Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
The burden of unintentional drowning: global,
regional and national estimates of mortality from the
Global Burden of Disease 2017 Study
►►Additional material is
published online only. To view
Richard Charles Franklin ‍ ‍,1,2 Amy E Peden ‍ ‍,2,3 Erin B Hamilton,4
please visit the journal online Catherine Bisignano,4 Chris D Castle,4 Zachary V Dingels,4 Simon I Hay,4,5 Zichen Liu,4
(http://​dx.​doi.o​ rg/​10.​1136/​
injuryprev-​2019-​043484). Ali H Mokdad,4,5 Nicholas L S Roberts,4 Dillon O Sylte,4 Theo Vos,4,5
For numbered affiliations see Gdiom Gebreheat Abady,6 Akine Eshete Abosetugn,7 Rushdia Ahmed,8,9
end of article. Fares Alahdab,10 Catalina Liliana Andrei,11 Carl Abelardo T Antonio,12,13 Jalal Arabloo,14
Correspondence to Aseb Arba Kinfe Arba,15 Ashish D Badiye,16 Shankar M Bakkannavar,17
Dr Spencer L James, Institute for
Health Metrics and Evaluation,
Maciej Banach,18,19 Palash Chandra Banik,20 Amrit Banstola,21
University of Washington, Suzanne Lyn Barker-­Collo,22 Akbar Barzegar,23 Mohsen Bayati,24 Pankaj Bhardwaj,25,26
Seattle, WA 98121, USA; ​
spencj@​uw.​edu Soumyadeep Bhaumik,27 Zulfiqar A Bhutta,28,29 Ali Bijani,30 Archith Boloor,31
Received 19 September 2019
Félix Carvalho,32 Mohiuddin Ahsanul Kabir Chowdhury,33,34 Dinh-­Toi Chu,35
Revised 4 December 2019 Samantha M Colquhoun,36 Henok Dagne,37 Baye Dagnew,38 Lalit Dandona,4,5,39
Accepted 6 December 2019
Rakhi Dandona,4,39 Ahmad Daryani,40 Samath Dhamminda Dharmaratne,4,41
Zahra Sadat Dibaji Forooshani,42 Hoa Thi Do,43 Tim Robert Driscoll,44
Arielle Wilder Eagan,45,46 Ziad El-­Khatib,47,48 Eduarda Fernandes,49 Irina Filip,50,51
Florian Fischer,52 Berhe Gebremichael,53 Gaurav Gupta,54 Juanita A Haagsma,55
Shoaib Hassan,56 Delia Hendrie,57 Chi Linh Hoang,58 Michael K Hole,59
Ramesh Holla,60 Sorin Hostiuc,61,62 Mowafa Househ,63,64 Olayinka Stephen Ilesanmi,65
Leeberk Raja Inbaraj,66 Seyed Sina Naghibi Irvani,67 M Mofizul Islam,68
Rebecca Q Ivers,69 Achala Upendra Jayatilleke,70,71 Farahnaz Joukar,72
Rohollah Kalhor,73,74 Tanuj Kanchan,75 Neeti Kapoor,16 Amir Kasaeian,76,77
Maseer Khan,78 Ejaz Ahmad Khan,79 Jagdish Khubchandani,80 Kewal Krishan,81
G Anil Kumar,39 Paolo Lauriola,82 Alan D Lopez,4,83 Mohammed Madadin,84
Marek Majdan,85 Venkatesh Maled,86,87 Navid Manafi,88,89 Ali Manafi,90
Martin McKee,91 Hagazi Gebre Meles,92 Ritesh G Menezes,93 Tuomo J Meretoja,94,95
Ted R Miller,57,96 Prasanna Mithra,97 Abdollah Mohammadian-­Hafshejani,98
Reza Mohammadpourhodki,99 Farnam Mohebi,100,101 Mariam Molokhia,102
Ghulam Mustafa,103,104 Ionut Negoi,105,106 Cuong Tat Nguyen,107
Huong Lan Thi Nguyen,107 Andrew T Olagunju,108,109 Tinuke O Olagunju,110
Jagadish Rao Padubidri,111 Keyvan Pakshir,112 Ashish Pathak,47,113 Suzanne Polinder,55
Dimas Ria Angga Pribadi,114 Navid Rabiee,115 Amir Radfar,116,117
Saleem Muhammad Rana,118,119 Jennifer Rickard,120,121 Saeed Safari,122
Payman Salamati,123 Abdallah M Samy,124 Abdur Razzaque Sarker,125
David C Schwebel,126 Subramanian Senthilkumaran,127 Faramarz Shaahmadi,128
Masood Ali Shaikh,129 Jae Il Shin,130,131 Pankaj Kumar Singh,132 Amin Soheili,133,134
Mark A Stokes,135 Hafiz Ansar Rasul Suleria,136 Ingan Ukur Tarigan,137
© Author(s) (or their
employer(s)) 2020. Re-­use Mohamad-­Hani Temsah,138,139 Berhe Etsay Tesfay,140 Pascual R Valdez,141,142
permitted under CC BY.
Published by BMJ.
Yousef Veisani,143 Pengpeng Ye,144 Naohiro Yonemoto,145 Chuanhua Yu,146,147
To cite: Franklin RC,
Hasan Yusefzadeh,148 Sojib Bin Zaman,33,149 Zhi-­Jiang Zhang,150 Spencer L James ‍ ‍4
Peden AE, Hamilton EB, et al.
Inj Prev Epub ahead of print:
[please include Day Month
Year]. doi:10.1136/
injuryprev-2019-043484

Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 1

 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
Abstract at global and local levels to address the circumstances leading to
Background Drowning is a leading cause of injury-­related mortality drowning to guide preventive efforts.
globally. Unintentional drowning (International Classification of When considering YLLs, within the unintentional injuries
Diseases (ICD) 10 codes W65-74 and ICD9 E910) is one of the 30 category, drowning has the second highest YLLs after road inju-
mutually exclusive and collectively exhaustive causes of injury-­related ries, demonstrating the significant impact that drowning has,
mortality in the Global Burden of Disease (GBD) study. This study’s particularly on children.9 Conversely, it has among the lowest
objective is to describe unintentional drowning using GBD estimates YLDs due to the high lethality of the injury.10 Recent data from
from 1990 to 2017. the USA shows that for each drowning death there are 2.5 hospi-
Methods Unintentional drowning from GBD 2017 was estimated for talisations (ratio 1:2.5), whereas for road injuries this ratio is
cause-­specific mortality and years of life lost (YLLs), age, sex, country, 1:88 and for falls even higher, 1:229.11 Thus, reducing drowning
region, Socio-­demographic Index (SDI) quintile, and trends from 1990 deaths will reduce the overall mortality burden but is likely to
to 2017. GBD 2017 used standard GBD methods for estimating have limited impact on morbidity of unintentional injuries.
mortality from drowning. There is limited information on drowning in LMICs. A recent
Results Globally, unintentional drowning mortality decreased by review of published literature on drowning in LMICs between
44.5% between 1990 and 2017, from 531 956 (uncertainty interval 1984 and 2015 identified 62 studies, with the majority of these
(UI): 484 107 to 572 854) to 295 210 (284 493 to 306 187) deaths. from Asia (56%), followed by Africa (26%).5 Further work is
Global age-­standardised mortality rates decreased 57.4%, from 9.3 required to understand the drowning burden in countries in
(8.5 to 10.0) in 1990 to 4.0 (3.8 to 4.1) per 100 000 per annum in Latin America, Africa, Asia, and the Pacific.
2017. Unintentional drowning-­associated mortality was generally There has been a reduction, over the last decade, in the esti-
higher in children, males and in low-­SDI to middle-­SDI countries. China, mated annual rate of drowning globally, with the latest GBD
India, Pakistan and Bangladesh accounted for 51.2% of all drowning study showing a reduction of 17.2% from 2007 to 2017.4 There
deaths in 2017. Oceania was the region with the highest rate of age-­ is no clear consensus as to why this reduction has occurred.
standardised YLLs in 2017, with 45 434 (40 850 to 50 539) YLLs per For LMICs, this is likely related to changes in development
100 000 across both sexes. and urbanisation,12 with the greatest reductions occurring in
Conclusions There has been a decline in global drowning rates. East Asia and southern sub-­Saharan Africa.9 A recent review of
This study shows that the decline was not consistent across countries. drowning in South Africa found an increase in drowning across
The results reinforce the need for continued and improved policy, the country with variation based on geography,13 demonstrating
prevention and research efforts, with a focus on low- and middle-­ that the drowning burden is not evenly distributed and that
income countries. choosing data from a region within a country or a country within
a region can potentially skew the findings.
This current study focuses on unintentional drowning, as
defined by those International Classification of Diseases (ICD-
9/10) codes that are primary unintentional drowning codes
Introduction
(E910 and W65-74). This does not include all drowning deaths,
Drowning has been identified as a public health priority by
meaning intentional (X71), disaster (X38) and transport (V90,
the World Health Organization (WHO)1 and is defined as the
V92)-­related drowning are not included, with unspecified effects
process of experiencing respiratory impairment from submer-
of immersion (T75.1) and undetermined intent (Y21) redistrib-
sion/immersion in liquid, with drowning outcomes classified
uted across drowning and other categories. It has been estimated
as death, morbidity and no morbidity.2 Fatal drowning is a
that the unintentional drowning ICD codes account for approx-
leading cause of unintentional injury-­related mortality world-
imately 40%–50% of all drowning mortality in high-­ income
wide, with approximately 300 000 drowning deaths per annum
countries.14 15 Given some countries have a higher burden of
globally.3 4 Drowning disproportionately impacts those from boating-­related16 17 and disaster-­
related18 drowning mortality,
low- and middle-­ income countries (LMICs), males and chil- defining unintentional drowning only by ICD codes W65-74
dren.5 6 Having accurate and timely data to aid in the allocation impacts accurate estimates. Similarly, intentional drowning
of public health resources and the monitoring of interventions is deaths19 must be included in overall all-­ cause estimates of
important for continued implementation of drowning preven- mortality to further enhance accuracy of estimates.
tion programs.3 7 The goal of this study was to report estimates from GBD
The Global Burden of Diseases, Injuries, and Risk Factors 2017 for mortality from unintentional drowning by region
(GBD) Study is a comprehensive assessment of health losses asso- and compare these to 1990 estimates. While GBD 2017 also
ciated with risk factors, and from morbidity and mortality. GBD reported estimates for non-­fatal cases of drowning, the purpose
2017 produced estimates of all-­cause mortality, cause-­specific of this study was to focus exclusively on cause-­specific mortality
mortality, years of life lost (YLLs), incidence, prevalence, years from unintentional drowning.
lived with disability (YLDs) and disability-­adjusted life years for
292 different causes of mortality and 354 different causes of
non-­fatal health loss for 195 countries and territories across a Methods
28-­year time span from 1990 to 20174 8. Drowning is one of 30 The GBD covers all nation-­states using a range of data sources to
mutually exclusive causes of injury-­related mortality in the GBD provide a comprehensive picture of global health. An overview
2017 study design and is nested within the unintentional injury of methods used for GBD 2017 is provided in online supple-
category of the GBD 2017 cause hierarchy.9 mentary appendix 1 and in other GBD literature.4 8 20–23 All
Within injury, drowning has the third highest unintentional analytical code used for GBD 2017 is available online (http://
mortality rate after road injuries and falls. Across all injury types www.​ghdx.​healthdata.​org). This study complies with the Guide-
(both intentional and unintentional), drowning has the fourth lines for Accurate and Transparent Health Estimates Reporting
highest overall mortality rate after road injuries, falls and inter- recommendations (online supplementary appendix 2). Analyses
personal violence.9 As such, there is a need for immediate action were completed using Python V.2.7, Stata V.13.1, or R V.3.3.
2 Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
Statistical code used for GBD estimation is publicly available Table 1 shows the covariates that were utilised for measuring
online at h ​ ealthdata.​org. An overview of methods specific to drowning mortality in CODEm, as well as priors for the spec-
drowning mortality is as follows: ified direction of causality (positive meaning that cause-­specific
mortality is expected to increase with increases in the level of the
Cause of death data for drowning covariate) and the level of causality expected for the relation-
GBD 2017 utilised all available cause of death data for every ship between the covariate and mortality (‘1’ referring to more
location in the GBD location hierarchy, which includes 195 proximal associations, and ‘3’ referring to more distal associa-
countries and territories.4 The GBD 2017 cause of death data- tions). These factors are based on common factors described in
base included vital registration (VR), verbal autopsy, police the literature and expert opinion. Further work exploring these
record data and mortuary data, among other data types. These factors and the impact on drowning would add value.
data undergo extensive quality checks and data processing to Once CODEm models are fit for drowning, the location-­
ensure comparability between different coding systems from specific estimates are scaled to fit within each parent location
different sources that were collected at different times—for (eg, all subnational drowning deaths sum to national drowning
example, to allow for both ICD9-­coded and ICD10-­coded data deaths, and national drowning deaths sum to global drowning
to be used in a comparable manner. In addition, the GBD cause deaths). In addition, cause-­specific estimates are scaled to fit
hierarchy is mutually exclusive and collectively exhaustive, within the parent cause (eg, the sum of deaths from each injury
meaning that each death has one and only one underlying cause equals the deaths for the overall injury cause, and the sum of
of death assigned. Cause of death data also undergo a process deaths from each individual cause in the cause hierarchy equals
known as garbage code redistribution, whereby ill-­defined cause the sum of all-­cause mortality). For drowning, these processes
of death codes are redistributed to underlying causes of death in led to an increase in 0.82% in terms of drowning deaths across
the GBD 2017 cause hierarchy. Such processes are described in all ages for both sexes in 2017. Results are presented followed
more detail elsewhere.4 For GBD 2017, drowning deaths were by a 95% uncertainty interval (UI) range (online supplementary
identified with ICD9 code E910 and ICD10 codes W65-­W74.9. appendix 1).
In addition, drowning was a cause of death in verbal autopsy
survey instruments, which were a source of cause of death data Calculating YLLs for drowning
in many locations including India. YLLs are defined as the difference between life expectancy
and the age at which a death occurs, based on life tables used
Cause-specific mortality modelling for drowning in GBD 2017 that estimate the remaining life expectancy for
Since GBD estimates outcomes for every location, year, age and each 5-­year age group in all populations greater than 5 million
sex in the GBD 2017 study design, statistical modelling processes in GBD 2017.22 YLLs are an important measure of drowning
are implemented to predict estimates for where data are missing. mortality, since drowning more commonly occurs at early ages,
GBD 2017 utilised standard GBD methods for modelling cause-­ which leads to more YLLs than a death occurring at an older
specific mortality from drowning.4 In particular, the study imple- age. The age-­specific life expectancies used to calculate YLLs are
mented the Cause of Death Ensemble model (CODEm) approach, provided in table 2.
which is described in more detail elsewhere4 but is summarised
as follows in terms of five key principles. First, CODEm uses all Socio-demographic Index
available data, as described above. Second, data processing steps The Socio-­ demographic Index (SDI) is an index of socio-­
are conducted to ensure comparability and quality of the dataset, economic development which is used as a covariate in various
also as described above. Third, CODEm develops a diverse set of GBD processes as well as a metric to demonstrate burden trends
plausible models using different regression forms and different for locations at different levels of development. The index is
sets of covariates. Fourth, CODEm measures the predictive calculated based on lag-­ distributed income per capita, which
validity of each model and of an ensemble of all models. Finally, is a smoothed gross domestic product per capita series, mean
the best model is chosen based on out-­ of-­sample predictive educational attainment in years and total fertility rate under age
validity testing. The CODEm framework is intended to develop 25 years. In its formulation, SDI increases as income and educa-
the best possible model for each cause given the availability of tion increase and fertility decreases. The derivation of SDI is
existing data and relationships with covariates to help inform described in more detail elsewhere.8
model estimates where data are absent or sparse.

Results
Globally, in 1990, there were 531 956 (UI: 484 107 to 572 854)
Table 1 Covariates used in unintentional drowning CODEm models deaths from unintentional drowning, which decreased by 44.5%
Covariate Level Direction (UI: 38.9% to 48.3%) to 295 210 (UI: 284 493 to 306 187)
Alcohol (litres per capita) 1 + deaths from unintentional drowning in 2017. The global age-­
Coastal population within 10 km (proportion) 1 + standardised cause-­ specific mortality rate for unintentional
Landlocked nation (binary) 1 –
drowning was 9.3 (UI: 8.5 to 10.0) per 100 000 in 1990, which
decreased by 57.4% (UI: 53.3% to 60.1%) to a rate of 4.0 (UI:
Log-­transformed summary exposure value for drowning 1 +
3.8 to 4.1) per 100 000 in 2017 (online supplementary table S1).
Rainfall, lowest quintile 1 –
Online supplementary table S1 provides all-­ age drowning-­
Rainfall, highest quintile 1 +
specific deaths and age-­ specific mortality
standardised cause-­
Elevation under 100 m (proportion 2 +
rates for 1990 and 2017, as well as the percentage change
Education (years per capita) 3 –
between 1990 and 2017 for every country, region and super-­
Lag-­distributed income per capita 3 No prior region in the GBD 2017 location hierarchy. Overall, there was a
Socio-­demographic Index 3 – decrease in drowning rates (except in Tonga) and numbers, with
CODEm, Cause of Death Ensemble model. the notable exception of Oceania, which had a near doubling
Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 3
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
15–49, 50–69 and 70 plus) are shown in figure 6, with Asia
Table 2 Age-­specific life expectancy used for years of life lost
(South, East and Southeast) having the highest YLLs. In 1990,
calculation
drowning caused 37 925 314 (UI: 33 968 621 to 41 237 399)
Age Life expectancy (years) total YLLs, a number which decreased by 56.3% (UI: 51.2% to
0 87.9 59.7%) to a total of 16 563 278 (UI: 15 784 185 to 17 349 952)
1 87.0 YLLs in 2017. The age-­standardised YLL rate in 1990 was 632.3
5 83.0 (UI: 568.8 to 685.6) per 100 000, which decreased by 63.9%
10 78.1 (UI: 59.8% to 66.6%) to 228.3 (UI: 217.2 to 239.7) YLLs per
15 73.1 100 000 in 2017. South Asia was the region with the greatest
20 68.1 number of YLLs in 2017, with 5 273 794 (UI: 4 819 285 to 5
25 63.2 744 698) YLLs across both sexes and all ages. Oceania, however,
30 58.2
was the region with the highest rate of age-­standardised YLLs in
2017, with 828.2 (UI: 672.7 to 997.9) YLLs per 100 000 across
35 53.3
both sexes (online supplementary table S2).
40 48.4
45 43.5
50 38.7 Socio-demographic Index
55 34.0 The trends over time in terms of age-­standardised cause-­
60 29.3 specific mortality rates from drowning for each super-­region
65 24.7 in GBD between 1990 and 2017 are shown in figure 7, with
70 20.3
decreases seen across all super-­ r egions. The numbers of
75 16.1
deaths and age-­standardised cause-­specific mortality rates
were noted to decrease in every SDI quintile and for most
80 12.2
countries. In 27 countries, mainly in the Oceania region,
85 8.8
there was no significant percentage change between 1990
90 6.1
and 2017 for age-­s tandardised cause-­s pecific mortality rates.
95 3.9
Countries outside of the Oceania region that did not expe-
100 2.2 rience decreases in age-­s tandardised cause-­s pecific mortality
105 1.6 rates were Lesotho, Zimbabwe and Cape Verde. Selected
110 1.4 countries with large populations were noted to still have
large numbers of deaths from unintentional drowning in
2017, specifically China, India, Bangladesh and Pakistan,
(80.1%) of numbers, led by a large increase in Papua New accounting for just over half of all drowning deaths (51.2%)
Guinea (93.4%). Online supplementary table S2 provides the (online supplementary table S1).
same information for YLLs.
Discussion
Age-standardised cause-specific mortality There were approximately 295 210 (UI: 284 493 to 3 06 187)
The age-­standardised cause-­specific mortality rates for unin- unintentional drowning deaths in 2017, an almost 50% reduc-
tentional drowning per 100 000 for each country in 2017 by tion (44.5%) in unintentional drowning deaths over the last
sex are shown in figure 1, with male rates generally higher 28 years. This study explored changes between 1990 and 2017
than female rates. Figure 2 shows all-­a ge death counts for in unintentional drowning mortality across the globe, and while
unintentional drowning for each country in 2017 by sex, there has been an expected reduction in unintentional drowning
with China, India, Pakistan and Bangladesh all having mortality, as has been seen in other studies,7 24 the reduction
over 10 000 deaths in 2017. Figure 3 shows the number of was not uniform. Countries from middle SDI groupings had
deaths by age, sex and super-­region, with children 1–4 years the greatest reduction (54.0%), again indicating that urban-
having the highest number of deaths. For females, there is isation and development are possible drivers of the decrease
an increase in the number of deaths in the 60–84 years age in drowning deaths; however, other drivers, such as greater
groups. investment in water safety, government recognition of the issue,
The distributions of deaths by region for five age groups changing social norms or coding frameworks, could be contrib-
(under 5, 5–14, 15–49, 50–69 and 70 plus) are shown in uting.1 2 14
figure 4. These figures emphasise how there is greater unin- Four countries in 2017 (China, India, Pakistan and Bangladesh)
tentional drowning mortality in males than females in most accounted for half of all drowning deaths, with most countries
age groups, with sex differences less pronounced in certain experiencing a decrease in drowning rates and numbers between
regions such as South Asia (online supplementary table S1). 1990 and 2017. All four countries in 2017 had rates higher than
Distributions of drowning deaths by age vary depending on the global average of 4.0 (3.6–4.1) and while Bangladesh had
region. For example, there are relatively more deaths from experienced the largest reduction in the drowning death rate (a
unintentional drowning in younger age groups in lower-­ reduction of 70.4%) it had a rate of 9.5 (7.6–12.1). The excep-
income regions, compared with other age groups in those tion was Oceania, which saw a near doubling (80.1% increase)
regions, such as the sub-­S aharan Africa, South Asia, North of drowning fatalities; however, it is unclear what is the cause
Africa, Middle East and Central Asia regions. of this increase. Such trends represent an ongoing challenge for
those working in drowning prevention in the Oceania region.
Years of life lost Children, residents of countries in Asia and Africa, males, and
YLLs by super-­region, sex and age group are shown in figure 5 low-­SDI and middle-­SDI countries account for the majority of
and YLLs by sex, region and five age groups (under 5, 5–14, drowning deaths. Unintentional drowning is preventable and is
4 Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.

Figure 1 Age-­standardised cause-­specific mortality rates per 100 000 by country for unintentional drowning in 2017.

Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 5

 Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.

Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484

Figure 2 All-­age mortality by country for unintentional drowning in 2017.
Original research

6
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
Figure 3 Incidence of unintentional drowning mortality by age group, sex and super-­region in 2017.

linked to exposure.6 25–27 Increasing economic prosperity and cognition and lack of swimming skills, remain over-­represented
urbanisation has brought about safer domestic and work-­related in unintentional drowning statistics. This is highlighted by the
conditions, with increasing exposure around recreational use of high number of YLLs from drowning.
water.7 16 28–30 Children, with their natural curiosity, evolving

Figure 4 Incidence of unintentional drowning mortality by region, sex and five age groups in 2017.
Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 7
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
Figure 5 Years of life lost to unintentional drowning by age group, sex and super-­region in 2017.

Child drowning prevention is a well-­ established area of jackets.36 However, it is noted that there is a need to take inter-
research,18 31 with multiple studies identifying strategies which ventions to scale18 and that strategies will need to take into
can effectively prevent unintentional drowning in high-­income account factors such as age, gender and access to safe aquatic
countries and LMICs, including restricting access to water,32 locations.37 Cardiopulmonary resuscitation (CPR) is also a poten-
supervision,33 learning to swim,34 crèches35 and the use of life tially lifesaving procedure capable of preventing fatal drowning,

Figure 6 Years of life lost to unintentional drowning by region, sex and five age groups, 2017.
8 Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
Figure 7 Age-­standardised cause-­specific mortality rate per 100 000 for unintentional drowning by year and super-­region (1990–2017).

rather than preventing the initial drowning incident, and needs can have a marked impact on mortality and morbidity rates. A
to be taught to supervisors of people in and around water, recent review5 of drowning in LMICs found only a small number
including parents.38 It also appears that while each strategy (ie, of studies from LMICs (n=62 studies) which met their inclu-
restricting access to water, supervision, learning to swim, use of sion criteria, noting that data used in a third of the included
lifejackets and CPR) is helpful, the best results are seen when all studies were from autopsy and medico-­legal records, with the
measures, combined, are enacted.33 Although a range of other rest from surveys and interviews, hospital records, media and
strategies have been proposed (such as alarms, cameras and child ambulance records. This highlights the paucity of published data
behaviour modification), there is not as strong an evidence base on drowning from LMICs, compounded by a lack of medico-­
for these interventions to date. legal, uniformly collected, comprehensive data. Strengthening
WHO, in recognising unintentional drowning as a public drowning data (monitoring) in LMICs should be prioritised,
health threat, has recently developed an implementation guide to leading to an increase in published studies, thereby enhancing
help countries address the challenge of unintentional drowning.3 the evidence base of drowning epidemiology and risk factors
This guide is designed to ensure that each country understands to better inform prevention efforts. The limited peer-­reviewed
the factors contributing to unintentional drowning and has a set studies from Africa also highlight the need for a greater focus
of strategies that can support unintentional drowning preven- on drowning in Africa, especially considering that 11 of the 16
tion. It is vital that all countries have quality data (comprehen- articles were from South Africa.5 Another review of drowning
sive, up-­to-­date, accurate and useable for planning) at national in South Africa showed large heterogeneity in drowning across
and subnational levels to inform a context-­specific and evidence-­ the country, a potential challenge for estimating the number of
based understanding of the burden of unintentional drowning. drowning deaths in Africa.13
Presenting the mortality rate of unintentional drowning across Future studies should focus analysis in regions and countries
the globe can help countries benchmark their unintentional with high rates of drowning mortality, particularly LMICs, as
drowning rate against that of similar countries (by size, loca- well as explore seasonal variations in drowning, the impact of
tion, SDI, etc) as well as provide an opportunity to discuss what alcohol, and cultural reasons for differing risks of drowning.
might be reasonable unintentional drowning reduction targets The impact of a country’s development index on drowning
to aim for in the development of a national water safety plan, as mortality rates, in addition to SDI, could be examined in
recommended by WHO.3 This information can be used to help future studies.
evaluate the success at a country level of a comprehensive unin-
tentional drowning prevention strategy.
While reducing drowning in low-­SDI and middle-­SDI coun- Limitations
tries is a challenge, recent work in Bangladesh34 and Thai- This study used data from the GBD project, combining a range
land29 39 has demonstrated that country-­level commitment and of data sources to provide the most comprehensive picture of
appropriate resourcing to unintentional drowning reduction unintentional drowning mortality across the globe. However,
Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 9
 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
it is not without its limitations. The use of mutually exclusive countries achieving greater reductions than others. Future
ICD codes means that not all drowning deaths are captured in GBD studies should evaluate actions at a country level to
the original data and recent work has shown that drowning inform future prevention efforts. Children, especially those
may not be the primarily coded cause of death, and while GBD from LMICs had higher fatal unintentional drowning rates,
redistribution methods should account for miscoding, more and achieving reductions in drowning in China, India, Bangla-
research and validation is required to continue improving data desh and Pakistan should be prioritised due to their population
quality.14 15 This study included unintentional drowning deaths size. There is a need for investment in drowning prevention
only, and as such does not include intentional drowning, strategies, policy and research, including the evaluation
drowning following aquatic transport or other transportation of preventive strategies to continue the downward trend in
incidents, or drowning due to natural disasters. Additionally, drowning rates seen in this study.
in countries, especially LMICs, where there is not a regular
collection of death data or which lack hospital data, there is Author affiliations
1
potential for misestimation of drowning deaths, which is an College of Public Health, Medical and Veterinary Science, James Cook University,
Douglas, Queensland, Australia
ongoing area of GBD research.40 2
Royal Life Saving Society, Sydney, New South Wales, Australia
This study captures greater uncertainty in data-­sparse areas, 3
School of Public Health and Community Medicine, Faculty of Medicine, University of
and this should be viewed alongside the point in time esti- New South Wales, Sydney, New South Wales, Australia
4
mates. As drowning is not evenly distributed across countries, Institute for Health Metrics and Evaluation, University of Washington, Seattle,
this study used models associated with drowning, such as coun- Washington, USA
5
Department of Health Metrics Sciences, School of Medicine, University of
tries with greater populations residing within 10 km of the
Washington, Seattle, Washington, USA
coast, with high rainfall, and those close to sea level (table 1). 6
College of Medicine and Health Sciences, Department of Nursing, Adigrat University,
These assumptions and covariates will benefit from ongoing Adigrat, Ethiopia
7
review and refinement. For example, some landlocked coun- 8
Department of Public Health, Debre Berhan University, Debre Berhan, Ethiopia
tries may have more exposure to drowning in inland water- James P Grant School of Public Health, BRAC University, Dhaka, Bangladesh
9
Health Systems and Population Studies Division, International Centre for Diarrhoeal
ways such as rivers and lakes.41 This study did not explore the Disease Research, Dhaka, Bangladesh
circumstances leading to drowning, aquatic location where the 10
Evidence Based Practice Center, Mayo Clinic Foundation for Medical Education and
incident occurred, or the impact of alcohol, which is a signif- Research, Rochester, Minnesota, USA
11
icant contributor to drowning.42–44 There is also further work 12
Carol Davila University of Medicine and Pharmacy, Bucharest, Romania
required to quantify the burden of drowning among migrants. Department of Health Policy and Administration, University of the Philippines
Manila, Manila, Philippines
Future drowning studies should also encompass all drowning 13
Department of Applied Social Sciences, Hong Kong Polytechnic University, Hong
deaths (ie, transport related, disaster related, unintentional, Kong, China
14
intentional and undetermined intent). Health Management and Economics Research Center, Iran University of Medical
Sciences, Tehran, Iran
15
Nursing Department, Wolaita Sodo University, Wolaita Sodo, Ethiopia
Conclusion 16
Department of Forensic Science, Government Institute of Forensic Science, Nagpur,
Unintentional drowning deaths across the globe continue to India
17
decrease, with the 2017 GBD (data from all national states) Department of Forensic Medicine and Toxicology, Manipal Academy of Higher
estimates showing 295 210 drowning deaths. The reduction Education, Manipal, India
18
Department of Hypertension, Medical University of Lodz, Lodz, Poland
in unintentional drowning has not been consistent, with some 19
Polish Mothers’ Memorial Hospital Research Institute, Lodz, Poland
20
Department of Non-­Communicable Diseases, Bangladesh University of Health
Sciences (BUHS), Dhaka, Bangladesh
21
Department of Research, Public Health Perspective Nepal, Pokhara-­Lekhnath
What is already known on the subject Metropolitan City, Nepal
22
School of Psychology, University of Auckland, Auckland, New Zealand
23
►► Mortality codes for unintentional drowning only capture Occupational Health Department, Kermanshah University of Medical Sciences,
Kermanshah, Iran
40%–50% of drowning deaths in high-­income countries. 24
Health Human Resources Research Center, Shiraz University of Medical Sciences,
►► Limited information exists on drowning in low-­income and Shiraz, Iran
25
middle-­income countries, especially in Latin America and Department of Community Medicine and Family Medicine, All India Institute of
Africa. Medical Sciences, Jodhpur, India
26
Department of Community Medicine, Datta Meghe Institute of Medical Sciences,
►► Drowning is a public health challenge requiring further
Deemed University, Wardha, India
research and sustained investment in evidence-­based 27
The George Institute for Global Health, New Delhi, India
interventions. 28
Centre for Global Child Health, University of Toronto, Toronto, Ontario, Canada
29
Centre of Excellence in Women and Child Health, Aga Khan University, Karachi,
Pakistan
30
Social Determinants of Health Research Center, Babol University of Medical
What this study adds Sciences, Babol, Iran
31
Department of Internal Medicine, Manipal Academy of Higher Education,
Mangalore, India
►► Almost 300 000 people died from unintentional drowning in 32
Research Unit on Applied Molecular Biosciences (UCIBIO), University of Porto,
2017, exclusive of drowning in natural disasters and due to Porto, Portugal
33
transport incidents. Maternal and Child Health Division, International Centre for Diarrhoeal Disease
►► The global age-­standardised mortality rate from Research, Dhaka, Bangladesh
34
Department of Epidemiology and Biostatistics, University of South Carolina,
unintentional drowning declined by 57% from 1990 to 2017. Columbia, South Carolina, USA
►► Trends over time by geographical regions and Socio-­ 35
Faculty of Biology, Hanoi National University of Education, Hanoi, Vietnam
demographic Index, from 1990 to 2017. Half (51.2%) of 36
Research School of Population Health, Australian National University, Action,
unintentional drowning-­related deaths in 2017 occurred in Australian Capital Territory, Australia
37
China, India, Pakistan and Bangladesh. Environmental and Occupational Health and Safety, University of Gondar, Gondar,
Ethiopia

10 Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484

 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
38 87
Department of Human Physiology, University of Gondar, Gondar, Ethiopia Health University, Rajiv Gandhi University of Health Sciences, Bangalore, India
39 88
Public Health Foundation of India, Gurugram, India Ophthalmology Department, Iran University of Medical Sciences, Tehran, Iran
40 89
Toxoplasmosis Research Center, Mazandaran University of Medical Sciences, Sari, Ophthalmology Department, University of Manitoba, Winnipeg, Manitoba, Canada
90
Iran Plastic Surgery Department, Iran University of Medical Sciences, Tehran, Iran
41 91
Department of Community Medicine, University of Peradeniya, Peradeniya, Sri Department of Health Services Research and Policy, London School of Hygiene &
Lanka Tropical Medicine, London, UK
42 92
Tehran University of Medical Sciences, Tehran, Iran Mekelle University, Mekelle, Ethiopia
43 93
Center of Excellence in Public Health Nutrition, Nguyen Tat Thanh University, Ho Forensic Medicine Division, Imam Abdulrahman Bin Faisal University, Dammam,
Chi Minh City, Vietnam Saudi Arabia
44 94
Sydney School of Public Health, University of Sydney, Sydney, New South Wales, Breast Surgery Unit, Helsinki University Hospital, Helsinki, Finland
Australia 95
University of Helsinki, Helsinki, Finland
45
Department of Global Health and Social Medicine, Harvard University, Boston, 96
Pacific Institute for Research & Evaluation, Calverton, Maryland, USA
Massachusetts, USA 97
Community Medicine, Manipal Academy of Higher Education, Mangalore, India
46
Department of Social Services, Tufts Medical Center, Boston, Massachusetts, USA 98
Department of Epidemiology and Biostatistics, Shahrekord University of Medical
47
Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden Sciences, Shahrekord, Iran
48
World Health Programme, Université du Québec en Abitibi-­Témiscamingue, Rouyn-­ 99
Department of Nursing, Shahroud University of Medical Sciences, Shahroud, Iran
Noranda, Québec, Canada 100
Non-­Communicable Diseases Research Center, Tehran University of Medical
49
REQUIMTE/LAQV, University of Porto, Porto, Portugal Sciences, Tehran, Iran
50
Psychiatry Department, Kaiser Permanente, Fontana, California, USA 101
Iran National Institute of Health Research, Tehran University of Medical Sciences,
51
School of Health Sciences, A.T. Still University, Arizona, Missouri, USA Tehran, Iran
52
Department of Population Medicine and Health Services Research, Bielefeld 102
Faculty of Life Sciences and Medicine, King’s College London, London, UK
University, Bielefeld, Germany 103
Department of Pediatric Medicine, Nishtar Medical University, Multan, Pakistan
53
Public Health Department, Haramaya University, Harar, Ethiopia 104
Department of Pediatrics & Pediatric Pulmonology, Institute of Mother & Child
54
Non-­Communicable Diseases (NCD), World Health Organization (WHO), New Delhi, Care, Multan, Pakistan
India 105
General Surgery Department, Carol Davila University of Medicine and Pharmacy,
55
Department of Public Health, Erasmus University Medical Center, Rotterdam, Bucharest, Romania
Netherlands 106
General Surgery Department, Emergency Hospital of Bucharest, Bucharest,
56
Centre for International Health and Section for Ethics and Health Economics, Romania
University of Bergen, Bergen, Norway 107
Institute for Global Health Innovations, Duy Tan University, Hanoi, Vietnam
57
School of Public Health, Curtin University, Perth, Western Australia, Australia 108
Department of Psychiatry and Behavioural Neurosciences, McMaster University,
58
Center of Excellence in Behavioral Medicine, Nguyen Tat Thanh University, Ho Chi Hamilton, Ontario, Canada
Minh City, Vietnam 109
Department of Psychiatry, University of Lagos, Lagos, Nigeria
59
Department of Pediatrics, Dell Medical School, University of Texas Austin, Austin, 110
Department of Pathology and Molecular Medicine, McMaster University, Hamilton,
Texas, USA Ontario, Canada
60
Kasturba Medical College, Manipal Academy of Higher Education, Manipal, India 111
Department of Forensic Medicine, Kasturba Medical College, Manipal Academy of
61
Department of Legal Medicine and Bioethics, Carol Davila University of Medicine Higher Education, Manipal, India
and Pharmacy, Bucharest, Romania 112
Parasitology and Mycology, Shiraz University of Medical Sciences, Shiraz, Iran
62
Clinical Legal Medicine Department, National Institute of Legal Medicine Mina 113
Department of Pediatrics, RD Gardi Medical College, Ujjain, India
Minovici, Bucharest, Romania 114
63 Health Sciences Department, Muhammadiyah University of Surakarta, Sukoharjo,
Division of Information and Computing Technology, College of Science and
Indonesia
Engineering, Hamad Bin Khalifa University, Doha, Qatar 115
64 Department of Chemistry, Sharif University of Technology, Tehran, Iran
Qatar Foundation for Education, Science, and Community Development, Doha, 116
College of Medicine, University of Central Florida, Orlando, Florida, USA
Qatar 117
65 College of Graduate Health Sciences, A.T. Still University, Mesa, Arizona, USA
Department of Community Medicine, University of Ibadan, Ibadan, Nigeria 118
66 University Institute of Public Health, University of Lahore, Lahore, Pakistan
Department of Family Medicine, Bangalore Baptist Hospital, Bangalore, India 119
67 Public Health Department, University of Health Sciences, Lahore, Pakistan
Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical 120
Surgery Department, University of Minnesota, Minneapolis, Minnesota, USA
Sciences, Tehran, Iran 121
68
School of Psychology and Public Health, La Trobe University, Melbourne, Victoria, Surgery Department, University Teaching Hospital of Kigali, Kigali, Rwanda
122
Australia Emergency Department, Shahid Beheshti University of Medical Sciences, Tehran,
69
School of Public Health and Community Medicine, University of New South Wales, Iran
123
Sydney, New South Wales, Australia Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences,
70
Institute of Medicine, University of Colombo, Colombo, Sri Lanka Tehran, Iran
124
71
Faculty of Graduate Studies, University of Colombo, Colombo, Sri Lanka Department of Entomology, Ain Shams University, Cairo, Egypt
125
72
Gastrointestinal and Liver Disease Research Center, Guilan University of Medical Health Economics, Bangladesh Institute of Development Studies (BIDS), Dhaka,
Sciences, Rasht, Iran Bangladesh
126
73
Social Determinants of Health Research Center, Qazvin, Iran Department of Psychology, University of Alabama at Birmingham, Birmingham,
74
Health Services Management Department, Qazvin University of Medical Sciences, Alabama, USA
127
Qazvin, Iran Emergency Department, Manian Medical Centre, Erode, India
128
75
Department of Forensic Medicine and Toxicology, All India Institute of Medical Department of Health Promotion and Education, Alborz University of Medical
Sciences, Jodhpur, India Sciences, Karaj, Iran
129
76
Hematology-­Oncology and Stem Cell Transplantation Research Center, Tehran Independent Consultant, Karachi, Pakistan
130
University of Medical Sciences, Tehran, Iran College of Medicine, Yonsei University, Seodaemun-­gu, South Korea
131
77
Pars Advanced and Minimally Invasive Medical Manners Research Center, Iran Division of Cardiology, Emory University, Atlanta, Georgia, USA
132
University of Medical Sciences, Tehran, Iran Department of Forensic Medicine, Kathmandu University, Dhulikhel, Nepal
133
78
Epidemiology Department, Faculty of Public Health and Tropical Medicine, Jazan Medical Surgical Nursing Department, Urmia University of Medical Science, Urmia,
University, Jazan, Saudi Arabia Iran
134
79
Epidemiology and Biostatistics Department, Health Services Academy, Islamabad, Emergency Nursing Department, Semnan University of Medical Sciences, Semnan,
Pakistan Iran
135
80
Department of Nutrition and Health Science, Ball State University, Muncie, Indiana, Department of Psychology, Deakin University, Burwood, Victoria, Australia
136
USA Department of Agriculture and Food Systems, University of Melbourne, Melbourne,
81
Department of Anthropology, Panjab University, Chandigarh, India Victoria, Australia
137
82
Institute of Clinical Physiology, National Research Council, Pisa, Italy Center for Health Resource and Services Research and Development, National
83
University of Melbourne, Melbourne, Victoria, Australia Institute of Health Research & Development, Jakarta, Indonesia
84 138
Pathology Department, College of Medicine, Imam Abdulrahman Bin Faisal Department of Pediatrics, King Saud University, Riyadh, Saudi Arabia
139
University, Dammam, Saudi Arabia College of Medicine, Alfaisal University, Riyadh, Saudi Arabia
85 140
Department of Public Health, Trnava University, Trnava, Slovakia Department of Public Health, Adigrat University, Adigrat, Ethiopia
86 141
Health Education and Research Department, SDM College of Medical Sciences & Argentine Society of Medicine, Buenos Aires, Argentina
142
Hospital, Dharwad, India Velez Sarsfield Hospital, Buenos Aires, Argentina

Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 11

 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
143
Psychosocial Injuries Research Center, Ilam University of Medical Sciences, Ilam, 6 Rahman A, Mashreky SR, Chowdhury SM, et al. Analysis of the childhood fatal
Iran drowning situation in Bangladesh: exploring prevention measures for low-­income
144
Division of Injury Prevention and Mental Health Improvement, National Center for countries. Inj Prev 2009;15:75–9.
Chronic and Non-­Communicable Disease Control, Chinese Center for Disease Control 7 Peden AE, Franklin RC, Clemens T. Exploring the burden of fatal drowning and data
and Prevention, Beijing, China characteristics in three high income countries: Australia, Canada and New Zealand.
145
Department of Psychopharmacology, National Center of Neurology and Psychiatry, BMC Public Health 2019;19:794.
Tokyo, Japan 8 James SL, Abate D, Abate KH, et al. Global, regional, and national incidence,
146
Department of Epidemiology and Biostatistics, Wuhan University, Wuhan, China prevalence, and years lived with disability for 354 diseases and injuries for 195
147
Global Health Institute, Wuhan University, Wuhan, China countries and territories, 1990-2017: a systematic analysis for the global burden of
148
Department of Health Economics and Management, Urmia University of Medical disease study 2017. Lancet 2018;392:1789–858.
Science, Urmia, Iran 9 Haagsma JA, Graetz N, Bolliger I, et al. The global burden of injury: incidence,
149 mortality, disability-­adjusted life years and time trends from the global burden of
Department of Medicine, School of Clinical Sciences at Monash Health, Monash
University, Melbourne, Victoria, Australia disease study 2013. Inj Prev 2016;22:3–18.
150 10 Quan L, Bierens JJLM, Lis R, et al. Predicting outcome of drowning at the scene: a
Department of Preventive Medicine, Wuhan University, Wuhan, China
systematic review and meta-­analyses. Resuscitation 2016;104:63–75.
11 Centers for Disease Control and Prevention. Web-­Based injury statistics query and
Acknowledgements Dr. Franklin reports other from Royal Life Saving Society -
reporting system (WISQARS): centers for disease control and prevention, 2019.
Australia, during the conduct of the study. Dr. Bhaumik reports grants from Royal Available: https://www.​cdc.g​ ov/​injury/​wisqars/​index.h​ tml [Accessed 6 Jun 2019].
National Lifeboat Institution, UK, outside the submitted work. Dr. Driscoll reports 12 Poumanyvong P, Kaneko S, Dhakal S. Impacts of urbanization on national transport
grants from World Health Organization, during the conduct of the study. Dr. Ivers and road energy use: evidence from low, middle and high income countries. Energy
reports grants from Royal National LifeBoat Institute, outside the submitted work. Policy 2012;46:268–77.
Dr. Khubchandani reports grants from Merck Research Laboratories, outside the 13 Saunders CJ, Sewduth D, Naidoo N. Keeping our heads above water: a systematic
submitted work. Dr. James reports grants from Sanofi Pasteur, outside the submitted review of fatal drowning in South Africa. S Afr Med J 2018;108:61–8.
work. 14 Peden AE, Franklin RC, Mahony AJ, et al. Using a retrospective cross-­sectional
study to analyse unintentional fatal drowning in Australia: ICD-10 coding-­based
Funding This study was funded by the Bill and Melinda Gates Foundation. RCF and methodologies verses actual deaths. BMJ Open 2017;7:e019407.
AEP were support by funds from the Royal Life Saving Society – Australia. 15 Lunetta P, Penttilä A, Sajantila A. Drowning in Finland: ’external cause’ and ’injury’
Map disclaimer The depiction of boundaries on the map(s) in this article do not codes. Inj Prev 2002;8:342–4.
imply the expression of any opinion whatsoever on the part of BMJ (or any member 16 Clemens T, Tamim H, Rotondi M, et al. A population based study of drowning in
of its group) concerning the legal status of any country, territory, jurisdiction or area Canada. BMC Public Health 2016;16:559.
or of its authorities. The map(s) are provided without any warranty of any kind, either 17 Lunetta P, Smith GS, Penttilä A, et al. Unintentional drowning in Finland 1970-2000: a
express or implied. population-b­ ased study. Int J Epidemiol 2004;33:1053–63.
18 Linnan M, Rahman A, Scarr J. Child Drowning: Evidence for a newly recognized cause
Competing interests Dr. Franklin reports receiving airfare reimbursement from of child mortality in low and middle income countries in Asia’, Working Paper 2012-
Royal Life Saving Society - Australia, during the conduct of the study. Dr. Bhaumik 07, Special Series on Child Injury. 2. Florence: UNICEF Office of Research, 2012. ISBN:
reports grants from Royal National Lifeboat Institution, UK, outside the submitted 1014-7837. https://www.​unicef-i​ rc.​org/​publications/​663-c​ hild-​drowning-​evidence-​
work. Dr. Driscoll reports grants from World Health Organization, during the for-​a-​newly-r​ ecognized-​cause-​of-​child-m
​ ortality-​in-​low-​and.h​ tml
conduct of the study. Dr. Ivers reports grants from Royal National LifeBoat Institute, 19 Haw C, Hawton K. Suicide and self-­harm by drowning: a review of the literature. Arch
outside the submitted work. Dr. Khubchandani reports grants from Merck Research Suicide Res 2016;20:95–112.
Laboratories, outside the submitted work. Dr. James reports grants from Sanofi 20 Dicker D, Nguyen G, Abate D, et al. Global, regional, and national age-­sex-­specific
Pasteur, outside the submitted work. mortality and life expectancy, 1950-2017: a systematic analysis for the global burden
Patient consent for publication Not required. of disease study 2017. Lancet 2018;392:1684–735.
21 Kyu HH, Abate D, Abate KH, et al. Global, regional, and national disability-­adjusted
Provenance and peer review Not commissioned; externally peer reviewed. life-­years (DALYs) for 359 diseases and injuries and healthy life expectancy (HALE) for
Data availability statement Data are available in a public, open access 195 countries and territories, 1990-2017: a systematic analysis for the global burden
repository. Data are available upon reasonable request. Data may be obtained from of disease study 2017. Lancet 2018;392:1859–922.
a third party and are not publicly available. Select input data and all outcome results 22 Murray CJL, Callender CSKH, Kulikoff XR, et al. Population and fertility by age and
are available at ​ghdx.​healthdata.​org. sex for 195 countries and territories, 1950-2017: a systematic analysis for the global
burden of disease study 2017. Lancet 2018;392:1995–2051.
Open access This is an open access article distributed in accordance with the
23 GBD 2017 Risk Factor Collaborators. Global, regional, and national comparative risk
Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits
assessment of 84 behavioural, environmental and occupational, and metabolic risks or
others to copy, redistribute, remix, transform and build upon this work for any
clusters of risks for 195 countries and territories, 1990-2017: a systematic analysis for
purpose, provided the original work is properly cited, a link to the licence is given,
the global burden of disease study 2017. Lancet 2018;392:1923–94.
and indication of whether changes were made. See: https://​creativecommons.​org/​
24 Lu T-­H, Lunetta P, Walker S. Quality of cause-­of-­death reporting using ICD-10
licenses/​by/​4.​0/.
drowning codes: a descriptive study of 69 countries. BMC Med Res Methodol
2010;10:30.
ORCID iDs
25 Morgan D, Ozanne-­Smith J. Measurement of a drowning incidence rate combining
Richard Charles Franklin http://​orcid.​org/​0000-​0003-​1864-​4552
direct observation of an exposed population with mortality statistics. Int J Inj Contr
Amy E Peden http://​orcid.​org/​0000-​0002-​6424-​1511
Saf Promot 2015;22:209–14.
Spencer L James http://​orcid.​org/​0000-​0003-​4653-​2507
26 Mitchell RJ, Williamson AM, Olivier J. Estimates of drowning morbidity and mortality
adjusted for exposure to risk. Inj Prev 2010;16:261–6.
27 Peden AE, Franklin RC, Leggat PA. Exploring visitation at rivers to understand
References drowning risk. Inj Prev 2019;25.
1 World Health Organization. Global report on drowning: preventing a leading killer. 28 Franklin R, Peden A. Improving pool fencing legislation in Queensland, Australia:
Geneva: World Health organization, 2014. ISBN: 978 92 4 156478 6. https://www.​ attitudes and impact on child drowning fatalities. Int J Environ Res Public Health
who.i​ nt/​violence_​injury_​prevention/g​ lobal_​report_​drowning/​en/ 2017;14:1450.
2 van Beeck EF, Branche CM, Szpilman D, et al. A new definition of drowning: towards 29 Prameprart M, Lim A, Tongkumchum P. Modeling unintentional drowning mortality
documentation and prevention of a global public health problem. Bull World Health rates in Thailand, 2000-2009. Asia Pac J Public Health 2015;27:NP2471–9.
Organ 2005;83:853–6. 30 Rahman A, Giashuddin SM, Svanström L, et al. Drowning--a major but neglected child
3 World Health Organization. Preventing drowning: an implementation guide. Geneva: health problem in rural Bangladesh: implications for low income countries. Int J Inj
World Health organization, 2017. ISBN: 978-92-4-151193-3. https://www.​who.​int/​ Contr Saf Promot 2006;13:101–5.
violence_​injury_​prevention/​drowning/​drowning_​prevention_​guide/​en/ 31 Brenner RA. Childhood drowning is a global concern. BMJ 2002;324:1049–50.
4 Roth GA, Abate D, Abate KH, et al. Global, regional, and national age-­sex-­specific 32 Wallis BA, Watt K, Franklin RC, et al. Interventions associated with drowning
mortality for 282 causes of death in 195 countries and territories, 1980-2017: prevention in children and adolescents: systematic literature review. Inj Prev
a systematic analysis for the global burden of disease study 2017. Lancet 2015;21:195–204.
2018;392:1736–88. 33 Bugeja L, Franklin RC. An analysis of stratagems to reduce drowning deaths of young
5 Tyler MD, Richards DB, Reske-­Nielsen C, et al. The epidemiology of drowning in low- children in private swimming pools and Spas in Victoria, Australia. Int J Inj Contr Saf
and middle-­income countries: a systematic review. BMC Public Health 2017;17:413. Promot 2013;20:282–94.

12 Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484

 Original research

Inj Prev: first published as 10.1136/injuryprev-2019-043484 on 20 February 2020. Downloaded from http://injuryprevention.bmj.com/ on March 2, 2020 by guest. Protected by copyright.
34 Rahman F, Bose S, Linnan M, et al. Cost-­effectiveness of an injury and drowning 39 Sansiritaweesook G, Muangsom N, Kanato M, et al. Effectiveness of community
prevention program in Bangladesh. Pediatrics 2012;130:e1621–8. participation in a surveillance system initiative to prevent drowning in Thailand. Asia
35 Gupta M, Rahman A, ul Baset K, et al. Complexity in implementing community Pac J Public Health 2015;27:NP2677–89.
drowning reduction programs in southern Bangladesh: a process evaluation protocol. 40 Foreman KJ, Lozano R, Lopez AD, et al. Modeling causes of death: an integrated
Int J Environ Res Public Health 2019;16:968. approach using CODEm. Popul Health Metr 2012;10:1.
41 Peden AE, Franklin RC, Leggat PA. Fatal river drowning: the identification of research
36 Wintemute GJ, Anton A, Andrada E, et al. Compliance with an ordinance requiring the
gaps through a systematic literature review. Inj Prev 2016;22:202–9.
use of personal flotation devices by children in public waterways. West J Emerg Med
42 Pajunen T, Vuori E, Vincenzi FF, et al. Unintentional drowning: role of medicinal drugs
2013;14:200–3. and alcohol. BMC Public Health 2017;17:388.
37 Olaisen RH, Flocke S, Love T. Learning to swim: role of gender, age and practice in 43 Peden AE, Franklin RC, Leggat PA. Alcohol and its contributory role in fatal drowning
Latino children, ages 3-14. Inj Prev 2018;24:129–34. in Australian rivers, 2002-2012. Accid Anal Prev 2017;98:259–65.
38 Suominen PK, Vähätalo R, Sintonen H, et al. Health-­related quality of life after a 44 Hamilton K, Keech JJ, Peden AE, et al. Alcohol use, aquatic injury, and unintentional
drowning incident as a child. Resuscitation 2011;82:1318–22. drowning: a systematic literature review. Drug Alcohol Rev 2018;37:752–73.

Franklin RC, et al. Inj Prev 2020;0:1–13. doi:10.1136/injuryprev-2019-043484 13