Introduction

Drowning is a complex, and often overlooked public health issue, where global mortality estimates exceed 236 000 deaths annually, excluding drowning in disaster and transportation.1 2 Methodological advancements in population-representative data, growing evidence for effective interventions and an elevated focus on high burdens in low-income and middle-income countries (LMICs), have improved the visibility of the problem and coalesced action for drowning prevention.3

Drowning prevention’s rise in health and sustainable development discourse is evidenced by successive Resolutions, first at the United Nations General Assembly (2021),4 and then at the World Health Assembly (2023).5 Significant increases in drowning research,6 and the publication of reports by WHO,7–9 reinforce further growth. Yet, there is no universally agreed definition for drowning prevention.10

Significant ambiguity in the conceptualisations of drowning prevention across peer-reviewed and grey literature has been observed.6 10 The terms drowning prevention and water safety are used interchangeably or with different or overlapping meanings.10 Discordance around terminologies and a lack of consensus-based agreement for key terms, has the potential to hinder cohesive progress.3 6 Reaching a consensus on the definition of drowning prevention has the potential to advance the research agenda,11 foster stronger partnerships with stakeholders in other health and sustainable development initiatives,10 and facilitate the pursuit of strategic priorities through the formulation of a global strategy.12

Previous consensus-based processes have had significant benefits for advancing drowning prevention.3 For example, the development of a standardised definition for drowning in 2002,13 was adopted by WHO and led to increased consistency in reporting of the drowning burden.14 The Delphi method has been used to form consensus-based agreement on preventative messaging in recreational drowning prevention settings,15 and to develop agreement on prevention strategies for drowning in rivers.16

Here, we aim to develop a conceptual definition for drowning prevention and reach consensus-based agreement for its adoption using the Delphi method. We seek to develop shared terminology to strengthen multisectoral action for drowning prevention.

MethodsStudy design

We applied the Delphi method in a five-step process (figure 1). The Delphi method is a well-established approach to forming a consensus-based agreement across subject matter experts. The Delphi method has been applied to create conceptual definitions for major trauma,17 advance care planning18 and running-related injury.19

Figure 1

Five-step Delphi consensus-based agreement process.

Advisory group

We formed an advisory group (n=6) by recruiting members from the existing networks of both authors. Advisory group members held expertise in research, policy and practice, and had experience in a breadth of drowning prevention contexts including LMICs. The advisory group met twice (virtually) to provide input into study design, participant recruitment, survey development and analysis of results. The advisory group established that consensus-based agreement would be ≥70% of participants agreeing to the final draft definition or the proposed statements.

Participant recruitment

Participants were identified through document analysis, from the existing research, policy and practitioner networks of both authors, and referrals from the advisory group. International and national drowning prevention groups including International Life Saving Federation, Royal Life Saving Society Commonwealth, and Royal Life Saving Society—Australia, as well as The George Institute for Global Health were engaged to promote the study by posting advertisements on their social media accounts, or by sending an email invitation to potential participants.

The inclusion criteria were based on expertise: policy makers, researchers, technical specialists or community members, and work setting: UN agencies, governments, non-government organisations, foundations, academic institutions or from communities affected.

Potential participants reviewed an online information and consent form. When participants enrolled in the study and provided consent, they were sent a link to the first-round Delphi survey, which we deployed using online software REDCap.20

Preparation: identifying key concepts for inclusion

We used document review of peer-review and grey literature to identify concepts used to describe drowning prevention. The search of both peer-review and grey literature aimed to capture contexts and terminology for drowning prevention across research, policy and practice.

First, we searched four databases: MEDLINE (via PubMed), The Cochrane Library, Web of Science and Embase for peer review literature using the search terms “drowning,” “drowning MESH” and “near drowning” in titles, abstracts and keywords for the period 1 January 2005 to 31 December 2020.

Then, we searched four databases: WHO library- Iris, UNICEF website database, Google, Relief Web for grey literature, including global, regional, or national-level reports and planning documents using the search terms “drowning” and “drowning and water safety” for the period 1 January 2014 and 31 December 2020. The WHO library-Iris, UNICEF and Relief Web were chosen to capture grey documents in the UN system, and various sustainable development agendas. The search method and data extraction is reported in our scoping review that identified opportunities for multisectoral action for drowning prevention.10

We conducted content analysis, by first extracting descriptive text blocks from the literature and grey documents where the term ‘drowning prevention’ was used. We then assigned each block of text a code, and then categorised the codes into frames. Frames are defined as the central idea being presented in the text.21 The thematically extracted frames informed the initial approach to drafting the subsections of potential draft definition. These subsections and the overall approach were reviewed by both authors and the advisory group, before the drafts were again discussed and confirmed in preparation for the first-round.

First-round: participant selection of concepts for inclusion

We designed the first-round survey in two parts. In part one, we proposed five sections for a potential definition. Once participants made their choices for each section, their choices were combined and displayed in draft definition format to allow for further consideration (online supplemental file 1).

In part two of the survey, we assessed participant views on the relationship between drowning prevention and water safety. Participants were presented with the statement—drowning prevention and water safety are synonyms and can be used interchangeably and asked to indicate their view on a five-point Likert scale (strongly agree—agree—neutral—disagree—strongly disagree). Open text fields were provided to collect feedback for both parts.

Second-round: building consensus-based agreement

In the second-round Delphi survey, we presented a draft conceptual definition, which was codeveloped using the highest scoring concepts from each section in the first-round survey, and after considering participant and advisory group feedback. We presented three statements aiming to address the feedback collected about confusion between the terms drowning prevention and water safety.

Patient and public involvement statement

There was no public/patient involvement in this study. Participants will receive a link to access the results.

Role of the funding source

The funding source (Australian Government Research Training Programme Scholarship) played no role in the study.

Discussion

Our study adopted a consensus-based approach to developing a conceptual definition for drowning prevention and aimed to strengthen the definition by addressing perceived confusion between the terms drowning prevention and water safety. The definition presents a starting point for framing and reframing drowning prevention in ways to inspire action, expand coalitions and strengthen institutions for multisectoral action.3

The concepts within our definition: (1) multidisciplinary approach, (2) evidence-informed measures, (3) reducing risk, (4) building resilience and (5) addressing hazards, exposure, vulnerability, provide a framework for understanding, developing and strengthening drowning prevention efforts. These concepts may aid the identification of cobenefits in sectors and agendas not yet engaged in drowning prevention and assist the development of interventions in ways not previously considered.

Promoting multidisciplinary approaches

Multidisciplinary approaches combine academic and/or professional specialisations to solve a problem. Through the utilisation of a multidisciplinary approach, our definition emphasises the need to integrate diverse sectors and stakeholders each with their own contexts, expertise and approaches. Examples of multidisciplinary expertise engaged in drowning prevention include technical domains such as lifeguarding, emergency response management or education; academic fields such as epidemiology or engineering; management disciplines such as communications, advocacy, or administration, and regulators such as those in occupational health, maritime, and swimming pool safety. It is hoped that further synergistic efforts across disciplines can combine to increase comprehension, scope and efficiency of drowning prevention measures.

More than desirable, a multidisciplinary approach may be critical to the leadership and coordination of drowning prevention efforts. A review of drowning prevention planning shows that national leadership and coordination is not homogeneous, and varies across sectors and disciplines, in some examples including ministries of public health (Thailand), labour, invalids and social affairs (Viet Nam) and an alliance of non-government organisations (Australia)10 In some cases responsibility is fluid. For example, the Malaysian Water Activity Safety Committee was launched in 2015 under the Ministry for Local Government (discipline - public administration),8 and then moved to the Department of Fire and Rescue (discipline - emergency management) in 2022. Strengthening capacity for work in whole-of-society models, combining the resources, skills, and capabilities of each sector and stakeholder, whether government or non-government, is critical to successful multisectoral action and to advancing strategic priorities for drowning prevention.10 12

Expanding evidence-informed measures

Our study presents the concept evidence-informed to describe drowning prevention measures, although there was some consideration of using evidence-based instead. Evidence-based practice has high relevance in medical settings where clinical expertise, and the best available research evidence is combined to make healthcare decisions.22 On the other hand evidence-informed practice combines research evidence, clinical experience, professional judgement, participant preferences and contextual factors to make decisions,23 and is designed to meet changing conditions with flexibility when evidence is weak or non-existent.24

The evidence base for drowning prevention is emergent, with many interventions commonly accepted by policy makers and practitioners considered to lack robust evidence.25 The challenge is that while these suggested interventions may lack the rigour of a scientific trial, they align with established best practices, and are often feasibility tested in the field. In this regard, the concept of evidence-informed aligns to the notion of drowning prevention as a multidisciplinary endeavour and necessitating the incorporation of contextual knowledge and methodologies from diverse sectors. A future research agenda should prioritise studies that explore greater contextualisation of drowning prevention, and measure the efficiency and efficacy of interventions.12

Reframing drowning risk

Our definition reinforces the importance of drowning risk—the consequence of the interaction between a hazard, the characteristics that make people vulnerable to drowning, and their exposure (context, frequency of occurrence, duration) to a drowning hazard. The basis for this conceptualisation is used in disaster risk reduction where it underpins the Sendai Framework,26 and in the Intergovernmental Panel on Climate Change.27

Disaster risk reduction and climate resilience are two agendas prioritised for the integration of drowning prevention efforts.12 Reframing drowning prevention and aligning to the language and concepts used in disaster risk reduction and climate resilience sectors may assist that integration and help to facilitate the identification of cobenefits and complimentary interventions.

Exploring hazard, vulnerability and exposure

The management of drowning risk more often focuses on mitigating hazards in aquatic environments. The consideration of exposure and vulnerability to drowning hazards may improve the assessment of risk and provide context essential to designing effective interventions.

A drowning hazard is an environmental or human induced feature or process that contributes to drowning risk. Drowning hazards can be (1) natural features for example, oceans, rivers, lakes and beaches, (2) natural processes including tropical storms, floods, tsunami, heatwaves, currents, waves steep or slippery banks or (3) human induced processes and features in the built environment or flaws in technological design, for example, modes of transportation, swimming pools, bathtubs and structurally deficient dams.

There is extensive research exploring the dangers of drowning hazards including rip currents28 and backyard swimming pools.29 Further research investigating the impacts of hydrometeorological hazards such as flooding,30 storm surge and heatwaves,31 on drowning risk is much needed.

Drowning exposure is the context, occurrence and duration in which an individual, community or population has contact with a specific drowning hazard. Exposure can be influenced by social, economic and environmental factors, and can occur in occupational, recreational or everyday living contexts. Calculating drowning risk based on exposure to hazards could supplement research based on population-based drowning estimates.32 Exposure is a concept important for monitoring adverse outcomes. A study found that learning to swim did not increase exposure or risk taking in Bangladeshi children.33 While exposure provides a basis for targeting interventions, high rates of exposure alone does not mean a person is at increased risk of drowning. For detailed assessment of drowning risk, exposure must be considered alongside vulnerability.

Drowning vulnerability is the susceptibility of an individual, community or population to drowning. Here, we use vulnerability as a dynamic concept that interacts with context, rather than as a label (vulnerable) applied to an individual or group.34 Vulnerability is best measured in layers and can be based on a range of physical, social, economic and environmental factors.

We caution against generalised use of the term vulnerable to apply to large cohorts without consideration of the individual traits that may contribute to vulnerability. Individual vulnerabilities include poor or insufficient swimming skills that might contribute to increased risk in a particular environment or water activity,35 medical conditions including heart disease,36 epilepsy37 and autism,38 and excessive alcohol or illicit drug use,39 which may increase susceptibility to drowning when a person is exposed to water.

Research investigating how social determinants increase exposure and vulnerabilities to drowning, including the impacts of where a person is born, the features of their neighbourhood and built environment in which they live, what access they have to education, safe working conditions and health services, is a high priority for the advancement of drowning prevention.12 A detailed analysis of the role of exposure and vulnerability may help in the targeting of prevention measures. For example, rock fisher drowning in Australia is an activity where participants are exposed to multiple hazards (rocks, waves, tidal changes), and where mortality data shows over representation of specific migrant populations—some of whom are thought to lack swimming skills, adding layers of vulnerability in the event of a fall into the wave zone.32 Studies have called for interventions that go beyond the delivery of translated messages, and address specific exposures and vulnerabilities of identified rock fisher populations.32 40 41 Those studies recommended direct advocacy via migrant community groups, provision of safety equipment to participants to reduce vulnerability in the event of a fall and increasing awareness of local weather forecasting when conditions are likely to increase exposure to hazards.

Strengthening resilience for drowning prevention

Resilience is a concept used frequently in disaster risk reduction and health system contexts. Disaster resilience is the ability of a system, community or society to resist, absorb, accommodate and recover from the effects of a hazard in a timely and efficient manner.26 Health system resilience tends to focus on preparations for health shocks, by strengthening the policies, planning, health workforce and surveillance needed to minimise the consequences of such disruptions.42 Community resilience is described as a community’s intrinsic capacity to resist and recover from a disturbance, where communication, risk awareness, adaptation and risk governance are important elements.43 Community resilience is often measured across social, economic, institutional, physical and natural domains.27

In our definition, drowning prevention resilience means the adaptive and robust capacity of an individual, community or system to avoid, resist or respond to drowning risk, including pre and post event. For example, strengthening individual-level drowning prevention resilience includes the development of swimming, water safety and basic rescue skills in school children, parental education on supervision requirements around water, and public awareness of lifejacket laws. Building community-level drowning prevention resilience includes increasing the availability of designated and supervised places to swim, enhancing local beach lifeguard services and ensuring causeways have appropriate signage or gates in times of forecast flooding. Reinforcing system-level drowning prevention resilience includes upgrading national pool fencing legislation, monitoring safety standards for waterways and swimming pools, and strengthening of a drowning data surveillance system.

Using the conceptual definition

The conceptual definition has many potential applications. Both researchers and safety practitioners can expand risk models to measure exposure, vulnerability, and hazards in calculating, reporting and addressing drowning risk. Policy makers can embrace multidisciplinary approaches in intervention planning and adopt whole-of-society coalitions for drowning prevention. Educators can reframe and redesign drowning prevention programmes to consider the concept of resilience. The use of the conceptual definition should be monitored in peer-review literature and grey documents. Adoption by WHO, the Global Alliance for Drowning Prevention,12 and international and national drowning prevention groups may increase use, and lead to further refinement and strengthening of the conceptual definition.

Developing an operational definition

There may be utility in operationalising the definition for use in research and evaluation. An operational definition sets indicators to measure the degree to which the definition threshold is achieved. While not an objective of this study, indicators in an operational definition could include the breakdown of hazards, exposures and vulnerabilities into variables. Such a framework could enhance evaluation. For example, a programme or policy targeting a hazard without considering the role of exposure, or vulnerability could be improved if all three factors were addressed.

Limitations

Our document review may have missed relevant studies not specifically mentioning ‘drowning’ in title, abstracts or key words. A more comprehensive search of grey literature and non-indexed journals may have identified additional data relevant to defining drowning prevention.

Our recruitment methods are a strength and a weakness. We sought to recruit participants that reflected diverse expertise, as studies show drowning prevention is not homogeneous, and involves various sectors.3 10 It is possible that the definition, while developed by consensus-based agreement, does not completely resonate with a specific sector. For example, leisure and education sector practitioners might place greater emphasis on individual knowledge or skill development.

The Delphi method captured widespread input, however, the process of anonymising feedback meant some points raised by participants could not easily be explored or progressed to develop deeper insight. Anonymising participants meant we were unable to identify representation from specific organisations, governments or academic centres for follow-up.

Our study was conducted in English, which may have impacted on non-native English speaker participation. Although they completed the surveys, 24.6% of our participants indicated that English was not their preferred language. Preferred languages included Bangla, (4.6%) French (3.7%), a Chinese language (3.0%), Portuguese (3.0%) or Spanish (3.0%). Translating and evaluating the definition in languages other than English may be beneficial.

Acknowledgments

Authors thank the Advisory Group members; Dr David Meddings (WHO, Switzerland), Kelly Larson (Bloomberg Philanthropies, United States), William Koon (School of Biological Earth and Environmental Sciences, University of New South Wales, Australia), Prof Emmanuel Kweku Naku, (Ghana), Gemma May (Royal National Lifeboat Institution, UK), and David Walker (Royal Society for the Prevention of Accidents, UK). We also thank the organisational that agreed to promote our study, including Royal Life Saving Society—Australia, International Life Saving Federation, Royal Life Saving Society Commonwealth, the Centre for Injury Prevention and Research, Bangladesh, and the George Institute for Global Health. We thank all 134 participants who voluntarily and openly contributed their time and expertise to our study.