Study design

A population-based cross-sectional study of suicide at coastal locations in Australia was conducted.

Inclusion and exclusion criteria

Cases were included in the study that met the following eligibility criteria:

• the incident occurred between 1 January 2005–31 December 2019 (inclusive)
• the incident occurred at a coastal location comprising between onshore and 12 nautical miles (nm) offshore where onshore locations included (but were not limited to) beaches and surrounding public amenities (e.g. nearby carparks and sand dunes), rocky cliff locations; nearshore environments, and offshore waters up to 12 nm
• the incident outcome was death and was notified to the Coroner
• the manner of death was determined by the Coroner as intentional self-harm
• the investigation was completed by the Coroner at the time of data extraction (31^st August 2020)
• The deceased was 18 years or older at the time of the incident.

Cases were excluded where the suicide followed a murder, thus removing both the deceased and offender. Cases classified with undetermined intent (n = 7) were also excluded from analyses.

Data sources

The primary data source for this study was the National Coronial Information System (NCIS), an electronic database of deaths notified to Australian Coroners from July 2000 (2001 in Queensland) and deaths notified to New Zealand Coroners from July 2007. The NCIS is an electronic database for which the Department of Justice and Community Safety as the source organisation of the data. The NCIS contains coded and free-text information on the decedent, the incident and cause and manner of death. The NCIS also includes up to four full text reports: police narrative of circumstances of the death; autopsy report; toxicology report; and coroners’ finding. Data from the NCIS was supplemented with Surf Life Saving Australia’s (SLSA) Surf Guard Incident Report Database (IRD). Surf Guard IRD stores information uploaded by surf lifesaving personnel from incident reports following the provision of medical treatment or search and rescue operations.

Case identification

Deaths were identified using the key word and “query design” in the NCIS. The cause of death field was searched for the key words “drown*” and “immersion” and a query was performed where the variable “Location Level 2” was classified as “beach, shore, bank of a body of water”, “large area of water”, “stream of water”, “marsh, swamp”, “drain, channel”, “lookout, viewpoint”, “base of precipice, cliff”, “wharf, pier, jetty”, “watercraft” or “boat ramp”. The search results included the variable “intent—completion” and “activity” which enabled records returned to identify cases where the Coroner determined the death was a result of intentional self-harm (hereafter referred to as suicide). A review of the Coroners’ finding for each case returned from the search was reviewed against the inclusion and exclusion criteria. Where a Coroners’ finding was not available in the NCIS, the police narrative of circumstances of the death and variables: “cause of death”, “intent–completion”, “location level 2” and “activity” were reviewed against the inclusion criteria.

Data extraction and management

For each death that met the inclusion criteria, the NCIS record was reviewed and information was extracted on the decedent, the incident and relevant historical and proximate circumstances. Information on the decedent comprised: age; sex; and birth continent. Incident factors comprised: year, season, month of year, day of week and time of day; public holiday; date of personal significance; location; remoteness (coded in accordance with the Australian Statistical Geographical Classification—Remoteness Areas); and external cause of death (coded in accordance with the WHO’s International Classification of Diseases, Tenth Revision [ICD-10]). Relevant historical and proximate circumstances were reviewed to identify risk factors which comprised: evidence of alcohol or drug consumption; evidence of mental ill health; and evidence of a history of suicidal behaviour. Raw data were only accessed, managed and analysed by JL and LS in accordance with ethical agreements.

Where an ICD-10 external cause code was not recorded in the NCIS, two members of the research team (LB, JL) independently assigned a code based on the cause of death. Conflicts in classification of the ICD-10 codes was resolved by a third member of the research team (AP). Where the incident time of day was estimated as within the preceding 24 hours, the midpoint was used and for cases where the incident time of day was estimated as greater than 24 hours, it was recorded as unknown. Evidence of mental ill health (diagnosed or suspected) and evidence of suicidal behaviour (ideation, previous attempts and self-harm) was identified from the reports attached to the NCIS (where available).

To determine the presence and impact of important dates (such as public holidays and birthdays) on suicide, the following coding was undertaken. Firstly, the date of the incident was identified from the NCIS database. If the specific date of the incident was not known, the following steps were undertaken:

• If the date range was greater than three days it was classified as unknown and excluded from the date analyses.
• If the date range was less than three days and there was a specific incident on the day last seen indicating that was most likely the day of death (such as the decedent sending a suicide text message to somebody) then that was used as the incident day.
• If there was no information clarifying when the incident occurred and the time frame was less than three days, the average day was used.

The decedent’s date of birth as listed in NCIS was used. If the date of birth was unknown this case was excluded from the data analyses. A separate database was compiled of all public holidays for each state for each year within the study period, including the holiday date, name and year of the holiday.

Data analysis

Ethics

This study was conducted with ethics approval from the Department of Justice and Community Safety Human Research Ethics Committee (JHREC CF/19/21185).

Describing Australian coastal suicides

Suicide is a complex issue, with extensive impacts that extend beyond the numbers to people, families and communities [49–52]. This study is the first to describe coastal suicide deaths in Australia framed by suicide method and sex of decedents. In general, the number of coastal suicide deaths in Australia are increasing and the number of suicide deaths was approximately three times higher in males than females, consistent with national suicide reporting [50,51]. Most decedents were Australian-born (59%, n = 391), however 23% were born overseas, and for 18% birth continent was unknown. This is higher than the recently reported proportion of Australian residents that were born overseas [30% in 2019; 53], supporting the framing of coastal environments as ‘lethal locations’ that are purposefully travelled to by decedents for increased anonymity and reduced chances of being interrupted [17,18]. While the highest number of coastal suicides occurred over the summer months and on Mondays, coastal suicides occurred consistently throughout the year, on any day and at any time during the day.

The circumstances under which people choose to take their own life are complex and often there is no single determining factor why an individual dies by suicide [54]. Locations away from home and certain method choices are suggested to have higher rates of lethality due to lower risk of intervention [17,18]. Jumping from a height was the most common method of suicide used in almost half of all coastal suicide deaths (43%), followed by drowning and then other means. The proportional use of these methods deviates considerably from the national trends [55] in which the most common suicide methods (hanging and self-poisoning) were less common in coastal suicides, whereas nationally jumping from a high place is fourth-highest and drowning and submersion is ninth [55]. Jumping from heights and suicidal drowning were more common in coastal environments [56], since the coastline may combine accessibility with availability due to the increased presence of the ocean or naturally elevated structures such as rocky cliffs [56]. In this study, rocky cliff locations recorded the highest number of coastal suicide deaths (50%) followed by beaches and bays, which would directly correlate with jumping from height and drowning methods. These results identify coastal environments as lethal locations but suggest a greater measure of lethality at rocky cliffs, beaches and bays, e.g. jumping from heights at a cliff may be considered more lethal than self-poisoning at home which could be more easily interrupted or have a delayed outcome or require more pre-planning. These results indicate that coastal suicides and their intervention or prevention strategies may require a different, or more localised, suite of responses and strategies with respect to suicides that occur inland or at home.

The costs of suicide to individuals, communities and society are extensive, with certain groups identified as being at greater risk than others [49,50]. Suicide is the leading cause of death for 15–44 year olds, especially men [10,50], with research investigating and supporting preventative strategies that are both community-based and broader, national public health approaches [50,57,58]. While this study excluded those under 18 years of age, our results support this with decedents aged between 25–59 representing the largest cohort, comprising almost two-thirds of all coastal suicides (59%). This trend has also been reported in the United States where suicide rates among adults aged 24–44 years old have exceeded those reported for older adults [65 years or older; 59]. This stage of adult life often encompasses intense employment or relationship development in combination with new responsibilities (i.e. parenthood) and is therefore likely to have multiple co-occurring stressors. For example, unemployment has been strongly associated with male suicide rates in Australia [especially those between 25–34 and 55–64 years of age; 29], however the relationship differed depending on the age group and the duration of unemployment–highlighting the challenge of deciphering causal factors within suicide deaths. The complexity surrounding suicide prevention, and the interrelated contributing factors within these incidents, suggests that individual and community-level strategies that aim to reduce suicidal behaviours through the incorporation of multiple protective factors are required to make communities and individuals more resilient and to reduce suicidal behaviour [29].

Fatal coastal suicide methods

Locations away from home are considered more lethal [18,60], but lethality also differs with the choice of suicide method. Practical aspects, particularly availability and accessibility, are hypothesised to be important drivers of suicide method choice [61,62] and essentially underpin prevention strategies that restrict access of hotspot areas [e.g. bridges; 63]. The impact of psychological and cultural factors on method choice have also been recognised [56] with highly-publicised suicide events, often at formidable locations, touted to increase the number of subsequent suicides that intentionally re-enact the publicised suicide in the months following the initial fatality [60,64]. The motives behind these incidents are suggested to be more symbolic or personally driven rather than by accessibility or availability alone.

Coastal areas in Australia comprise impressive rocky cliffs against wild unpredictable oceans, and given the coast is considered to be part of the Australian psyche, it is plausible that coastal environments combine practical drivers (accessibility and availability) with more symbolic motivators behind suicidal choices that are made by the decedent. The familiarity of, and the perceived capacity to effectively carry out the method, may also influence method choice [64]. For example, familiarity with firing a firearm would be generally lower across Australia since there is strict legislation in place regarding firearm ownership. This legislation significantly limits firearm availability and familiarity, and reduces the rate of Australian firearm-related suicides, in contrast with those recorded in firearm heavy countries [e.g. the United States; 65,66]. Since research that specifically investigates suicide at the individual method level is limited [but see 21,26,35,67–70], this section discusses coastal suicide deaths in the context of suicide method.

Jumping from a high place (X80)

Jumping from a high place was the most common coastal suicide method recorded and was increasing more than other methods. It was the highest cause of death for males, females, most age groups (except for 25–39 year olds and those aged 70 years or older) and was more prevalent in Australian-born decedents. Suicide by jumping from a height was significantly higher at rocky cliff locations and major cities which is likely to reflect the availability of naturally elevated locations along coastal environments with a higher perception of lethality and access by large numbers of people [71,72]. Elevated locations are also prevalent in outer-regional and remote coastal locations, yet jumping from high places was lower in comparison to other means. Since rates of suicide are generally considered to be higher in rural settings [73,74], this suggests that other drivers (e.g. population size, social determinants, or greater accessibility to other means such as firearms) may influence suicide method choice more in regional and remote areas [73]. This highlights a difference in the prevalence of coastal suicides by jumping in relation to regionality, and that suicide prevention strategies to circumvent suicide by jumping maybe be more relevant for coastal locations in major cities than in regional or remote areas.

Suicide by jumping from high places can involve extremely traumatic mechanisms of injury resulting in significant trauma and is strongly associated with an increased incidence of death, which some may consider more lethal than other means [71]. Jumping may also be considered as a more public act of suicide [75,76] and, by some, as a method that reflects a greater intent to die [75]. However, it should be acknowledged that the reasons behind chosen suicide methods are complex and apparently less-lethal means of suicide do not necessarily reflect a lesser desire to die, nor that there is consideration about lethality. With respect to jumping, some naturally elevated sites have gained a level of notoriety as suicide ‘hotspots’. Practical characteristics including accessibility, perceived lethality and symbolic qualities such as media attention or natural features (e.g. overlooking water) are suggested to contribute to this notoriety [72]. Much research has demonstrated that restricting access to these notorious sites leads to a reduction in suicide deaths from jumps at those sites [75,77], with little evidence to suggest substitution of location to alternative sites [63,77]. Since it is impractical and undesirable to restrict access to Australia’s extensive coastline, the implementation of exclusion strategies that target identified coastal suicide hotspots could provide an opportunity for the individual in distress to reconsider their actions [77]. While a national strategy is important to guide prevention efforts, these results emphasise the need for localised strategies that integrate local knowledge and build awareness of protective factors at the community level.

Drowning and submersion (X71)

Drowning-related suicides increased over the study period but at a lower rate than other methods. Decedents aged 70 years and older were more likely to suicide by drowning (57% of all fatalities of people aged 70 years and over), consistent with previous findings [67], while decedents aged 25–39 years were less likely (19%). This may be related to increasing intentional drowning among elderly males and a preference for less violent means among elderly females [67].

Decedents born in Asia were more likely to intentionally drown (57% of all Asian-born coastal suicide deaths), while Australian-born decedents were less likely (22% of all fatalities). This may be linked to cultural practices around suicide, with drowning considered to be a preferred method among people from Asia, in particular females [78]. This preference for suicide by drowning could also be linked to poorer swimming skills and higher drowning rates commonly seen in Asian nations [79,80], since a fatal outcome would be more likely if the decedent does not know how to swim.

Linked to the location-based focus of this study, is research that indicates that drowning and jumping are methods of suicide more commonly seen in natural environments [56]. In this study of coastal environments in Australia, drowning was the second leading method of intentional death (31%), after jumping from a high place (43%). Of note, drowning-related suicides were more likely to occur in bays (72% of all bay suicides) and at beaches (47% of all beach suicides). This provides further strength to the need to consider suicide risk (and associated prevention strategies) from a location perspective [8,22,56], including building awareness of suicidal drowning and the development of suicide-response training by surf lifesaving volunteers.

Other methods in coastal suicides

The ‘other’ means used in coastal suicide deaths included several different methods with hanging, self-poisoning and firearm discharge being the top three causes of death (combined totalling 17% of all coastal suicides). The number of suicide deaths that incorporate other means increased over time, but given the dominance of jumping and drowning as suicide methods at coastal locations, these other means are comparatively low, especially when compared to national trends [55]. Other means of suicide were higher at beaches and offshore waters, locations which would provide greater opportunity for a range of methods that take into account the geography or nature of the environment [64]. For example, it could be considered easier to find a secluded beach with trees and sand dunes to suicide privately in a seemingly less violent manner by poisoning or hanging. Suicide by other means was higher for oceanic-born decedents and more common in outer regional, remote and very remote areas, suggesting that social and economic determinants may be greater drivers of suicide method choice in these communities. Differences in access may also explain this result, especially given that firearm use in Australia is strongly linked to rural locations [81] and regional and remote communities would be more familiar with discharging a firearm [64], therefore increasing its availability as a suicide method compared to urban communities. These results highlight the diversity within coastal suicides and emphasise that prevention strategies may be more effective if developed and implemented at a community level.

Risk factors of coastal suicides

Coastal suicides were higher in males than females, consistent with nationally reported trends [50,51], otherwise differences observed by sex in this study related to evidence of mental health and suicide behaviour reported across many of the incidents. Evidence of mental ill health or a history of suicidal behaviours was reported less for incidents involving males, while a greater proportion of female suicide deaths reported a higher incidence of these behaviours, both individually and in combination. While it is well established that males are less likely to seek help, interpretation of our results must be considered with caution as they do not establish this since it may not have been investigated or reported in the coronial documents. This result is however consistent with national reports where suicidal behaviour is more prevalent in females, while males suicide deaths are more numerous [51]. Evidence presented in the National Survey of Mental Health and Wellbeing (2007) suggests that mental ill health and suicidal behaviours (e.g. ideations, self-harm and previous attempts) are widespread in the community, with an estimated impact on 2.1 million Australians [82]. This suggests that behavioural results from this study are probably more reflective of national suicide patterns and unlikely to be location or method specific.

Birthdays can be emotional times which could have associated triggers, which would be worth acknowledging their potential influence. The risk of coastal suicide was more than 10 times higher during the seven-day period prior to their birthday with a similar trend for the week following their birthday. This result contradicts previous findings where no birthday effects were found [83,84], but supports research from Japan where birthdays, especially those celebrating significant milestones, can increase suicide risk [85,86]. These findings further acknowledge the influence of cultural and social factors in suicide surveillance. While this result is unlikely to be linked to location or method choices within suicide, it is important with respect to implementing more protective measures to high-risk individuals in the period leading up to their birthday. A strategy that implements this would be more effective at the individual level, but could potentially be used at higher levels (i.e. community or national) to prompt professional services to check in around high-risk times (i.e. with medical professionals or within databases where personal information is retained after previous records of suicidal behaviours).

Alcohol and drugs are known suicide risk factors, due to their propensity to impair decision making, dramatically influence our response capacity and impact on our emotional state [87]. Moreover, it has been proposed that suicides in natural areas are more likely to involve softer drugs (alcohol and prescription drugs) instead of illicit drugs [56], a hypothesis that is largely supported by our results. Nearly 40% of coastal suicides with known toxicology results were under the influence of alcohol and/or drugs, in spite of the conservative toxicological inclusion strategy implemented in this study where therapeutic doses of psychiatric medications were excluded. This result is comparable to other suicide research [88], but higher than toxicology findings reported in unintentional coastal drowning deaths (24%) and other unintentional coastal fatalities [22%; 33]. Men were more likely to have consumed alcohol, with an average blood alcohol concentration of 0.18 g/100mL (S2 Table), more than three times the legal driving limit. A similar average blood alcohol concentration was found for females (0.22 g/100mL), but there were fewer cases. Alcohol is a known depressant so understanding its role in suicides is significant, these results indicate that alcohol use in coastal suicides is similar to national trends and is not location specific.

Prescriptions drugs played a greater role in female suicide deaths at coastal locations, largely due to increased contribution of anti-depressants and prescription opioids. This result supports previous research where females had a higher prevalence of both anti-depressant and opioid use in suicide [89,90]. In this study, females had a higher serum concentration of antidepressants than males, which could be due to differences in metabolic enzymes [91]. This result has also been shown in naturalistic clinical settings [91] and, given the higher uptake of these prescription drugs due to increased rates of diagnosis [92], it has been proposed that females may exhibit higher tolerance to anti-depressants [92] and therefore may use higher doses. Females are also reported to have an increased incidence of chronic pain compared to males and may manage this pain with prescription opioids [93]. This, alongside higher prevalence of mental ill health conditions such as depression and anxiety [94] and an increased likelihood to receive treatment by women [and prescribed medication; 95,96], could explain the increased prescription drug concentrations found in female coastal suicide deaths.

Implications

Understanding suicide by location will help develop location and community specific prevention strategies and ensure local knowledge is incorporated into these strategies. Coastal locations are unique, especially in Australia, where there is a workforce (volunteer and paid) who patrol many of the beaches and nearby coastal areas. This workforce is called on to rescue people in trouble, help in the search and rescue of a missing person and also retrieve bodies [33,34]. In the course of their work, they also respond to incidents occurring in the coastal environment but outside of the water (i.e. on rock platforms or in beach carparks). Surf lifesavers, lifeguards and other first responders are often required to attend to intentional incidents, in many cases without knowing the incident was intentional. The adverse impacts of such job-related occupational stressors among first responders, such as post-traumatic stress disorder (PTSD), suicidal thoughts and behaviours, are well-documented [93–100], but as yet unknown for Surf Life Saving [34]. Surf Life Saving Australia, as an organisation, has an obligation to better understand these situations and ensure its members are provided with necessary and appropriate training and equipment to attend such incidents [34]. It is also important to ensure the workforce is adequately supported during and after such incidents, using evidence-based postvention support strategies. Research to better understand impacts of trauma on surf lifesavers while on patrol is currently underway by Surf Life Saving Australia [34].

As part of a broader commitment to saving lives, this present study has identified an opportunity for Surf Life Saving to collaborate with suicide prevention organisations through various means (such as research and information sharing) to identify locations with a high incidence of suicide to guide prevention interventions [77]. The diversity within coastal suicides highlight the importance of integrating local knowledge and community values within national suicide prevention strategies. While it is important to have an overarching national strategy to guide prevention, building awareness and understanding at the community or local level can create greater ownership and therefore implementation of protective measures that could be community or location specific. With the identification of notorious suicide hotspots and a greater understanding of suicide risk factors, the community (including the surf life saving workforce) will be better equipped to not only respond to such incidents, but also to recognise and prevent them from occurring as well as helping to develop localised prevention strategies.

Strengths and limitations

This paper is novel and makes the first steps towards better understanding suicides along the Australian coast. The use of multiple population-based data sources, including the NCIS and Surf Life Saving’s IRD provides rich detail enabling the identification of risk factors to guide prevention efforts of this complex cause of death. There are, however, limitations associated with this study. Data fields on cases still under coronial investigation are subject to change. At the time of analysis, 97.3% (n = 648) of cases of interest were closed. Cases in more recent years were less likely to have intent completion coding, as such, 2019 data were removed from temporal trend analysis. Some variables are poorly reported within the coronial system, such as birth continent and time of incident, leading to large numbers of unknowns. Similarly, incomplete information regarding the rescue or body retrieval effort and the decedent’s history of mental health and suicidal behaviours also represents limitations of this study and areas of improvement for coronial services. Although out of scope for this study was an examination of the residential or tourist status of overseas-born decedents, the initial findings reported in this study support the need for future investigation into the propensity for distances travelled to a particular suicide location.

There were several cases that did not have intentional self-harm ICD-10 codes as their underlying cause of death despite the precipitating factors being an act of suicide. This highlights a challenge with respect to coding deaths and a concern of omitting data when using ICD-10 codes alone for analyses. Similarly, due to the systematic nature of coronial data collection related to cause of death, protective factors are not included in the data. It is important to note that the role they play in an individual’s life is extremely pertinent [52]. Further work is required to increase awareness across data custodians of the importance of complete records and their use post classification.

With respect to the dates of significance analyses, all public holidays were used, as opposed to placing a value judgement on a particular day that may be more significant (i.e. Christmas Day). For incidents which occurred within a date range, the public holidays or birthdays may have been missed, as they may have occurred in the week before or after. This study has not undertaken an exhaustive analysis of psychosocial factors, and it must be acknowledged that suicidal triggers differ at the individual level, therefore the findings of this study represent generalisations that should be interpreted with caution.

The toxicological analyses do not account for complex interactions between drugs or in combination with alcohol (for example when multiple central nervous system depressants such as opioids, benzodiazepines and alcohol are combined an individual is at an increased risk of overdose due to the additive interactions between the substances [101]. Similarly, the concentration analysis in this study assumes a linear dose response and does not account for an individual’s tolerance to a drug or any potential drug redistribution after death.

Toxicological analyses conducted in this paper are conservative with therapeutic levels classified as non-contributary as therapeutic doses of psychiatric mediation generally decreases the chance of suicide [40]. The authors would like to acknowledge that therapeutic levels of drugs used to treat mental health conditions can cause a range of side effects in patients, which can increase suicide risk. However, it would since mental health medication are generally effective at decreasing mental health symptoms and preventing suicide, it would be erroneous to classify the therapeutic presence of any as being contributory [40].