Drowning deaths between 1861 and 2000 in Victoria, Australia

Abstract Objective To identify the long-term patterns of drowning mortality in the state of Victoria, Australia, and to describe the historical context in which the decrease occurred. Methods We obtained data on drowning deaths and population statistics from the Australian Bureau of Statistics and its predecessors for the period 1861 to 2000. From these data, we calculated drowning death rates per 100 000 population per year, by gender and age. We reviewed primary and secondary historical resources, such as government and newspaper archives, books and the Internet, to identify changes or events in the state that may have affected drowning mortality. Findings From 1861 to 2000, at least 18 070 people drowned in Victoria. Male drowning rates were higher than those for females in all years and for all ages. Both sexes experienced the highest drowning rate in 1863 (79.5 male deaths per 100 000 population and 18.8 female death per 100 000 population). The lowest drowning rate was documented in 2000 (1.4 male deaths per 100 000 population and 0.3 female deaths per 100 000 population). The reduction patterns of drowning mortality occurred within a historical context of factors that directly affected drowning mortality, such as the improvement in people’s water safety skills, or those that incidentally affected drowning mortality, like infrastructure development. Conclusion We identified patterns of reduction in drowning mortality, both in males and females and across age groups. These patterns could be linked to events and factors that happened in Victoria during this period. These findings may have relevance to current developing communities.


Introduction
Death due to drowning is a global problem. In 2012, an estimated 372 000 people drowned worldwide 1 and the majority of such deaths occurred in low-and middle-income countries. The prevention of drowning deaths requires a complete understanding of the contributing factors, but the knowledge about drowning deaths in developing communities is sparse. Although epidemiological data have begun to emerge, they provide only a brief snapshot of the existing conditions. Some aspects of prevention, such as the development of water survival skills, for example swimming and rescue, or mechanisms to prevent drowning, for example the removal of hazards such as water-filled holes or the construction of barriers around hazards, require a fuller understanding of the changes in drowning patterns over time.
Injury prevention research has made little use of historical research methods. 2 A few studies have examined drowning trends, 3-7 but using relatively short time frames. One study has investigated flood-related drowning deaths between 1788 and 1996. 8 Historical perspectives allow a depth of understanding that comes from seeing the phenomena in continuous and constant change and identifying the dynamics of that change. 9 This study uses this perspective in a retrospective examination of long-term drowning death trends in a developing community, as the community evolved from exhibiting a high drowning risk to a low drowning risk.
In the financial year 2014-2015, the Australian state of Victoria had an unintentional drowning mortality rate of 0.67 deaths per 100 000 population. 10 However, in 1905, Victoria's drowning rate was 12 deaths per 100 000 population (Staines C and Ozanne-Smith J, Monash University, unpublished data, September 2006) resembling the rates in the World Health Organization (WHO) African Region (7.9 deaths per 100 000 population) and the WHO South-East Asia Region (7.4 deaths per 100 000 population) in 2012. 1 Here we examine the changes in drowning incidence in Victoria between 1861 and 2000 and associate trends with historical factors. The results could be an epidemiological basis for a broader programme of research investigating the factors contributing to drowning deaths in this period.

Drowning mortality
We obtained drowning mortality and population data for Victoria from records of the Australian Bureau of Statistics and its predecessors for the period 1861 to 2000. No drowning mortality data were available for three years (1906, 1914 and 1931). The gender of those who drowned was available for all years except 1862 and 1917.
We calculated drowning death rates (deaths per 100 000 population per year) using either Australian Bureau of Statistics historical population data or estimates based on these data as the denominator.
To determine trend lines, we fitted structural time series models to the drowning data using the STAMP software package version 7.04 (Timberlake Consultants Ltd., London, United Kingdom of Great Britain and Northern Ireland). 11

Age
Because of variation in the reporting practices of the statistical agencies over time, age specific drowning data, and corresponding age specific population statistics, were available for only a limited number of years in the 19th and 20th centuries. We used the available data to determine drowning mortality Objective To identify the long-term patterns of drowning mortality in the state of Victoria, Australia, and to describe the historical context in which the decrease occurred. Methods We obtained data on drowning deaths and population statistics from the Australian Bureau of Statistics and its predecessors for the period 1861 to 2000. From these data, we calculated drowning death rates per 100 000 population per year, by gender and age. We reviewed primary and secondary historical resources, such as government and newspaper archives, books and the Internet, to identify changes or events in the state that may have affected drowning mortality. Findings From 1861 to 2000, at least 18 070 people drowned in Victoria. Male drowning rates were higher than those for females in all years and for all ages. Both sexes experienced the highest drowning rate in 1863 (79.5 male deaths per 100 000 population and 18.8 female death per 100 000 population). The lowest drowning rate was documented in 2000 (1.4 male deaths per 100 000 population and 0.3 female deaths per 100 000 population). The reduction patterns of drowning mortality occurred within a historical context of factors that directly affected drowning mortality, such as the improvement in people's water safety skills, or those that incidentally affected drowning mortality, like infrastructure development. Conclusion We identified patterns of reduction in drowning mortality, both in males and females and across age groups. These patterns could be linked to events and factors that happened in Victoria during this period. These findings may have relevance to current developing communities.
Drowning deaths were analysed in 5-year age groupings, in accordance with those employed by the Australian Bureau of Statistics.

Historical context
To provide contextual information and to identify changes or events in the state that may have affected drowning mortality during the period of interest, we accessed a range of historical and contemporary resources. These included data from primary sources such as government department archives, primarily retrieved from Victoria's Public Record Office and secondary sources, such as newspaper archives (available as either hard copies from newspaper and library archives or online sources, primarily the Australian online library database Trove). We also accessed historical and contemporary books and other literature from university or public libraries and electronic sources were accessed through the Internet.

Results
From 1861 to 2000, at least 18 070 people died due to drowning in Victoria.
Before 1968, drowning statistics combined unintentional deaths with those where intent of death was undetermined. Therefore, the drowning death rates for the combined statistics (1861-1967) are shown in Fig. 1 and unintentional drowning death rates for the period 1968-2000 in Fig. 2. The overall pattern shows a reduction in drowning mortality, with total rates showing marked reductions from a maximum of 53.5 deaths per 100 000 population in 1863 to 0.8 deaths per 100 000 population by 2000. However, there were deviations from the general trend of reduction, there being some years of transient increase and periods of relatively little change.
A substantial proportion of the reduction occurred by the end of the 19th century. Based on 10-year averages, the total drowning rate decreased by 31.5%,   Year 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 Fig. 3 shows age-specific drowning rates for males and Fig. 4 shows rates for females. Males aged 65 years or older had the highest drowning rates until the early-20th century. However, by the mid-20th century this had shifted to children younger than 5 years ( Fig. 5 and Fig 6). The drowning mortality rates in the mid-19th century were high for males aged 65 years or older and females aged 70 years or older, 65.6 deaths per 100 000 population and 32.6 deaths per 100 000 population, respectively. However, the proportions of total drowning deaths were small with only 1.7% (20/1169) of male and 1.4% (4/281) of female deaths. The pattern of high mortality rates for older people continued in the late-19th century for males and females aged 70 years or older, 66.6 deaths per 100 000 population and 30.3 deaths per 100 000 population, respectively, comprising 4% (43/1075) of male deaths and 5.6% (13/232) of female deaths.

Age
While differing in magnitude, the drowning rates showed similar age-related patterns for males and females ( Fig. 5 and Fig 6). Only in the early-20th century, were there marked differences between males and females in the patterns for the younger and older age groups.

Sex
Both males and females showed reductions in drowning deaths over the period. For all years, the mortality rates for males were greater than for females. For both sexes, we observed the highest rates in 1863, 79.5 male deaths per 100 000 population and 18.8 female deaths per 100 000 population. The lowest rates were reported in 2000 with 1.4 male deaths per 100 000 population and 0.3 female deaths per 100 000 population.

Males
In the mid-19th century, all age groups had a very high drowning mortality of at least 31.3 male deaths per 100 000 population per year (20-24 year olds). The drowning mortality decreased in each subsequent era. Rates remained high until the early-20th century for all age groups (although unidentified intentional deaths may inflate the rates for this era). By the mid-20th century, only boys younger than 5 years had a high mortality, 9.4 deaths per 100 000 population per year (Fig. 3). There were several exceptions to the pattern of reduction. In the late-19th century, men aged 55 years or older continued to experience mortal-ity rates resembling those of the previous era. Similarly, there was little change for [15][16][17][18][19] year olds from the late-19th century to the early-20th century and for boys younger than 5 years from the early-to mid-20th century.

Research
Drowning deaths in Victoria, Australia Carolyn Staines & Joan Ozanne-Smith The age-specific patterns for drowning mortality differed between the eras. For example, for males, the age group showing the lowest level of drowning mortality varied over time: 20-24 year olds in the mid-19th century; [15][16][17][18][19] year olds in the late-19th century; 30-34 year olds in the early 20th century; and 10-14 year olds in the mid-20th century (Fig. 3).
Boys younger than 5 years had high drowning mortality in all eras, with the mid-19th century showing the highest levels of mortality of 51.4 deaths per 100 000 population per year.

Females
The patterns of female drowning deaths are difficult to interpret. For most age groups there were so few cases that the patterns showed considerable variability. However, as with boys, girls younger than 5 years had relatively high levels of drowning mortality in all eras, with the mid-19th century showing the highest mortality (30.2 deaths per 100 000 population per year; Fig. 4). Fig. 7 shows a summary of the historical context of the drowning trends. In the 19th century, there were few coordinated efforts addressing drowning. While there were several individual factors that may have contributed to a reduction in drowning mortality, such as the establishment of bathing facilities providing safer venues for swimming and the commencement of swimming clubs, these did not show a synergistic effect until the 20th century. Much of the reduction in mortality in the 19th century was more likely due to the development of general infrastructure and the urbanization of the population.

Infrastructure development
Several aspects of the development of the physical infrastructure in the state could have affected drowning mortality. The development of a piped water supply system 12 removed the need for reliance on hazardous water supplies or storage systems. Safe bridges and roadways, street lighting, the removal of water hazards, the construction of safety barriers, more substantial housing with perimeter fencing and the improved management of waterways would all be expected to have contributed to reductions in drowning mortality.

Urbanization
In the 1860s, the majority of residents lived in rural locations and 42% of people, including large numbers of children, lived in gold mining areas. 13 By the 1880s, the population began to move away from rural locations, where they were exposed to many unprotected water hazards, and by the 1890s, almost half of the population was living in urban Melbourne. 13

Other factors
The introduction of mandatory school attendance in 1872 could have contributed to a reduction in child drowning, by reducing the amount of time that children were unsupervised and in hazardous environments. In 1910, a coordinated programme for swimming and lifesaving instructions was established in state schools. The programme also included certification of swimming teachers, which increased the state's capacity to provide children with water safety skills.
Much of the development of water survival skills in Victoria was associated with a supporting infrastructure of lifesaving agencies. In 1874, the Victorian Humane Society was established and while not having drowning prevention as its sole focus, it laid the basis for the later establishment of organizations whose primary aim was the prevention of drowning deaths. The Royal Life Saving Society Australia, Victoria Branch, took over the drowning prevention functions from the Victorian Humane Society in 1904 and in1947, the Surf Life Saving Society of Victoria joined this prevention effort. These two organizations have operated continuously in Victoria since their inception and have played important roles in the development of a culture of water safety, Drowning deaths in Victoria, Australia Carolyn Staines & Joan Ozanne-Smith instruction in swimming, water safety and rescue and resuscitation skills. They also provide lifeguard services at recreational swimming venues. One of the early barriers to developing water survival skills was the lack of safe swimming venues. However, in the mid-19th century, in response to the lack of private washing facilities in Melbourne, several public bathing houses were opened. 14 While the initial focus was the improvement of hygiene, people started using the bathhouses for more recreational purposes and an interest in recreational bathing and swimming began to develop. In the 1870s, the first clubs for the promotion of swimming as a competitive sport were established. In the following century, hundreds of swimming clubs opened in urban and rural locations, which provided an infrastructure for population-wide exposure to water safety training.
The Games of the XVI Olympiad, held in Victoria in 1956, further led to the development of safe swimming venues. Building on the state's strong interest in competitive swimming, the Olympic Games provided a further focus for the development of swimming infrastructure and resulted in a marked increase in the construction of Olympic sized swimming pools across the state. 15 We also found that newspapers played a role in addressing the problem of drowning mortality. In the 19th century, newspapers reported coronial death inquests in considerable detail. The publications of drowning incident details could have contributed to an increased awareness of drowning hazards and may have generated an expectation in the community that authorities could and should do something to address the problems.
Newspapers continued to have a role in raising awareness throughout the 20th century. One metropolitan newspaper, in partnership with the education department and lifesaving agencies, established the state-wide Herald Learnto-Swim programme in 1928.

Discussion
Here we demonstrate that Victoria experienced a substantial reduction in drowning deaths between 1861 and 2000, both for males and females and across all ages. Contemporary drowning death statistics show that, worldwide, males experience higher drowning mortality than females. 16 The results of this study concur with this pattern and that, in Victoria, the pattern has prevailed at least since 1861.
In this study, we used historical drowning statistics reported by the Australian Bureau of Statistics and its predecessors. Using historical statistics can be associated with problems due to changes in reporting practices. Before 1968, the Bureau aggregated unintentional drowning deaths and undetermined drowning deaths. Therefore, we reported the data separately for the years 1861-1967 and 1968-2000, since undetermined drowning deaths might have inflated mortality statistics before 1968. A review of a sample of undetermined deaths suggested that while 19th century statistics were relatively unaffected by undetermined deaths, early 20th century statistics may have included substantial proportions of these deaths. 17 These findings suggest that one should interpret 20th century statistics before 1968 with caution.  1862  1864  1866  1868  1870  1872  1874  1876  1878  1880  1882  1884  1886  1888  1890  1892  1894  1896  1898  1900 1902 1904 1906 1908 1910 1912 1914 1916 1918 1920 1922 1924 1926 1928 1930 1932 1934 1936 1938 1940 1942 1944 1946 1948 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 Year We found very high levels of drowning mortality in children younger than 5 years in the mid-19th century. Studies done in contemporary developing communities have also reported similar or higher mortality levels. In Bangladesh, in 2003, drowning death rates for children younger than 5 years were reported to be 33.1 per 100 000 population in urban locations and 158.7 per 100 000 population in rural locations. 18 In the Inner Mongolia Region of China, in 2012, children aged 1-4 years were found to have a drowning rate of 31.1 per 100 000 population. 19 We also found that people aged 70 years or older had the highest drowning rates until the early-20th century. However, these results should be interpreted with caution, since the population size of this age group was small in this period and hence could have contributed to an overestimation of the mortality rates.
The mortality trends identified in this study show that while there is an overall decline in drowning mortality, the path is neither smooth nor steady. Some periods show a sharp decline, whereas others show an increase or no change. These changes in pattern are of potential interest, as they suggest that events might have happened that have induced these changes. Therefore, further investigating these changes would help to determine the possible underlying factors.
In Victoria, a substantial reduction in the number of drowning deaths occurred in the 19th century when there were few drowning prevention strategies in operation. Instead, mortality reduction at this time appeared to be an incidental outcome caused by sev-eral factors that indirectly resulted in a reduced exposure to hazards.
The implementation of specific drowning prevention strategies began in the late-19th century, but it appeared to take time before these strategies were working effectively. Coordinated and sustained drowning prevention strategies were not implemented until the 20th century. Following the establishment of life saving agencies, water skills' programmes, community awareness strategies and regulatory interventions, drowning mortality continued to decline during the 20th century.