95th 4 0 (-3,2)
≥99th - - -
121
Chapter 5: Results paper three 121
Table 5-5c. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2003 to 2007
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 22 0 (-1,1)
≥99th 2 0 (-3,3)
≥3
≥95th 9 0 (-1,1)
≥99th - - -
≥4
≥95th 4 -1 (-2,1)
≥99th - - -
122 Chapter 5: Results paper three
Table 5-5d. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2004 to 2008
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 24 0 (-1,1)
≥99th 3 0 (-4,4)
≥3
≥95th 10 -1 (-2,1)
≥99th 1 2 (-9,13)
≥4
≥95th 4 -1 (-3,1)
≥99th - - -
123
Chapter 5: Results paper three 123
Table 5-5e. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2005 to 2009
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 29 0 (-1,1)
≥99th - - -
≥3
≥95th 14 5 (-8,18)
≥99th - - -
≥4
≥95th 9 -2 (-4,1)
≥99th - - -
124 Chapter 5: Results paper three
Table 5-5f. Paediatric diarrhoea due to the added effect of heat waves in Brisbane, Australia, from 2006 to 2010
Heat Waves
No. of Consecutive Days Percentile Days Diarrhoea 95% CI
≥2
≥95th 33 0 (-1,1)
≥99th 2 7* (1,14)*
≥3
≥95th 19 0 (-1,1)
≥99th 1 13* (1,25)*
≥4
≥95th 12 0 (-2,2)
≥99th - - -
*P<0.05
Chapter 5: Results paper three 125
5.4 Discussion
The spatial variability of temperature across a city has been well documented in the
literature (Kestens et al. 2011). Existing studies quantifying the impact of
temperature on childhood diarrhoea many used data collected from ground monitors
in a city (Checkley et al. 2000). Due to the limited number of ground monitoring
sites, the temperature collected may not well represent the exposure of whole
population, possibly resulting in measurement bias in the effect estimates. In this
study, we used satellite remote sensing data to minimise this problem. This study
examined the effects of both high and low temperatures as well as heat waves on
childhood diarrhoea, while our previous work only examined the effects of
temperature variation on childhood diarrhoea (Xu et al. 2013). Both heat and cold
were associated with increase in EDVs for childhood diarrhoea in Brisbane. An
added effect of heat waves on childhood diarrhoea was found, though this effect
varied greatly across the study period. The effect of high temperature on childhood
diarrhoea showed a decreasing trend over time.
Both heat and cold have been found to be associated with increases in EDVs for
childhood diarrhoea in Brisbane, which may be partially explained by three reasons.
First, high temperature promotes the growth of bacteria (Hashizume et al. 2007),
while low temperature increases the replication and survival of virus, e.g., rotavirus
(D'Souza et al. 2008). Second, high temperature may impact the food chain, from
food preparation stage to production process (D'Souza et al. 2004), and expose
children more to contaminated food. Third, extremely low and high temperatures
may alter children’s hygiene behaviours (e.g., water drinking behaviour).
As climate change continues, global surface average temperature will increase, and
heat-related diarrhoea burden may increase accordingly, but cold related diarrhoea
126 Chapter 5: Results paper three
burden may decrease, especially for viral diarrhoea which favours cold temperature
(Xu et al. 2012). Hence, it is essential to explore the balance between cold and heat
effects on childhood diarrhoea. In this study, we found cold effect on childhood
diarrhoea was greater than heat effect, which might be explained by the fact that in
industrialized countries, interventions to improve hygiene and sanitation may
decrease the occurrence of diarrhoea caused by bacteria and parasites, but for
rotavirus-related diarrhoea which is spread from person-to person, these
interventions may be less effective, and virus may be the dominant aetiological
pathogen in these regions (Olesen et al. 2005; Malek et al. 2006; Parashar et al.
2009). This finding implies that EDVs for childhood diarrhoea in Brisbane related to
the main effect of temperature may not increase greatly as climate change progresses.
In this study, we found Indigenous children were particularly vulnerable to the
impact of temperature on diarrhoea, which corresponds to the findings from a cohort
study reporting that Indigenous Australians were very sensitive to high and low
temperatures (Guo et al. 2013). Indigenous children require more public health
attention in Australia. They have restricted access to medical service and climate
change adaptation infrastructures, and high or cold temperature may trigger or
exacerbate their existing health problems. The poor housing conditions may also
render their greater vulnerability to heat or cold impact (Bailie et al. 2010). The RR
of diarrhoea in male children during extreme temperatures was greater than female
children, which might be partially due to their body composition (Maeda et al. 2005)
and behaviours (White-Newsome et al. 2011). Basu et al. argued that differences in
the effect of temperature on males and females varied among different locations and
populations (Basu and Samet 2002), and we believe that the differences of
temperature sensitivity between boys and girls may even vary with disease types.
127
Chapter 5: Results paper three 127
An added effect of heat waves on childhood diarrhoea has been observed in this
study, and this effect increased with the intensity and duration of heat waves,
suggesting that the burden of childhood diarrhoea associated with heat waves may
increase as more frequent, intense and longer-lasting heat waves are projected to
occur in the future (Meehl and Tebaldi 2004; IPCC 2013). Parents and caregivers
should be educated and made aware of this risk and take precautionary measures to
protect their children during heat waves, and the government may also consider the
development of an heat early warning system as it will substantially decrease
children’s disease burden in heat waves (Xu et al. 2013). Interestingly, we found that
the effect of high temperature on childhood diarrhoea had a declining trend across
the study period, but the added effect of heat waves appeared to increase in recent
years. The decreasing trend in the main effect of heat on childhood diarrhoea may be
partially explained by the decreasing mean temperatures in the last four years (Figure
1). The finding also imply that children in Brisbane might be experiencing better
hygiene standards and/or have increasingly adapted to mild heat in recent years, but
persistent extremely hot days still pose a huge challenge to the health of their
intestinal system.
There are several strengths of this study. This is the first study to apply the satellite
remote sensing technology to quantifying the temperature-diarrhoea association,
which minimized the measurement bias. Our study examined the balance between
heat and cold effects on childhood diarrhoea and firstly reported the added effect of
heat waves on childhood diarrhoea. Two main weaknesses should also be
acknowledged. First, the ecological design restricts us to explore the possible
confounders (people’s drinking behaviour, etc.) and may cause ecological fallacy.
128 Chapter 5: Results paper three
Second, we did not have the pathogen data and thus could not specifically analyse
the relation between temperature and diarrhoea caused by different pathogens.
In conclusion, both hot and cold temperatures were associated with childhood
diarrhoea, and male children and Indigenous children appeared to be at higher risk.
Heat waves had an added effect on childhood diarrhoea, which increased with
intensity and duration of heat waves. Parents, caregivers, schools and the government
should take action to enhance the children’s intestinal health particularly during
extreme temperatures and promote protective measures in advance.
5.5 References
Anderson B, Bell M. 2009. Weather-related mortality: how heat, cold, and heat
waves affect mortality in the United States. Epidemiology 20(2):205-213.
Bailie R, Stevens M, McDonald E, Brewster D, Guthridge S. 2010. Exploring cross-
sectional associations between common childhood illness, housing and social
conditions in remote Australian Aboriginal communities. BMC Public Health
10(1):147.
Basu R, Samet JM. 2002. Relation between elevated ambient temperature and
mortality: A review of the epidemiologic evidence. Epidemiol Rev
24(2):190-202.
Bentham G, Langford I. 1995. Climate change and the incidence of food poisoning
in England and Wales. Int J Biometeorol 39(2):81-86.
Bentham G, Langford I. 2001. Environmental temperatures and the incidence of food
poisoning in England and Wales. Int J Biometeorol 45(1):22-26.
129
Chapter 5: Results paper three 129
Checkley W, Epstein L, Gilman R, Figueroa D, Cama R, Patz J, et al. 2000. Effect of
El Niño and ambient temperature on hospital admissions for diarrhoeal
diseases in Peruvian children. Lancet 355(9202):442-450.
Chui KKH, Jagai JS, Griffiths JK, Naumova EN. 2011. Hospitalization of the elderly
in the United States for nonspecific gastrointestinal diseases: A search for
etiological clues. Am J Public Health 101(11):2082-2086.
D'Souza RM, Becker N, Hall G, Moodie K. 2004. Does ambient temperature affect
foodborne disease? Epidemiology 15(1):86-92.
D'Souza RM, Hall G, Becker NG. 2008. Climatic factors associated with
hospitalizations for rotavirus diarrhoea in children under 5 years of age.
Epidemiol Infect 136(01):56-64.
Estes M, Al-Hamdan M, Crosson W, Estes S, Quattrochi D, Kent S, et al. 2009. Use
of remotely sensed data to evaluate the relationship between living
environment and blood pressure. Environ Health Perspect 117(12):1832-
1838.
FitzGerald G, Toloo S, Rego J, Ting J, Aitken P, Tippett V. 2012. Demand for public
hospital emergency department services in Australia: 2000–2001 to 2009–
2010. Emerg Med Australas 24(1):72-78.
Gasparrini A, Armstrong B. 2011. The impact of heat waves on mortality.
Epidemiology 22(1):68-73.
Guo Y, Wang Z, Li S, Tong S, Barnett A. 2013. Temperature sensitivity in
indigenous Australians. Epidemiology 24(3):471-472.
Hajat S, Armstrong B, Baccini M, Biggeri A, Bisanti L, Russo A, et al. 2006. Impact
of high temperatures on mortality: Is there an added heat wave effect?
Epidemiology 17(6): 632-638.
130 Chapter 5: Results paper three
Hashizume M, Armstrong B, Hajat S, Wagatsuma Y, Faruque AS, Hayashi T, et al.
2007. Association between climate variability and hospital visits for non-
cholera diarrhoea in Bangladesh: effects and vulnerable groups. Int J
Epidemiol 36(5):1030-1037.
IPCC. 2013. Summary for policymakers. In: Climate change 2013: the physical
science basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge.
Kestens Y, Brand A, Fournier M, Goudreau S, Kosatsky T, Maloley M, et al. 2011.
Modelling the variation of land surface temperature as determinant of risk of
heat-related health events. Int J Health Geogr 10(1):7.
Knowlton K, Rotkin-Ellman M, King G, Margolis H, Smith D, Solomon G, et al.
2009. The 2006 California heat wave: impacts on hospitalizations and
emergency department visits. Environ Health Perspect 117(1):61-67.
Laaidi K, Zeghnoun A, Dousset B, Bretin P, Vandentorren S, Giraudet E, et al. 2012.
The impact of heat islands on mortality in Paris during the August 2003 heat
wave. Environ Health Perspect 120(2):254-259.
Ma W, Yang C, Chu C, Li T, Tan J, Kan H. 2013. The impact of the 2008 cold spell
on mortality in Shanghai, China. Int J Biometeorol 57(1):179-184.
Maeda T, Sugawara A, Fukushima T, Higuchi S, Ishibashi K. 2005. Effects of
lifestyle, body composition, and physical fitness on cold tolerance in humans.
J Physiol Anthropol Appl Human Sci 24(4):439-443.
Malek MA, Curns AT, Holman RC, Fischer TK, Bresee JS, Glass RI, et al. 2006.
Diarrhea- and rotavirus-associated hospitalizations among children less than
5 years of age: United States, 1997 and 2000. Pediatrics 117(6):1887-1892.
131
Chapter 5: Results paper three 131
McMichael AJ. 2013. Globalization, climate change, and human health. N Engl J
Med 368(14): 1335-1343.
Meehl GA, Tebaldi C. 2004. More intense, more frequent, and longer lasting heat
waves in the 21st century. Science 305(5686):994-997.
Naumova EN, Jagai JS, Matyas B, Demaria A, Macneill IB, Griffiths JK. 2007.
Seasonality in six enterically transmitted diseases and ambient temperature.
Epidemiol Infect 135(02):281-292.
Olesen B, Neimann J, Böttiger B, Ethelberg S, Schiellerup P, Jensen C, et al. 2005.
Etiology of diarrhea in young children in Denmark: a case-control study. J
Clin Microbiol 43(8):3636-3641.
Parashar UD, Burton A, Lanata C, Boschi-Pinto C, Shibuya K, Steele D, et al. 2009.
Global mortality associated with rotavirus disease among children in 2004. J
Infect Dis 200(S1):S9-S15.
Sheffield P, Landrigan P. 2011. Global climate change and children's health: threats
and strategies for prevention. Environ Health Perspect 119(3):291-298.
Tong S, Wang XY, Barnett AG. 2010. Assessment of heat-related health impacts in
Brisbane, Australia: Comparison of different heatwave definitions. PLoS
ONE 5(8):e12155.
White-Newsome JL, Sánchez BN, Parker EA, Dvonch JT, Zhang Z, O’Neill MS.
2011. Assessing heat-adaptive behaviors among older, urban-dwelling adults.
Maturitas 70(1):85-91.
Xu Z, Huang C, Turner LR, Su H, Qiao Z, Tong S. 2013b. Is diurnal temperature
range a risk factor for childhood diarrhea? PLoS ONE 8(5):e64713.
Xu Z, Sheffield P, Su H, Wang X, Bi Y, Tong S. 2014. The impact of heat waves on
children’s health: a systematic review. Int J Biometeorol 58(2):239-247.
132 Chapter 5: Results paper three
Xu Z, Sheffield PE, Hu W, Su H, Yu W, Qi X, et al. 2012. Climate change and
children’s health—A call for research on what works to protect children. Int J
Environ Res Public Health 10;9(9):3298-316.
Zhang K, Oswald EM, Brown DG, Brines SJ, Gronlund CJ, White-Newsome JL, et
al. 2011. Geostatistical exploration of spatial variation of summertime
temperatures in the Detroit metropolitan region. Environ Res 111(8):1046-
1053.
Chapter 6: Results paper four 133
Chapter 6: Results paper four
Temperature variability and childhood pneumonia: an
ecological study
Zhiwei Xu, Wenbiao Hu, Shilu Tong
Citation: Xu Z, Hu W, Tong S (2014). Temperature variability and childhood
pneumonia: an ecological study. Environmental Health, 13(1):51.
134 Chapter 6:Results paper four
Abstract
Background: Few data on the relationship between temperature variability and
childhood pneumonia are available. This study attempted to fill this knowledge gap.
Methods: A quasi-Poisson generalized linear regression model combined with a
distributed lag non-linear model was used to quantify the impacts of diurnal
temperature range (DTR) and temperature change between two neighbouring days
(TCN) on emergency department visits(EDVs) for childhood pneumonia in Brisbane,
from 2001 to 2010, after controlling for possible confounders.
Results: An adverse impact of TCN on EDVs for childhood pneumonia was
observed, especially in winter, and the magnitude of this impact increased greatly
from the first five years (2001–2005) to the second five years (2006–2010). Children
aged 5–14 years, female children and Indigenous children were particularly
vulnerable to TCN impact. However, there was no significant association between
DTR and EDVs for childhood pneumonia.
Conclusions: As climate change progresses, unstable weather pattern may occur,
and parents and caregivers of children should be aware of the high risk of pneumonia
posed by big TCN and take precautionary measures to protect children, especially
those with a history of respiratory diseases.
135
Chapter 6:Results paper four 135
6.1 Introduction
Pneumonia is the top cause of mortality in children under five years (Walker et al.
2013). It is estimated that in 2010, worldwide, there were 120 million episodes of
pneumonia in children younger than five (Walker et al. 2013). Pneumonia is highly
preventable, and hence it is particularly important to explore the risk factors which
drive the incidence of pneumonia and further to prevent children from being exposed
to these risk factors.
Many nutritional, socioeconomic and environmental factors are involved in the
occurrence of pneumonia (Black et al. 2008; Fonseca et al. 1996; Paynter et al.
2010). The possible impact of climatic factors on pneumonia transmission has
attracted increasing research attention (Paynter et al. 2010; Paynter et al., 2013).
Both high and low temperatures have been reported to be associated with increased
pneumonia incidence (Ebi et al. 2001; Green et al. 2010). However, the potential
impact of temperature variability on childhood pneumonia has not been researched
yet, even though big temperature changes may stress on the respiratory system (Song
et al. 2008).
There are several ways to define temperature variability (Karl et al. 1995). For
example, the difference in daily maximum and minimum temperatures (i.e., diurnal
temperature range (DTR)) (Kan et al. 2007), and the mean temperature difference
from one day to the next (i.e., temperature change between two neighbouring days
(TCN)) (Guo et al. 2011; Lin et al. 2013). Previous studies have highlighted that big
DTR or TCN may pose a big threat to the respiratory system of human (Guo et al.
2011; Kan et al. 2007; Lin et al. 2013), especially for children (Xu et al. 2013c). We
hypothesized that great DTR or TCN might be associated with increase in childhood
136 Chapter 6:Results paper four
pneumonia cases, and we used the data on emergency department visits (EDVs) for
childhood pneumonia in Brisbane from 2001 to 2010 to test our hypothesis.
6.2 Methods
Data collection
Data on EDVs from 1st January 2001 to 31st December 2010 classified according to
the International Classification of Diseases, 9th version and10th version (ICD 9 and
10) were supplied by Queensland Health. The details of the Brisbane data (selected
hospitals and covered regions, etc.) have been clarified in Chapter 4. We extracted
those cases coded as pneumonia (ICD 9 codes: 480–486; ICD 10 codes: J12–J18) in
children aged 0–14 years. Data on climate variables, including maximum and
minimum temperatures, rainfall and relative humidity, were obtained from Australian
Bureau of Meteorology. DTR was calculated as daily maximum temperature minus
daily minimum temperature (Kan et al. 2007). Daily mean temperature was the
average of daily maximum and minimum temperatures, and TCN was calculated as
mean temperature of the current day minus mean temperature of the previous day
(Guo, et al., 2011). Data on air pollutants, including daily average particular matter ≤
10 µm (PM10) (µg/m3), daily average nitrogen dioxide (NO2) (µg/m3) and daily
average ozone (O3) (ppb), were retrieved from the Queensland Department of
Environment and Heritage Protection.
Data analysis
Distributed lag non-linear model (DLNM) was developed to incorporate both lagged
and the non-linear effects of temperature on mortality or morbidity (Gasparrini et al.
2010). The detailed mechanism underlying DLNM has been introduced in (Xu et al.
2013d). A quasi-Poisson generalized linear regression combined with DLNM was
137
Chapter 6:Results paper four 137
used to quantify the association between DTR (or TCN) and EDVs for childhood
pneumonia.
Yt ~ quasiPoisson(μt)
Log (μt) = α + βDTRt,l (βTCNt,l)+ns(Tt,l, 3)+ns(RHt, 3) + ns(PM10t, 3) + ns(O3t, 3)
+ ns(NO2t, 3) +ns(Timet,8)+ η1Holiday + η2Day of Weekt
Where t is the day of the observation; Yt is the observed daily childhood pneumonia
on day t; α is the model intercept; DTR t,l is a matrix obtained by applying DLNM to
DTR, TCN t,l is a matrix obtained by applying DLNM to TCN, Tt,l is a matrix
obtained by applying the DLNM to temperature; β is vector of coefficients for Tt,l,
and l is the lag days; ns(RHt, 3) is a natural cubic spline with three degrees of
freedom for relative humidity; ns(PM10t, 3) is a natural cubic spline with three
degrees of freedom for PM10; ns(O3t, 3) is a natural cubic spline with three degrees of
freedom for O3; ns(NO2t, 3) is a natural cubic spline with three degrees of freedom
for NO2; ns(Timet,8) is a natural cubic spline with eight degrees of freedom per year
for long-term trend and seasonality; Holiday is the public holiday, and Day of Weekt
is the categorical day of the week with a reference day of Sunday.
Previous studies have revealed that there might be a lagged effect of temperature
variability on human health, and the relationship between temperature variability and
respiratory diseases appears to be non-linear (Guo et al. 2011; Lin et al. 2013; Xu et
al. 2013c; Xu et al. 2013d). Thus, we used DLNM to incorporate the non-linear and
lagged effect (Gasparrini, et al., 2010). Specifically, DTR (or TCN) and lag were
incorporated using a “natural cubic spline–natural cubic spline” approach. The model
included lags up to 21 days for DTR (or TCN) and mean temperature (Xu et al.
2013a). For all other confounders (i.e., relative humidity, PM10, O3, and NO2), we
used lags up to 10 days.
138 Chapter 6:Results paper four
All data analysis was conducted using the R statistical environment (v 2.15), and
“dlnm” package was used to fit the regression. In the sensitivity analysis, we changed
the df for DTR, TCN and time. We also excluded 2009 data as there was a big
pneumonia spike in 2009.
6.3 Results
Figure 6-1 shows the daily distributions of childhood pneumonia, mean temperature,
DTR and TCN, revealing that there was a seasonal trend in childhood pneumonia,
mean temperature, and DTR. The 2009 pneumonia peak (due to H1N1 flu pandemic)
is also revealed in this figure. Table 6-1 depicts the Spearman correlations between
climate variables, air pollutants and childhood pneumonia. There was a negative
correlation between DTR and mean temperature (r= -0.44, P<0.01). TCN was
positively correlated with mean temperature (r= 0.14, P<0.01). Further, DTR was
positively correlated with childhood pneumonia (r=0.21, P<0.01), while no
significant correlation between TCN and childhood pneumonia was observed.
139
Chapter 6:Results paper four 139
010
2030
4050
pneu
mon
ia10
2030
tem
pera
ture
05
1020
30
DTR
-10
-50
5
2002 2004 2006 2008 2010
TCN
Time
Figure 6-1. The daily distributions of EDVs for paediatric pneumonia, mean
temperature, DTR and TCN in Brisbane, from 2001 to 2010
140 Chapter 6:Results paper four
Table 6-1. Spearman’s correlation between daily weather conditions, air pollutants and paediatric pneumonia in Brisbane, Australia, from 2001–2010
Mean temperature DTR† TCN† Relative
humidity Rainfall PM10 O3 NO2 Pneumonia
Mean temperature 1.00
DTR† -0.44* 1.00
TCN† 0.14* 0.03 1.00
Relative humidity -0.06* -
0.33* 0.09* 1.00
Rainfall 0.16* -0.49* -0.03 0.38* 1.00
PM10 0.16* 0.31* 0.08* -0.29* -0.31* 1.00
O3 0.03 0.20* 0.03 -0.28* -0.07* 0.31* 1.00
NO2 -0.67* 0.42* -0.02 0.13* -0.14* 0.01 -0.10* 1.00
Pneumonia -0.37* 0.21* -0.01 0.03 0.06* 0.07* 0.10* 0.28* 1.00
* P<0.01; DTR, Diurnal temperature range; TCN, temperature change between two neighbouring days
141
Chapter 6:Results paper four 141
Figure 6-2 shows the exposure-response relationship between temperature variability
and childhood pneumonia. No significant association between DTR and childhood
pneumonia was observed. While, negative TCN (< -2 °C), meaning a big
temperature decrease from one day to the next, increased the risk of childhood
pneumonia. There were more than 50 days every year with TCN below -2 °C
(|TCN|>2), with most days with temperature drop > 2 °C occurring in the second half
of each year (June to December) (Figure 6-3). Figure 6-4 shows the pattern of lagged
effect of TCN impact on childhood pneumonia, revealing that TCN effect lasted for
nearly three weeks. It can be seen from Figure 6-5 that children aged 5–14 years,
female children and Indigenous children appeared to be more vulnerable to the TCN
impact compared with children of other groups.
As there was a distinct seasonality in childhood pneumonia, with peak in winter, we
specifically analysed the TCN impact on childhood pneumonia in summer
(December, January, and February) and winter (June, July and August), and found
that this impact mainly occurred in winter (Figure 6-6).
To test whether there was a change over time in the effect of TCN on childhood
pneumonia, we splitted the ten years into two periods (2001–2005 and 2006–2010).
It can be seen from Figure 6-7 that the effect of TCN on childhood pneumonia during
period two was much greater than it was during period one.
Figure 6-8 shows the results excluding the 2009 data, indicating that the magnitude
of TCN effect on childhood pneumonia in Brisbane declined greatly without 2009
pneumonia spike, although the shape of the TCN-pneumonia relationship was
similar.
142 Chapter 6:Results paper four
0 5 10 15 20 25 30
0.5
1.5
2.5
3.5
DTR (°C)
RR
(ped
iatri
c pn
eum
onia
)
-10 -5 0 5
05
1015
20
TCN (°C)
RR
(ped
iatri
c pn
eum
onia
)
Figure 6-2. The overall effects of DTR and TCN on paediatric pneumonia in
Brisbane, from 2001 to 2010
143
Chapter 6:Results paper four 143
Figure 6-3. Monthly average number of days with TCN < -2 °C (|TCN|>2)
144 Chapter 6:Results paper four
0 5 10 15 20
1.00
1.05
1.10
1.15
1.20
TCN effect
Lag
RR
(ped
iatri
c pn
eum
onia
)
Figure 6-4. The lagged effect of TCN on childhood pneumonia
145
Chapter 6:Results paper four 145
Figure 6-5. The effect of TCN on the total-, age-, gender- and ethnic-specific
childhood pneumonia in Brisbane, from 2001 to 2010 (Relative risk: The risk of EDVs for childhood pneumonia on days with temperature drop =5.7 °C relative to
days with temperature drop= 2.0 °C)
146 Chapter 6:Results paper four
-10 -5 0 5
02
46
810
Summer
TCN (°C)
RR
(ped
iatri
c pn
eum
onia
)
-10 -5 0 5
05
1015
20
Winter
TCN (°C)
RR
(ped
iatri
c pn
eum
onia
)
Figure 6-6. The effects of TCN on childhood pneumonia in summer and winter
147
Chapter 6:Results paper four 147
-5 0 5
02
46
810
2001-2005
TCN (°C)
RR
(ped
iatri
c pn
eum
onia
)
-5 0 5
02
46
810
2006-2010
TCN (°C)
RR
(ped
iatri
c pn
eum
onia
)
Figure 6-7. The overall effects of TCN on childhood pneumonia during two periods
148 Chapter 6:Results paper four
-10 -5 0 5
02
46
810
TCN (°C)
RR
(ped
iatri
c pn
eum
onia
)
Figure 6-8. The overall effect of TCN on childhood pneumonia in Brisbane, from
2001 to 2010 (excluding 2009)
149
Chapter 6:Results paper four 149
6.4 Discussion
This study has quantified the impacts of both DTR and TCN on childhood
pneumonia and yielded several notable findings. A big temperature decrease from
one day to the next (|TCN|>2) may increase the EDVs for childhood pneumonia, and
this effect lasted for around three weeks. Every year, children were at the risk of big
TCNs for more than 50 days, and big TCNs mainly occurred in the second half of
each year, especially in winter. Children aged 5–14 years, female children and
Indigenous children were at particular risk. Further, there was a change in the effect
of TCN on childhood pneumonia over time, and big TCN may contribute to the 2009
pneumonia spike. No significant relationship between DTR and childhood
pneumonia was observed.
As climate change progresses, not only the global surface average temperature, but
also the frequency of unstable weather patterns (e.g., sharp increase/decrease in
temperature) will increase, as projected (Hansen et al. 2012), which poses a
significant challenge to public health sectors. Children are particularly vulnerable to
both extreme temperatures (Xu et al. 2013b) and temperature variation (Xu et al.
2013c), due partially to their relatively less-developed thermoregulation capability
(Xu et al. 2012). In this study, we found a sharp temperature drop was followed by
significantly increased EDVs for childhood pneumonia, and the TCN impact lasted
for roughly three weeks. When big temperature change happens, children’s
temperature regulation system will firstly cope with the adverse impact. While,
temperature change which exceeds certain limits may trigger the symptoms of
existing respiratory diseases in a few days, and it may take another a few days from
the occurrence of symptoms to seeking for healthcare. The greater effect of TCN on
150 Chapter 6:Results paper four
childhood pneumonia in winter than summer was also detected in this study, which
may be due to the greater number of extreme TCNs in winter.
With regards to prior studies looking at the impact of TCN on human health, some
studies have observed significantly increased respiratory-related mortality associated
with big TCN (Lin et al. 2013) while some others did not find significant results
(Guo et al. 2011). Our study stands out of previous studies by specifically focusing
on childhood pneumonia and controlling for a range of possible confounders. Most
previous studies assessing the impact of DTR on mortality or morbidity due to
respiratory diseases found significant effects of DTR (Cheng et al. 2014), while our
study did not find a significant relation between DTR and childhood pneumonia. The
disparity between findings of prior studies and our study may be explained by
different socioeconomic factors which may modify TCN effects on health (Malik et
al. 2012), different thermoregulatory capacity between children and adults (Xu et al.
2012), and different health seeking patterns between people of different regions.
Immunity plays an important role in respiratory infections (Curriero et al. 2002), and
infectious respiratory diseases are mainly caused by the infections of pathogens.
Weather change may affect humoral and cellular immunity (Bull 1980) as well as the
survival and replication of pathogens. From 2001 to 2010, every year, there were
more than 50 days with a great negative TCN, especially in winter, meaning that
parents and caregivers should take precautionary measures to protect their children,
especially those children with pre-existing respiratory conditions, if weather
forecasting reports a big temperature decrease in the coming couple of days.
In this study, we also found age, gender and Indigenous-status modified the
relationship between TCN and pneumonia. The school-aged children (5–14 years)
were more sensitive to TCN compared with children in other age groups, which
151
Chapter 6:Results paper four 151
might be because they play outdoors more often and thus exposed more to the
outdoor temperature change. Another reason of school-aged children’s greater
vulnerability to big temperature change may be because school-aged children,
especially girls, are more likely to wear trendy clothes, and sometimes focus more on
the style than the thickness of those clothes. The difference in vulnerability to TCN
between two genders may be due to their body composition (Frascarolo et al. 1990),
though some researchers argued that differences in the effect of temperature on
males and females is variable among different locations and populations (Basu
2009). Our results also suggest that Indigenous children were more sensitive to TCN
effect compared with non-Indigenous children. Previous studies have reported that
the burden of pneumonia in Indigenous children is 10 to 20 fold higher than non-
Indigenous children, and they have longer hospital admissions and are more likely to
have multiple admissions with pneumonia (Burgner and Richmond 2005).
Indigenous children have limited access to infrastructures, and experience more
poverty than non-Indigenous children, possibly resulting in their greater vulnerability
to both extreme temperatures and temperature variation (Ford 2012).
We found that the effect of TCN on childhood pneumonia during 2006–2010 was
much greater than it was during 2001–2005, which may partially be attributable to
the TCN impact on 2009 pneumonia spike. This finding indicates that children might
be impacted more by sharp temperature decrease in the future if unstable weather
patterns occur as projected. Elucidating the impact of temperature variability on
children’s health is essential for the improvement of public health. The findings of
our study not only remind parents and children’s caregivers to take good care of
children in the days of big TCN, but also imply that government should take
temperature variability into account while developing the future early warning
152 Chapter 6:Results paper four
systems because TCN may substantially impact children’s health, independent of
temperature.
This study has two major strengths. First, this is, to our knowledge, the first study to
look at the impact of DTR and TCN on childhood pneumonia. Second, the change
over time in the TCN effect on childhood pneumonia which we observed in this
study may encourage future studies to explore the temporal variability of TCN
impact on children’s health in a longer time period. Two weaknesses should be
acknowledged. First, this is a one city study, which means further generalization of
our findings to regions of different climates should be cautious. Second, the
aggregated data on temperature and EDVs for childhood pneumonia we used may
result in some biases in exposure and/or outcome measures.
6.5 Conclusions
No relationship between DTR and childhood pneumonia was observed. A sharp
temperature decrease from one day to the next had an adverse impact on childhood
pneumonia, especially in winter, and the magnitude of this impact increased in recent
years. Big TCNs may have contributed to the 2009 pneumonia peak in Brisbane.
Children aged 5–14 years, female children and Indigenous children are at particular
risk. As climate change continues, unstable weather patterns may be more frequent,
and the findings of this study have important implications for public health policy to
protect children from the impact of pneumonia.
6.6 References
Basu R. 2009. High ambient temperature and mortality: a review of epidemiologic
studies from 2001 to 2008. Environ Health 8(1):40.
153
Chapter 6:Results paper four 153
Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, et al. 2008.
Maternal and child undernutrition: global and regional exposures and health
consequences. Lancet 371(9608):243-260.
Bull G. 1980. The weather and deaths from pneumonia. Lancet 1(8183):1405-1408.
Burgner D, Richmond P. 2005. The burden of pneumonia in children: an Australian
perspective. Paediatr Respir Rev 6(2):94-100.
Cheng J, Xu Z, Zhu R, Wang X, Jin L, Song J, et al. 2014. Impact of diurnal
temperature range on human health: a systematic review. Int J Biometeorol,
doi: 10.1007/s00484-014-0797-5
Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. 2002. Temperature
and mortality in 11 cities of the eastern United States. Am J Epidemiol
155(1):80-87.
Ebi K, Exuzides K, Lau E, Kelsh M, Barnston A. 2001. Association of normal
weather periods and El Niño events with hospitalization for viral pneumonia
in females: California, 1983-1998. Am J Public Health 91(8):1200-1208.
Fonseca W, Kirkwood B, Victora C, Fuchs S, Flores J, Misago C. 1996. Risk factors
for childhood pneumonia among the urban poor in Fortaleza, Brazil: a case--
control study. Bull World Health Organ 74(2):199-208.
Ford JD. 2012. Indigenous health and climate change. Am J Public Health
102(7):1260-1266.
Frascarolo P, Schutz Y, Jequier E. 1990. Decreased thermal conductance during the
luteal phase of the menstrual cycle in women. J Appl Physiol 69(6):2029-
2033.
Gasparrini A, Armstrong B, Kenward MG. 2010. Distributed lag non-linear models.
Stat Med 29(21):2224-2234.
154 Chapter 6:Results paper four
Green R, Basu R, Malig B, Broadwin R, Kim J, Ostro B. 2010. The effect of
temperature on hospital admissions in nine California counties. Int J Public
Health 55(2):113-121.
Guo Y, Barnett AG, Yu W, Pan X, Ye X, Huang C, et al. 2011. A large change in
temperature between neighbouring days increases the risk of mortality. PLoS
ONE 6(2):e16511.
Hansen J, Sato M, Ruedy R. 2012. Perception of climate change. Proc Natl Acad Sci
U S A 109(37):E2415-23
Kan H, London SJ, Chen H, Song G, Chen G, Jiang L, et al. 2007. Diurnal
temperature range and daily mortality in Shanghai, China. Environ Res
103(3):424-431.
Karl TR, Knight RW, Plummer N. 1995. Trends in high-frequency climate
variability in the twentieth century. Nature 377(6546):217-220.
Lin H, Zhang Y, Xu Y, Xu X, Liu T, Luo Y, et al. 2013. Temperature changes
between neighboring days and mortality in summer: A distributed lag non-
linear time series analysis. PLoS ONE 8(6):e66403.
Malik S, Awan H, Khan N. 2012. Mapping vulnerability to climate change and its
repercussions on human health in Pakistan. Global Health 8(1):31.
Paynter S, Ware RS, Weinstein P, Williams G, Sly PD. 2010. Childhood pneumonia:
a neglected, climate-sensitive disease? Lancet 376(9755):1804-1805.
Paynter S, Weinstein P, Ware RS, Lucero MG, Tallo V, Nohynek H, et al. 2013.
Sunshine, rainfall, humidity and child pneumonia in the tropics: time-series
analyses. Epidemiol Infect 141(06):1328-1336.
155
Chapter 6:Results paper four 155
Song G, Chen G, Jiang L, Zhang Y, Zhao N, Chen B, et al. 2008. Diurnal
temperature range as a novel risk factor for COPD death. Respirology
13(7):1066-1069.
Walker CLF, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. 2013. Global
burden of childhood pneumonia and diarrhoea. Lancet 381(9875):1405-1416.
Xu Z, Etzel RA, Su H, Huang C, Guo Y, Tong S. 2012. Impact of ambient
temperature on children's health: A systematic review. Environ Res 117:120-
131.
Xu Z, Hu W, Su H, Turner LR, Ye X, Wang J, et al. 2013a. Extreme temperatures
and paediatric emergency department admissions. J Epidemiol Community
Health 68(4):304-311.
Xu Z, Huang C, Hu W, Turner LR, Su H, Tong S. 2013b. Extreme temperatures and
emergency department admissions for childhood asthma in Brisbane,
Australia. Occup Environ Med 70(10):730-735.
Xu Z, Huang C, Su H, Turner L, Qiao Z, Tong S. 2013c. Diurnal temperature range
and childhood asthma: a time-series study. Environ Health 12(1):12.
Xu Z, Huang C, Turner LR, Su H, Qiao Z, Tong S. 2013d. Is diurnal temperature
range a risk factor for childhood diarrhea? PLoS ONE 8(5):e64713.
157
Chapter 7:Results paper five 157
Chapter 7: Results paper five
The impact of temperature variability on childhood
diarrhoea
Zhiwei Xu, Wenbiao Hu, Shilu Tong
158 Chapter 7:Results paper five
Abstract
Background: The association between temperature and diarrhoea has been well
researched. However, the relationship between temperature variability and childhood
diarrhoea has not been unveiled.
Methods: A quasi-Poisson generalized linear model combined with a distributed lag
non-linear model (DLNM) was used to examine the effect of temperature variability
(Diurnal temperature range (DTR), and temperature change between two
neighbouring days (TCN)) on EDVs for childhood diarrhoea, controlling for
potential confounders.
Results: High DTR and TCN were significantly associated with an increase in EDVs
for childhood diarrhoea in Brisbane. Every year, from May to September, especially
July, children were at high risk posed by high DTR and low TCN. Male children
were particularly vulnerable to the adverse impacts of DTR and TCN on diarrhoea.
Conclusions: As climate change continues, unstable weather patterns may occur
more often, and it would be necessary to develop effective strategies to protect
children from being harmed by climate impacts.
159
Chapter 7:Results paper five 159
7.1 Introduction
There is a wide spread consensus that climate is changing due to anthropogenic
activities, and climate change poses a huge threat to human well-being (McMichael
2013). Children are particularly vulnerable to the adverse impact of climate change
(Sheffield and Landrigan 2011). Paediatric infectious diseases (e.g., diarrhoea) may
increase due to the increasing global surface average temperature (Xu et al. 2012b).
The effect of high temperature on childhood diarrhoea has been well-documented
(Checkley et al. 2000; Green et al. 2010). However, the relationship between
temperature variability and childhood diarrhoea has not been unveiled yet, even
though the frequency of unstable weather patterns will increase as climate change
progresses (Epstein 2005), and big temperature variability may stress on immune
system (Bull 1980) and jeopardize children’s resistance to intestinal aetiological
agents.
Temperature variability can be defined in several ways. Diurnal temperature range
(DTR), temperature change from one day to the next (TCN), and the standard
deviation of temperature in a certain period of time, are the most frequently used
definitions (Karl et al. 1995). Herein, we adopted the first two temperature indexes
(i.e., DTR and TCN) as the indicators of temperature variability. We hypothesized
that big temperature variability may increase the cases of childhood diarrhoea, and
used data on emergency department visits (EDVs) in Brisbane, Australia, from 2001
to 2010 to testify our hypothesis.
7.2 Methods
Data collection
160 Chapter 7:Results paper five
Data on EDVs were supplied by Queensland Health. The details of the data (selected
hospitals and covered regions, etc.) have been clarified in Chapter 4. We selected as
the following codes for diarrhoea in children aged 0–14 years: ICD-9 codes: 001–
003, 004, 005, 006.0–006.2, 007.0–007.5, 008–009; ICD–10 codes: A00–A03, A04,
A05, A06.0–A06.3, A06.9, A07.0–A07.2, A07.9, A08–A09. Data on daily maximum
temperature, minimum temperature, relative humidity, and rainfall were obtained
from Australian Bureau of Meteorology. DTR was calculated as daily maximum
temperature minus daily minimum temperature (Kan et al. 2007). We averaged the
daily maximum temperature and minimum temperature to get the daily mean
temperature, and TCN was calculated as mean temperature of the current day minus
mean temperature of the previous day (Guo et al. 2011).
Data analysis
A quasi-Poisson generalized linear model combined with a distributed lag non-linear
model (DLNM) was used to examine the effect of temperature variability
(DTR/TCN) on EDVs for childhood diarrhoea. A natural cubic spline with four
degrees of freedom (df) was used to capture a potentially non-linear effect of
DTR/TCN on childhood diarrhoea. Mean temperature was controlled for using a
“natural cubic spline–natural cubic spline” approach with lags up to 21 days (Xu
2013a). Rainfall and relative humidity were controlled for by using a natural cubic
spline with four df. Seasonal patterns and long-term trends were controlled by using
a natural cubic spline with seven df per year of data. Day of week was controlled for
as a categorical variable. Public holiday was also controlled for in the model.
There might be a lagged effect of temperature variability on human health, and we
used DLNM to incorporate the non-linear and lagged effect (Gasparrini et al. 2010).
161
Chapter 7:Results paper five 161
Specifically, temperature variability (DTR/TCN) and lag were incorporated using a
“natural cubic spline–natural cubic spline” approach with lags up to 10 days.
All data analysis was conducted using the R statistical environment (v 2.15), and
“dlnm” package was used to fit the regression. Sensitivity analysis was conducted by
changing the df for DTR, TCN and time.
7.3 Results Table 7-1 presents the summary statistics of climate variables and EDVs for
childhood diarrhoea. The mean value of DTR was 14.7 °C (Range: 0.7–30.2 °C), and
the mean value of TCN was 0 °C (-9.5–6.8 °C). The daily average value of EDVs for
childhood diarrhoea was 15.9, with a range of 0 to 91. Figure 7-1 shows the daily
distribution of mean temperature, DTR, TCN and EDVs for childhood diarrhoea.
There was a distinct seasonal trend for childhood diarrhoea, mean temperature and
DTR.
Chapter 7: Results paper five 162
Table 7-1. Summary statistics for climatic variables, air pollutants and paediatric diarrhoea in Brisbane, Australia, 2001–2010
Variables Mean SD Min
Percentile
Max
25 50 75
Mean temperature
(°C) 20.1 4.9 6.9 16.2 20.4 23.8 34.9
DTR (°C)† 14.7 5.0 0.7 11.2 14.7 18.4 30.2
TCN (°C)† 0 2.0 -9.5 -1.0 0.2 1.4 6.8
Relative humidity
(%) 65.0 15.0 13.0 56.0 65.0 75.0 100.0
Rainfall (mm) 2.2 8.3 0 0 0 0.4 149.0
Diarrhoea 15.9 8.8 0 10.0 14.0 20.0 91.0
†DTR, Diurnal temperature range; TCN, temperature change between two neighbouring days
Chapter 7: Results paper five 163
020
4060
80
diar
rhea
1015
2025
3035
mea
ntem
p0
510
1520
2530
dtr
-10
-50
5
2002 2004 2006 2008 2010
tcn
Time
Figure 7-1. The daily distributions of EDVs for paediatric diarrhoea, mean temperature, DTR and TCN in Brisbane, from 2001 to 2010
164 Chapter 7:Results paper five
Table 7-2 shows the Spearman correlations between climate variables and EDVs for
childhood diarrhoea, revealing that childhood diarrhoea was positively correlated
with both DTR and TCN. No correlation value was over 0.5, meaning that multi-
collinearity was not likely an issue in the subsequent modelling.
Table 7-2. Spearman’s correlation between daily weather conditions, air pollutants and paediatric diarrhoea in Brisbane, Australia, from 2001–2010
Mean temperature DTR TCN Relative
humidity Rainfall Diarrhoea
Mean temperature
1.00
DTR† -0.44* 1.00
TCN† 0.14* 0.03 1.00
Relative humidity
-0.06* -0.33* 0.09* 1.00
Rainfall 0.16* -0.49* -0.03 0.38* 1.00
Diarrhoea -0.02 0.09* 0.05* -0.11* 0.01 1.00
* P<0.01; DTR, Diurnal temperature range; TCN, temperature change between two neighbouring days
165
Chapter 7:Results paper five 165
The cumulative effects of DTR/TCN on EDVs for childhood diarrhoea were shown
in Figure 7-2. DTR over 17 °C or TCN lower than -2 °C, was associated with a
significant increase in EDVs for childhood diarrhoea. Every year, there were more
than 50 days with TCN below -2 °C, and more than 120 days with DTR over 17 °C
during 2001 and 2010. Figure 7-3 reveals that days with both high DTR and TCN
occurred mainly from May to September (especially July).
The effects of DTR/TCN on EDVs for childhood diarrhoea by age, gender and
Indigenous status was presented in Figure 7-4. Male children appeared to be more
vulnerable to both DTR and TCN compared with female children.
166 Chapter 7:Results paper five
0 5 10 15 20 25 30
0.5
1.0
1.5
2.0
DTR (°C)
RR
(ped
iatri
c di
arrh
ea)
-5 0 5
1.0
2.0
3.0
4.0
TCN (°C)
RR
(ped
iatri
c di
arrh
ea)
Figure 7-2. The overall effects of DTR and TCN on paediatric diarrhoea in Brisbane, from 2001 to 2010
167
Chapter 7:Results paper five 167
1 2 3 4 5 6 7 8 9 10 11 12
DTR>17°C & TCN<-2°C
Month
Num
ber o
f day
s
05
1015
2025
30
Figure 7-3. The monthly distribution of days when DTR > 17 °C and TCN < -2 °C
168 Chapter 7:Results paper five
Figure 7-4a. The effect of DTR on the total-, age-, gender- and ethnic-specific
childhood diarrhoea in Brisbane, from 2001 to 2010
Figure 7-4b. The effect of TCN on the total-, age-, gender- and ethnic-specific
childhood diarrhoea in Brisbane, from 2001 to 2010
169
Chapter 7:Results paper five 169
7.4 Discussion As climate change progresses, not only the global surface average temperature, but
also the frequency of unstable weather patterns (e.g., sharp increase/decrease in
temperature) will increase, as projected (Epstein 2005), which poses a big challenge
to public health sectors. Children are particularly vulnerable to both extreme
temperatures (Xu et al. 2013b) and temperature variation (Xu et al. 2013c), due
partially to their relatively less-developed thermoregulation system (Xu et al. 2012a).
This study found that high DTR and TCN were significantly associated with an
increase in EDVs for childhood diarrhoea in Brisbane. Every year, from May to
September, especially July, children were at high risk posed by big DTR and TCN.
Male children were particularly vulnerable to the adverse impact of DTR and TCN
on diarrhoea.
The results of this study support our hypothesis that there was a significant
association between DTR and childhood diarrhoea. Currently, the exact mechanism
by which exposure to a large DTR can increase the risk of childhood morbidity
remains largely unknown. Bull argued that sudden changes in weather conditions
may affect either humoral or cellular immunity (Bull 1980). Very young children
have a relatively immature immune system (Gerba et al. 1996) and low self-care
capacity (Xu 2012a), which might result in a greater vulnerability to temperature
change. Studies have found that sudden changes in the temperature of inhaled air are
associated with the release of inflammatory mediators by mast cells (Graudenz et al.
2006; Togias et al. 1985), which could also be related to higher diarrhoea prevalence
(Feng et al. 2007; Ramsay et al. 2010). Further study is necessary to determine which
biomarkers are affected by DTR.
170 Chapter 7:Results paper five
In this study, we also found a sharp temperature drop was followed by significantly
increased EDVs for childhood diarrhoea. With regards to prior studies looking at the
impact of TCN on human health, some studies have observed significantly increased
mortality associated with big TCN (Guo et al. 2011; Lin et al. 2013). In this study,
the TCN effect increased rapidly below -2 °C, highlighting that both parents and
medical staff should be made aware of the particularly high risk posed by large
temperature drop between two days and diarrhoea-related morbidity in children. In
Brisbane during the study period, children were exposed to the risk of both large
DTR and TCN from May to September, indicating that caregivers of children should
take precautionary measures in winter to protect children from being attacked by
large DTR and TCN.
This study is the first study to examine the relation between childhood diarrhoea and
TCN. Additionally, we used advanced methods to assess this association, which was
able to incorporate not only non-linear effects of temperature and DTR but also
lagged effects in the one model. Several limitations of this study should be
acknowledged. This is an ecological study, and therefore some bias due to exposure
misclassification may be inevitable. Only one city was examined, which might limit
the generalizability of our results. However, the findings of this research may
encourage a large scale, multi-centre study in the future.
7.5 Conclusions This study demonstrates a significant relationship between DTR/TCN and childhood
diarrhoea. Large DTR and TCN may trigger and increase childhood diarrhoea among
children. As climate change continues, unstable weather patterns may occur more
171
Chapter 7:Results paper five 171
often, and it would be necessary to develop effective strategies to protect children
from being harmed by climate impacts.
7.6 References Bull G. 1980. The weather and deaths from pneumonia. Lancet 1(8183):1405-1408.
Checkley W, Epstein L, Gilman R, Figueroa D, Cama R, Patz J, et al. 2000. Effect of
El Niño and ambient temperature on hospital admissions for diarrhoeal
diseases in Peruvian children. Lancet 355(9202):442-450.
Epstein PR. 2005. Climate change and human health. N Engl J Med 353(14):1433-
1436.
Feng, BS, He SH, Zheng PY, Wu L, Yang PC. 2007. Mast cells play a crucial role in
staphylococcus aureus peptidoglycan-induced diarrhea. Am J Pathol
171(2):537-547.
Gasparrini A, Armstrong B, Kenward MG. 2010. Distributed lag non-linear models.
Stat Med 29(21):2224-2234.
Gerba C, Rose J, Haas C. 1996. Sensitive populations: who is at the greatest risk? Int
J Food Microbiol 30(1-2):113-123.
Graudenz GS, Landgraf RG, Jancar S, Tribess A, Fonseca SG, Faé KC, et al. 2006.
The role of allergic rhinitis in nasal responses to sudden temperature changes.
J Allergy Clin Immunol 118(5):1126-1132.
Green R, Basu R, Malig B, Broadwin R, Kim J, Ostro B. 2010. The effect of
temperature on hospital admissions in nine California counties. Int J Public
Health 55(2):113-121.
172 Chapter 7:Results paper five
Guo Y, Barnett AG, Yu W, Pan X, Ye X, Huang C, et al. 2011. A large change in
temperature between neighbouring days increases the risk of mortality. PLoS
ONE 6(2):e16511.
Kan H, London SJ, Chen H, Song G, Chen G, Jiang L, et al. 2007. Diurnal
temperature range and daily mortality in Shanghai, China. Environ Res
103(3):424-431.
Karl TR, Knight RW, Plummer N. 1995. Trends in high-frequency climate
variability in the twentieth century. Nature 377(6546):217-220.
Lin H, Zhang Y, Xu Y, Xu X, Liu T, Luo Y, et al. 2013. Temperature changes
between neighboring days and mortality in summer: A distributed lag non-
linear time series analysis. PLoS ONE 8(6):e66403.
McMichael AJ. 2013. Globalization, climate change, and human health. N Engl J
Med 368(14): 1335-1343.
Ramsay DB, Stephen S, Borum M, Voltaggio L, Doman, DB. 2010. Mast cells in
gastrointestinal disease. Gastroenterol Hepatol 6(12):772-777.
Sheffield P, Landrigan P. 2011. Global climate change and children's health: threats
and strategies for prevention. Environ Health Perspect 119(3):291-298.
Togias AG, Naclerio RM, Proud D, Fish JE, Adkinson NF, Kagey-Sobotka A, et al.
1985. Nasal challenge with cold, dry air results in release of inflammatory
mediators. Possible mast cell involvement. J Clin Invest 76(4):1375-1381.
Xu Z, Etzel RA, Su H, Huang C, Guo Y, Tong S. 2012a. Impact of ambient
temperature on children's health: A systematic review. Environ Res
117(0):120-131.
173
Chapter 7:Results paper five 173
Xu Z, Hu W, Su H, Turner LR, Ye X, Wang J, Tong S. 2013a. Extreme temperatures
and paediatric emergency department admissions. J Epidemiol Community
Health 68(4):304-311.
Xu Z, Huang C, Hu W, Turner LR, Su H, Tong S. 2013b. Extreme temperatures and
emergency department admissions for childhood asthma in Brisbane,
Australia. Occu Environ Med 70(10):730-735.
Xu Z, Huang C, Su H, Turner L, Qiao Z, Tong S. 2013c. Diurnal temperature range
and childhood asthma: a time-series study. Environ Health 12(1):12.
Xu Z, Sheffield PE, Hu W, Su H, Yu W, Qi X, et al. 2012b. Climate change and
children’s health—A call for research on what works to protect children. Int J
Environ Res Public Health 9(9):3298-3316.
Chapter 7: Results paper five 174
175
Chapter 8: Discussion and conclusions 175
Chapter 8: Discussion and conclusions
Previous chapters (Chapters three to seven) have presented the main results of each
study, as well as discussions and conclusions. In this chapter, I discuss the key
findings of this thesis as an integrated body of work, and explore the mechanisms
and implications of the findings and make recommendations for future research
direction.
8.1 Key findings
In the spatiotemporal analysis (Chapter four), I found children suffering pneumonia
and diarrhoea who were admitted to emergency departments in Queensland from
2007 to 2011 were mainly from central west, northwest and far north of Queensland.
Mount Isa was the region at highest risk for both childhood pneumonia and
diarrhoea.
The subsequent time-series studies (Chapters four and five) using satellite remote
sensing temperature data revealed that both high and low temperatures were
associated with an increase in emergency department visits (EDVs) for childhood
pneumonia and diarrhoea in Brisbane from 2001 to 2010. Heat waves and cold spells
had added effects on childhood pneumonia, and the magnitude of the effects
increased with the intensity and duration of heat waves and cold spells. Heat waves
also had an added effect on childhood diarrhoea. The high temperature effects on
both childhood pneumonia and diarrhoea experienced a decreasing trend from 2001
to 2010, while the cold effects on pneumonia and diarrhoea changed little over this
176 Chapter 8: Discussion and conclusions
period. Indigenous children appeared to be more vulnerable to the impacts of high
and low temperatures on pneumonia and diarrhoea than non-Indigenous children.
Several findings from the two studies examining the effects of temperature
variability (diurnal temperature range (DTR) and temperature change between two
neighbouring days (TCN)) on childhood pneumonia and diarrhoea (Chapters six and
seven) are noteworthy. Big TCNs (<-2 °C) were associated with increases in EDVs
for childhood pneumonia and diarrhoea, and big DTR may increase EDVs for
childhood diarrhoea. Every year, from May to September, especially July, children
were at high risk posed by big DTRs and TCNs.
8.2 Mechanisms and implications
Mining industry is one of the mainstay industries of Australia. Mount Isa, the city
located adjacent to Mount Isa Mines, is the largest emitter of sulphur dioxide, lead
and other metals in Australia (National Pollution Inventory, 2010). The adverse
impacts of exposure to environmental hazards on human health in Mount Isa,
especially increased blood lead level in children because of high lead exposure, have
attracted public health attention since early 1990s (Munksgaard et al. 2010).
Considerable evidence has shown that some of the negative health, intellectual and
socio-behavioural effects due to high blood lead level can be lifelong (Lanphear et al.
2005; Mark et al. 2011).
The results of our study indicated that there was a highest risk for childhood
pneumonia and diarrhoea in Mount Isa, suggesting that there might be some risk
factors associated with both pneumonia and diarrhoea in this area. The semi-arid
freshwater aquatic system contaminated by mining may cause more diarrhoea
177
Chapter 8: Discussion and conclusions 177
episodes in children (Mark et al. 2009), and air pollutants emitted by mining may
increase childhood pneumonia cases (Barnett et al. 2005). How to minimize the
health impacts of environmental hazard exposure among Mount Isa children has been
discussed by not only public health sector but also industry and government
authorities (Mark et al. 2010). Beyond environmental risk factors, social
disadvantages (Walker et al. 2013), as well as cultural barriers to health care
(McBain-Rigg and Veitch 2011), may also affect the health of children in Mount Isa.
Difference in access to health care matter, as do differences in lifestyle, but the key
determinants of social inequalities in health lie in the circumstances where children
are born, grow, live and work.
In the context of climate change, two pivotal steps of public health surveillance are to
monitor temperature-health exposure-response function and to monitor the
vulnerabilities (social vulnerabilities, environmental vulnerabilities and underlying
health conditions) (Pascal et al. 2012). It has been documented that climate change
may impact respiratory infections and diarrhoeal diseases in Australia (Harley et al.
2011). Thus, I conducted the time-series studies to quantify the effects of both
extreme temperatures and prolonged extreme temperatures on childhood pneumonia
and diarrhoea and to identify the populations vulnerable to these effects. In these
studies, I used the satellite remote sensing data, which have been proven to be more
accurate than monitor-based data especially for those areas without extensive
monitoring networks (Lee et al. 2012). Prior research mainly focused on quantifying
the impact of hot temperature on children’s health (Green et al. 2010). However, it is
essential to examine the balance of hot and cold temperature-related morbidity
because future increasing temperature may also reduce the occurrence of cold-related
diseases (Xu et al. 2012). In this study, I found the magnitude of the main effects of
178 Chapter 8: Discussion and conclusions
heat and cold on childhood pneumonia was similar, and the main effect of cold on
childhood diarrhoea was even greater than heat effect, indicating that the burden of
childhood pneumonia and diarrhoea in Brisbane due to the main effects of
temperature may not substantially increase as climate change progresses, even
though there are still a lot of uncertainties.
The findings of our study also suggested that the effects of heat on childhood
pneumonia and diarrhoea appeared to decline in the past decade, even though the
relatively short study period restricts us to fully unveil the long-term trend (e.g., over
100 years) in the relationship between heat and childhood pneumonia and diarrhoea.
The decreasing trend in the heat effects on childhood pneumonia and diarrhoea may
partially be explained by two reasons: First, children in Brisbane may have gradually
adapted to heat effect as Brisbane normally has a very long summer every year.
Second, the increasing use of air conditioning in recent years may also reduce
children’s exposure to heat (Ostro et al. 2010). In our study, heat waves and cold
spells were found to have added effects on childhood pneumonia, and heat waves
may also increase EDVs for childhood diarrhoea in Brisbane. It is projected that, as
climate change progresses, there will be more frequent, more intense and longer-
lasting heat waves (Meehl and Tebaldi 2004), and therefore, it is critical to develop a
comprehensive heat early warning system in Brisbane targeting to protect children
from the adverse impact of heat (Xu et al. 2014).
Indigenous children were found more vulnerable to the impacts of extreme
temperatures on pneumonia and diarrhoea compared with non-Indigenous children in
this thesis. Indigenous Australians are generally the least healthy population of all
Indigenous populations in the world (Australian Institute of Health and Welfare
2012), and they face enormous disadvantages compared with general Australians. An
179
Chapter 8: Discussion and conclusions 179
Indigenous child is 2.5 times more likely to be born into the lowest income family,
and has a one in two chance of living in a one-parent family when compared with the
general population (National Mental Health Commission. Australian Government.
2012), and all these factors make them vulnerable to environmental hazards. Climate
change may exacerbate current health disparities between Indigenous and non-
Indigenous children (Fritze et al. 2008), and hence future climate change adaptation
and mitigation plans aiming to prevent Australians from adverse impacts of extreme
temperatures should focus more on Indigenous children. The health status of
Australia’s Aboriginal and Torres Strait Islander children is a significant public
health concern. The overall hospital admission rate for acute lower respiratory
infections in Indigenous children of northwest Queensland was approximately four
times higher than the rate for non-Indigenous children (Janu et al. 2014). Climate
change may exacerbate current health disparities between Indigenous and non-
Indigenous children (Fritze et al. 2008), and hence future climate change adaptation
and mitigation plans aiming to protect Australian children from adverse impacts of
extreme temperatures should focus more on Indigenous communities, especially for
those who are living in the distant regions (e.g., Mount Isa).
Climate change will not only increase global surface mean temperature but also
impact temperature variability (Schar et al. 2004). It has been found that increasing
temperature variability largely contributed to the record-breaking heat wave affecting
the whole Europe in 2003 (Schar et al. 2004). There are three widely accepted ways
to define temperature variability: DTR (Xu et al. 2013), TCN (Lin et al. 2013) and
standard deviation of mean temperature within a certain period of time (one week,
one month or one year) (Xu et al. 2014a). Our studies looking at the impacts of DTR
and TCN on childhood pneumonia and diarrhoea found that when mean temperature
180 Chapter 8: Discussion and conclusions
decreased for more than 2 °C from one day to the next, EDVs for childhood
pneumonia and diarrhoea increased significantly, suggesting that parents and
caretakers of children in Brisbane should be more aware of the temperature change in
two neighbouring days, especially in July. DTR over 17 °C was also associated with
an increase in EDVs for childhood diarrhoea, highlighting that more research and
targeted health policies and programs are needed to minimize the risk posed by great
DTRs.
The association between infectious diseases and weather conditions has long been
appreciated. Back to early 20th century, a study in Netherlands has demonstrated the
relationship between the increase in upper respiratory tract infections and outdoor
cold temperature (van Loghem1928). “Winter gastroenteritis” was a recognized
illness of early childhood before rotavirus was identified. In a review of 34 studies
conducted prior to 1990, Cook et al. have summarized that rotavirus infection
typically occur in winter, which puts forth the hypothesis that cold temperature may
increase rotavirus diseases (Cook et al. 1990). While subtle changes in local
temperature may play a role in explaining the seasonal cycling of childhood
pneumonia and diarrhoea in some settings, some other climatic factors, including
relative humidity (Moe and Shirley 1982), rainfall (Ansari et al. 1991) and sunshine
(Paynter et al. 2013), may also contribute to the occurrence of these infectious
diseases. Further, the transmission of these infectious diseases involves a lot of
factors, such as host behaviour and susceptibility, as well as spread and survivability
of pathogens etc., and climatic factors alone cannot fully explain its complexity. It is
still far from satisfactory to unfold the drivers of the transmission of childhood
infectious diseases.
181
Chapter 8: Discussion and conclusions 181
8.3 Strengths and Limitations
The impact of climate change on children’s health is scarcely researched so far. This
thesis has several strengths. First, it utilized the satellite remote sensing data to
quantify the effects of temperature on childhood pneumonia and diarrhoea, which
greatly minimized the measurement bias and therefore drew more accurate
conclusions. Second, it examined the balance between high and low temperature
effects on children’s health, which provided pivotal information on “how
temperature-related burden of childhood pneumonia and diarrhoea may change as
climate change proceeds”. Third, to the best of my knowledge, it examined the
effects of heat waves and cold spells on childhood pneumonia and diarrhoea for the
first time, and identified the relative importance of extremely high (low)
temperatures and sustained high (low) temperatures in the occurrence of childhood
pneumonia and diarrhoea. Finally, it assessed the impacts of temperature variation on
childhood pneumonia and diarrhoea, which suggests that the public health sector in
Brisbane take precautionary measures ahead of not only extreme temperatures and
but also big temperature variations.
This thesis also has several weaknesses. First, the ecological design which used the
aggregated data on temperature and EDVs for childhood pneumonia and diarrhoea in
Brisbane may result in some biases in exposure and/or outcome measures. The
ecological design also largely restricted us to examine the causation between
temperature and childhood pneumonia and diarrhoea. Second, due to the data
availability issue, we used the patients’ postcodes to do the spatial analysis (Chapter
3), which is not ideal compared with data by Statistical Local Areas (SLA) and Local
Government Areas (LGA). We are collecting data on EDVs for childhood diseases
182 Chapter 8: Discussion and conclusions
by SLA for implementing further studies. Third, there were some missing values in
the dataset, especially for the Indigenous status of the patients, which might to some
extent impact the conclusions we drew in the thesis. Forth, the time-series studies
(Chapters 4 to 7) only focused on one subtropical city, and thus it needs to be
cautious to generalize our findings to regions of other climates. Fifth, the range of
climatic variables we used is limited compared with a previous study looking at the
impacts of temperature, sun hours, rainfall and relative humidity on childhood
pneumonia (Paynter et al. 2013). Sixth, we did not have the pathogen data, which
restricted us to quantify the effects of temperature on specific pathogens.
8.4 Future research directions
Existing studies looking at the impacts of temperature on childhood pneumonia and
diarrhoea adopted various statistical approaches and yielded different types of
outputs, which makes it hard to quantitatively combine the findings together. Further,
majority of previous studies are one-city-only study, whereas children’s vulnerability
to the adverse impacts of extreme temperatures and large temperature variability
depends largely on their access to the climate change adaptation infrastructures of the
city they live, and thus it is necessary to conduct a large scale study using a
consistent statistical method to assess the heterogeneity in the relationship between
temperature and childhood pneumonia and diarrhoea across different cities. Even for
the children living in the same city but different suburbs, they may have different
sensitivities to temperature impact. Therefore, it is also strongly needed to evaluate
the with-in-city heterogeneity in children’s sensitivity to the adverse impacts of
183
Chapter 8: Discussion and conclusions 183
temperature on childhood pneumonia and diarrhoea, as well as to identify the drivers
behind, facilitating the future resource allocation to the priority subpopulations.
Different pathogens of childhood pneumonia and diarrhoea have different seasonal
variations, and the incubation period varies for viral and bacterial infections as well,
which requires future studies using lab-confirmed pathogen data to more accurately
quantify the effects of temperature on childhood pneumonia and diarrhoea caused by
different pathogens. Further, under reporting issue exists in surveillance data. Some
children with mild symptoms may do not want to seek medical care. Milinovich et al.
(Milinovich et al. 2014) highlighted using Internet resource to monitor emerging
infectious diseases, which can also be applied to the traditional children’s diseases
area. Using data collected from Internet-based surveillance system to quantify the
relationship between climate change and childhood pneumonia and diarrhoea may
more adequately capture the complexity of this relationship.
As climate change continues, the temperature-related burden of childhood
pneumonia and diarrhoea may vary accordingly (Walker et al. 2013). It is important
to project the burden of childhood pneumonia and diarrhoea under different climate
change scenarios, especially for those regions with a great risk for these diseases
(Huang et al. 2011).
8.5 Conclusions
In conclusion, this thesis contributes to the scientific knowledge in three ways: I) it
identified the risk areas of EDVs for childhood pneumonia and diarrhoea in
Queensland, Australia; II) it quantified the impacts of extreme temperatures (ie., heat
and cold) and prolonged extreme temperatures (ie., heat waves and cold spells) on
184 Chapter 8: Discussion and conclusions
EDVs for childhood pneumonia and diarrhoea; III) it examined the effects of
temperature fluctuation on EDVs for childhood pneumonia and diarrhoea.
Australia shoulders a big burden of childhood pneumonia and diarrhoea. Mount Isa
was observed as the high risk area where EDVs for both childhood pneumonia and
diarrhoea, even though EDVs for childhood pneumonia and diarrhoea experienced a
big decrease from 2007 to 2011. In light of the fact that children in Mount Isa are
exposed to enormous environmental hazards (polluted air and contaminated water,
etc.) due to mining activity, precautionary initiatives should be taken to protect
children in this area from illness.
Climate change will increase global average temperature, and cause more frequent,
intense, and longer-lasting heat waves, which may result in increases in the EDVs for
both childhood pneumonia and diarrhoea in Brisbane. Although, as climate change
continues, childhood pneumonia and diarrhoea cases due to mild high and low
temperatures may not increase substantially, burden of childhood pneumonia and
diarrhoea associated with prolonged extreme temperatures may increase in the future.
Other than absolute temperature, temperature variability, especially a temperature
decrease over 2 °C from one day to the next which most frequently occurred in July
every year, may increase EDVs for childhood pneumonia and diarrhoea.
The fundamental motivation of this thesis is to assist with the development of climate
change adaptation strategies aiming to protect children in Queensland. In particular,
the main and added effects of heat on childhood pneumonia and diarrhoea in the past
decade, the change over time in the impact of heat waves on childhood pneumonia
and diarrhoea, the temperature variation effects on childhood pneumonia and
diarrhoea, and the high vulnerability of Indigenous children to all these effects,
185
Chapter 8: Discussion and conclusions 185
should be taken into account while stakeholders are designing and implementing
future climate change adaptation plans for Queensland.
8.6 References
Australian Institute of Health and Welfare. 2012. Australia’s Health 2012. Canberra.
Ansari SA, Springthorpe VS, Sattar SA. 1991. Survival and vehicular spread of
human rotaviruses: possible relation to seasonality of outbreaks. Rev Infect
Dis 13(3):448-461.
Barnett AG, Williams GM, Schwartz J, Neller AH, Best TL, Petroeschevsky AL, et
al. 2005. Air pollution and child respiratory health. Am J Respir Crit Care
Med 171(11):1272-1278.
Cook SM, Glass RI, Lebaron CW, Ho MS. Global seasonality of rotavirus infections.
Bull World Health Organ 68(2):171-177.
Fritze J, Blashki G, Burke S, Wiseman J. 2008. Hope, despair and transformation:
Climate change and the promotion of mental health and wellbeing. Int J Ment
Health Syst 2(1):13.
National Mental Health Commission. Australian Government. 2012. A contributing
life: the 2012 national report card on mental health and suicide prevention.
Green R, Basu R, Malig B, Broadwin R, Kim J, Ostro B. 2010. The effect of
temperature on hospital admissions in nine California counties. Int J Public
Health 55(2):113-121.
Harley D, Bi P, Hall G, Swaminathan A, Tong S, Williams C. 2011. Climate change
and infectious diseases in Australia: Future prospects, adaptation options, and
research priorities. Asia Pac J Public Health 23(2s):54S-66S.
186 Chapter 8: Discussion and conclusions
Huang C, Barnett A, Wang X, Vaneckova P, FitzGerald G, Tong S. 2011. Projecting
future heat-related mortality under climate change scenarios: a systematic
review. Environ Health Perspect 119(12):1681-1690.
Janu EK, Annabattula BI, Kumariah S, Zajaczkowska M, Whitehall JS, Edwards MJ,
et al. Paediatric hospitalisations for lower respiratory tract infections in
Mount Isa. Med J Aust 200(10):591-594.
van Loghem JJ. 1928. An epidemiological contribution to the knowledge of the
respiratory diseases. J Hyg (Lond) 28(1):33-54.
Lanphear B, Hornung R, Khoury J, Yolton K, Baghurst P, Bellinger D, et al. 2005.
Low-level environmental lead exposure and children's intellectual function:
an international pooled analysis. Environ Health Perspect 113(7):894-899.
Lee S, Serre M, Donkelaar DV, Martin R, Burnett R, Jerrett M. 2012. Comparison of
geostatistical interpolation and remote sensing techniques for estimating
long-term exposure to ambient PM2.5 concentrations across the continental
United States. Environ Health Perspect 120(12):1727-1732.
Lin H, Zhang Y, Xu Y, Xu X, Liu T, Luo Y, et al. 2013. Temperature changes
between neighboring days and mortality in summer: A distributed lag non-
linear time series analysis. PLoS ONE 8(6):e66403.
Mcbain-Rigg KE, Veitch C. Cultural barriers to health care for Aboriginal and
Torres Strait Islanders in Mount Isa. Aust J Rural Health 19(2): 70-74.
Meehl GA, Tebaldi C. 2004. More intense, more frequent, and longer lasting heat
waves in the 21st century. Science 305(5686):994-997.
Milinovich GJ, Williams GM, Clements ACA, Hu W. 2014. Internet-based
surveillance systems for monitoring emerging infectious diseases. Lancet
Infect Dis 14(2):160-168.
187
Chapter 8: Discussion and conclusions 187
Moe K, Shirley J. 1982. The effects of relative humidity and temperature on the
survival of human rotavirus in faeces. Arch Virol 72(3):179-186.
Munksgaard NC, Taylor MP, Mackay A. 2010. Recognising and responding to the
obvious: the source of lead pollution at Mount Isa and the likely health
impacts. Med J Aust 193(3):131-132.
National Pollution Inventory (2010).
http://www.npi.gov.au/npidata/action/load/individual-facility-
detail/criteria/year/2009/browse-type/Company/regbusiness-
name/MOUNT%2BISA%2BMINES%2BLTD/jurisdiction-
facility/Q020MIM001. Retrieved October 1st 2013
Ostro B, Rauch S, Green R, Malig B, Basu R. 2010. The effects of temperature and
use of air conditioning on hospitalizations. Am J Epidemiol 172(9):1053-
1061.
Pascal M, Viso AC, Medina S, Delmas MC, Beaudeau P. 2012. How can a climate
change perspective be integrated into public health surveillance? Public
Health 126(8):660-667.
Paynter S, Weinstein P, Ware RS, Lucero MG, Tallo V, Nohynek H, et al. 2013.
Sunshine, rainfall, humidity and child pneumonia in tropics: time-series
analyses. Epidemiol Infect 141(6):1328-1336.
Schar C, Vidale PL, Luthi D, Frei C, Haberli C, Liniger MA, Appenzeller C. 2004.
The role of increasing temperature variability in European summer
heatwaves. Nature 427(6972):332-336.
Taylor MP, Mackay A, Kuypers T, Hudson-Edwards K. 2009. Mining and urban
impacts on semi-arid freshwater aquatic systems: the example of Mount Isa,
Queensland. J Environ Monit 11(5):977-986.
188 Chapter 8: Discussion and conclusions
Taylor MP, Schniering CA. 2010. The public minimization of the risks associated
with environmental lead exposure and elevated blood lead levels in children,
Mount Isa, Queensland, Australia. Arch Environ Occup Health 65(1):45-48.
Taylor MP, Schniering CA, Lanphear BP, Jones AL. 2011. Lessons learned on lead
poisoning in children: One-hundred years on from Turner's declaration. J
Paediatr Child Health 47(12):849-856.
Walker CLF, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. 2013. Global
burden of childhood pneumonia and diarrhoea. Lancet 381(9875):1405-1416.
Xu Z, Etzel RA, Su H, Huang C, Guo Y, Tong S. 2012. Impact of ambient
temperature on children's health: A systematic review. Environ Res 117:120-
131.
Xu Z, Hu W, Wang X, Huang C, Tong S. 2014a. The impact of temperature
variability on years of life lost. Epidemiology 25(2):313-314.
Xu Z, Huang C, Su H, Turner L, Qiao Z, Tong S. 2013. Diurnal temperature range
and childhood asthma: a time-series study. Environmental Health 12(1):12.
Xu Z, Sheffield P, Su H, Wang X, Bi Y, Tong S. 2014b. The impact of heat waves
on children’s health: a systematic review. Int J Biometeorol 58(2):239-247.
Xu Z, Sheffield PE, Hu W, Su H, Yu W, Qi X, et al. 2012b. Climate change and
children’s health—A call for research on what works to protect children. Int J
Environ Res Public Health 9(9):3298-3316.
Appendices 189
Appendices
Conference:
“2013 Conference of International Society for Environmental Epidemiology”, 20-23
August 2013, Basel, Switzerland (Poster presentation: Extreme temperatures and
paediatric emergency department admissions)
Journal Articles:
1. Xu Z, Hu W, Zhang Y, Wang X, Zhou M, Su H, Huang C, Tong S,
Guo Q (2015). Exploration of diarrhoea seasonality and its drivers in China.
Scientific Reports, 5:8241.
2. Xu Z, Hu W, Tong S (2015). The geographic co-distribution and
socio-ecological drivers of childhood pneumonia and diarrhea in Queensland,
Australia. Epidemiology and Infection, 143(5):1096-104.
3. Cheng J, Wu J, Xu Z, Zhu R, Wang X, Li K, Wen L, Yang H, Su H
(2014). Associations between extreme precipitation and childhood hand, foot
and mouth disease in urban and rural areas in Hefei, China. Science of the
Total Environment, 497-498: 484-90.
4. Cheng J, Zhu R, Xu Z, Wang X, Li K, Su H. Temperature variation
between neighboring days and mortality: a distributed lag non-linear analysis
(2014). International Journal of Public Health, 59(6):923-31.
5. Xu Z, Hu W, Zhang Y, Wang X, Tong S, Zhou M (2014).
Spatiotemporal pattern of bacillary dysentery in China from 1990 to 2009:
What is the driver behind? PLoS ONE, 9(8):e104329.
190 Appendices
6. Xu Z, Zhang W, Hu W, Tong S (2014). Seasonal amplitude of
hemorrhagic fever with renal syndrome in China: a call for attention to
neglected regions. Clinical Infectious Diseases, 59(7):1040-2.
7. Xu Z, Hu W, Tong S (2014). Temperature variability and childhood
pneumonia: an ecological study. Environmental Health, 13(1):51.
8. Xu Z, Liu Y, Ma Z, Toloo GS, Hu W, Tong S (2014). Assessment of
the temperature effect on childhood diarrhoea using satellite imagery.
Scientific Reports, 4:5389.
9. Xu Z, Liu Y, Ma Z, Li S, Hu W, Tong S (2014). Impact of
temperature on childhood pneumonia estimated from satellite remote sensing.
Environmental Research, 132:334-341.
10. Cheng J, Xu Z, Zhu R, Wang X, Jin L, Song J, Su H (2014). Impact
of diurnal temperature range on human health: a systematic review.
International Journal of Biometeorology, doi: 10.1007/s00484-014-0797-5
11. Xu Z, Hu W, Wang X, Huang C, Tong S. (2014). The impact of
temperature variability on years of life lost. Epidemiology, 25(2): 313-314.
12. Xu Z, Sheffield PE, Su H, Wang X, Bi Y, Tong S. (2014). The impact
of heat waves on children’s health: a systematic review. International Journal
of Biometeorology, 58(2):239-247.
13. Xu Z, Hu W, Su H, Turner LR, Ye X, Wang J, Tong S. (2013).
Extreme temperatures and paediatric emergency department admissions.
Journal of Epidemiology and Community Health, 68(4):304-311.
191
Appendices 191
14. Xu Z, Hu W, Williams G, Clements AG, Kan H, Tong S. (2013). Air
pollution, temperature and pediatric influenza in Brisbane, Australia.
Environmental International, 59:384-388.
15. Xu Z, Huang C, Hu W, Turner LR, Su H, Tong S. (2013). Extreme
temperatures and emergency department admissions for childhood asthma in
Brisbane, Australia. Occupational and Environmental Medicine, 70(10): 730-
735.
16. Xu Z, Huang C, Hu W, Turner LR, Su H, Qiao Z, Tong S. (2013). Is
diurnal temperature range a risk factor for childhood diarrhoea? PLoS ONE
8(5):e64713.
17. Huang C, Barnett AG, Xu Z, Chu C, Wang X, Turner LR, Tong S.
(2013). Managing the health effects of temperature in response to climate
change: challenges ahead. Environmental Health Perspectives, 121(4): 415-
419.
18. Xu Z, Huang C, Hu W, Turner LR, Su H, Qiao Z, Tong S. (2013).
Diurnal temperature range and childhood asthma: a time-series study.
Environmental Health, 12:12.
19. Xu Z, Sheffield PE, Hu W, Su H, Yu W, Qi X, Tong S. (2012).
Climate change and children’s health: a call for research on what works to
protect children. International Journal of Environmental Research and Public
Health, 9(9): 3298-3316.
20. Xu Z, Etzel RA, Su H, Huang C, Guo Y, Tong S. (2012). Impact of
temperature on children’s health: a systematic review. Environmental
Research, 117: 120-131.