Epidemiology, Risk Factors and Measures for Preventing Drowning in Africa: A Systematic Review

Background and Objectives: Drowning is a leading cause of unintentional injury related mortality worldwide, and accounts for roughly 320,000 deaths yearly. Over 90% of these deaths occur in low- and middle-income countries with inadequate prevention measures. The highest rates of drowning are observed in Africa. The aim of this review is to describe the epidemiology of drowning and identify the risk factors and strategies for prevention of drowning in Africa. Materials and Methods: A review of multiple databases (MEDLINE, CINAHL, PsycINFO, Scopus and Emcare) was conducted from inception of the databases to the 1st of April 2019 to identify studies investigating drowning in Africa. The preferred reporting items for systematic review and meta-analysis (PRISMA) was utilised. Results: Forty-two articles from 15 countries were included. Twelve articles explored drowning, while in 30 articles, drowning was reported as part of a wider study. The data sources were coronial, central registry, hospital record, sea rescue and self-generated data. Measures used to describe drowning were proportions and rates. There was a huge variation in the proportion and incidence rate of drowning reported by the studies included in the review. The potential risk factors for drowning included young age, male gender, ethnicity, alcohol, access to bodies of water, age and carrying capacity of the boat, weather and summer season. No study evaluated prevention strategies, however, strategies proposed were education, increased supervision and community awareness. Conclusions: There is a need to address the high rate of drowning in Africa. Good epidemiological studies across all African countries are needed to describe the patterns of drowning and understand risk factors. Further research is needed to investigate the risk factors and to evaluate prevention strategies.


Introduction
The World Health Organization (WHO) defines drowning as "the process of experiencing respiratory impairment from either immersion or submersion in liquid" [1]. Drowning is the third leading cause of unintentional injury related cause of mortality worldwide, accounting for 7% of all injury related deaths. It is a global under recognized and neglected public health burden that claims the lives of 320,000 people every year [2]. More than 90% of these deaths occur in low-and middle-income countries with inadequate prevention measures [3]. It is among the ten leading causes of deaths in children and young people in the world with children aged less than five years at increased risk [4]. Between 1990 and 2013, drowning rates declined by 52.2% globally [5], however, despite this decline, the highest rates of drowning were observed in Africa [3].

Literature Search
The systematic review was conducted in accordance to the preferred reporting items for systematic review and meta-analysis guidelines (PRISMA) [18]. The PRISMA flow chart for the review is shown in Supplementary Figure S1. A literature search was conducted using Ovid Medline, Emcare, Cumulative Index to Nursing and Allied Health (CINAHL), PsycINFO, and Scopus for original research articles published in English from inception until the 30th of November 2018. The search was updated on the 1st April 2019. We included all articles focusing on drowning in Africa. There were slight variations in the search terms depending on the database. Search terms involved a combination of free text words and Medical Subject Headings (MeSH) terms. General search terms were "drown*" and "Africa". The search strategy for Medline is shown in Supplementary Table S1. The study protocol was registered in PROSPERO with registration number CRD42019092758.

Eligibility Criteria
The studies included in this review are published original research reporting drowning in African countries. We applied no limits to the year of publication and included all age groups. In addition, we included studies that reported drowning as part of other injuries studies to capture all data from the region. Studies excluded were review articles, drowning because of suicide or homicide, non-fatal drowning or near drowning or hospitalization due to drowning or where fatal drowning could not be distinguished from non-fatal drowning.

Data Extraction
Faith O. Alele (F.O.A.) and Theophilus I. Emeto (T.I.E.) identified all included studies from the search strategy. Uncertainties about the included studies was discussed until consensus was reached. FOA and TIE extracted general and study specific characteristics from the included studies and Lauren Miller (L.M.) and Richard C. Franklin (R.C.F.) crosschecked the data.

Quality of Methods Assessment
The methodological quality of the included studies was assessed by FOA and TIE using the modified quality assessment tool for studies with diverse designs (QATSDD) critical appraisal tool [19]. The tool assesses the validity, reliability and generalizability of studies. The included studies were a mix of cross-sectional, descriptive and case-control studies and each study design were assessed using the appraisal tool. The tool was modified to exclude two items that were not applicable to the included studies. The excluded items comprised of statistical assessment of reliability and validity of measurement tool(s) (Quantitative only), fit between stated research question and format and content of data collection tool e.g., interview schedule (Qualitative), assessment of reliability of analytical process (Qualitative only) and evidence of user involvement in design. In the modified QATSDD tool each criterion was awarded a score of 0 to 3 with 0 = not at all, 1 = very slightly, 2 = moderately and 3 = complete. The scores of the criteria were summed up to assess the methodological quality of included studies with a maximum score of 36. For ease of interpretation, the scores were converted to percentages and were categorised as excellent (>80%), good (50-80%) and low (<50%) quality of evidence based on the overall score (Supplementary Table S2).

Data Synthesis
Drowning was reported exclusively or as part of a wider study such as injury studies. Approximately 28% (11) of the included studies reported drowning as unintentional. In studies where drowning was unspecified, we reported the drowning as intentional. Measures used to report drowning were proportions and incidence rates. The incidence rates and proportions of drowning were reported using frequency tables. The risk factors for drowning were identified in two articles, one of which only reported drowning as part of a wider study. Therefore, the risk factors identified were extrapolated to drowning. However, given the paucity of information on risk factors associated with drowning, we identified the potential risk factors based on previously identified factors documented in the literature [3,4,8] and based on the reported rates of the potential factors. A meta-analysis was not conducted due to the heterogeneity of the included studies.

Drowning Rates in Africa
In Table 1, twelve (12) investigated drowning exclusively in different regions of Africa [21-27, 30,43,50,54,60]. However, there were variable methods of reporting drowning across the different studies. Among population-based studies, the proportion of drowning fatalities ranged from 0.019% to 1.2% [23,26]. In studies where all submersion events were reported, unintentional drowning accounted for 80% of drowning deaths in one study [50], while accidental drowning accounted for 10.7% of all submersion (near drowning and drowning) events in another study [30].
The incidence rates of drowning across the different studies ranged from a low of 0.33/100,000 population to a high of 502/100,000 population [21,22,24,25,27,43,54]. However, the denominators for each study varied. Two studies were conducted using the total population in the country as the denominator [21,22], five studies were conducted within specific cities and towns and total population of the cities were used as denominators [23,24,27,43,54], while one study investigated drowning across five cities [25].

Prevention Strategies
Sixteen (16) studies proposed prevention strategies to reduce drowning rates in Africa [23-27, 30,31,42,44-46,48,53,54,59,60]. These prevention strategies include increased supervision of children around bodies of water, aquatic education and training about basic life support measures, training about life skills in communities, community awareness and implementation of legislation to prevent drowning (Table 4). Using the hierarchy of controls [64], fourteen of the sixteen studies proposed administrative control/preventive measures, which included education/training on basic life support, legislative laws and increasing public awareness [23,25,27,30,31,42,[44][45][46]48,53,54,59,60]. Two studies proposed engineering control measure that include building life safety facilities and the use of barriers and safety nets around swimming pools [24,26].

Discussion
Drowning is a significant public health burden in Africa and the findings of this systematic review suggest that there is a huge variation in drowning mortality across Africa. The highest proportion of drowning (approximately 80%) was reported in Nigeria [50], while the highest rate reported (502/100,000 population) was observed in Uganda [54]. Although only two studies identified risk factors which includes being a fisherman, and older age [48,58]; we identified potential risk factors based on previous evidence [3]. The limited evidence suggests that male gender and young people are at higher risk for drowning especially children and adolescents. In addition, other potential risk factors identified were being of black African ethnicity, alcohol use, access to bodies of water, age of boat and carrying capacity of the boat, weather and summer season. This systematic review has highlighted the need for more data on drowning prevalence, together with good epidemiological studies across all African countries to describe the patterns of drowning and understand risk factors to guide prevention initiatives.
Due to the limited data available, quantifying the prevalence of drowning in Africa was challenging. This finding was echoed in a study of drowning in low-and middle-income countries by Tyler et al. who reported that inconsistences in data collection for drowning poses as a challenge for data synthesis [8]. Although the estimated rates and proportions may be considered high, in majority of the studies, drowning was reported as part of a wider injury study. Of the 54 countries in Africa, only 15 countries had some published data on drowning with the majority (57%) of the literature originating from South Africa. In many African countries, cases go unreported and the lack of an injury surveillance system as seen in many LMIC also contributes to the limited data [65]. This is consistent with the findings of two recent systematic reviews describing the burden of drowning in South Africa and Tanzania [16,17]. Saunders et al. and Sarrassat et al., reported that strengthening the existing surveillance systems or establishing new ones are needed for consistent and detailed drowning surveillance [16,17]. In addition, as many of the drowning cases result in death at the time of the event, only a small proportion present at the hospital or medical facilities [66]. Both the lack of the injury surveillance system and underreporting of drowning cases prevent accurate documentation of drowning mortality in health records. According to WHO, approximately 90% of global drowning deaths occur in LMICs. Africa as a region had an estimated 73,635 drowning deaths in 2016 which accounted for approximately 23% of the total drowning deaths globally [67]. However, data collection in the region is limited, and hence the statistics from Africa underrepresents the true burden of drowning in the region [4].
Many of the potential risk factors associated with drowning identified in this review are similar to those reported in previous systematic reviews. Drowning occurred more frequently in males between the ages of 0 to 15 years. Specifically, highest drowning occurrences were found in the 0-5-year age group. In LMICs, drowning rates among children were higher among children aged 1-4 years, followed by children aged 5-9 years with males being twice as likely as females to drown [8]. In addition, younger children were found to drown in private pools or baths, whereas older children were found to drown in public swimming pools, rivers, dams or in the ocean. Given that a majority of the studies originated from South Africa, drowning in swimming pools occurred more in white African children, whereas drowning in dams and rivers were found to occur in older black African children [24,31]. Although it is not evident that socioeconomic status is a potential risk factor in this review, children from low-income households may not have access to private swimming pools and are more likely to access natural bodies of water around the house as reported in other LMICs [8]. Evidence suggests that a lack of child supervision and the lack of safety barriers has been associated with high drowning rates among children [66]. Furthermore, access to other water bodies using boats or through fishing, depending on occupational or recreational purposes was also considered as a potential risk factor [48,54]. Specifically, children and adults from fishing households, are more likely to access these types of water bodies daily, requiring significant surveillance and awareness strategies for children in such settings. Other potential risk factors such as alcohol consumption has been shown to be associated with drowning especially among adolescents and adult [68]. As blood alcohol concentration level rises, judgement, balance and vision may be impaired, increasing the risk of drowning. Binge drinking is common in some African countries and has been reported to be associated with drowning among adult men [23][24][25][26]50]. This calls for increased awareness of the risk of alcohol consumption in conjunction with swimming.
The prevention strategies proposed by sixteen studies includes focusing on pool safety such as restricting access to private pools for young children, education and training at schools on life skills, increasing public awareness through media campaigns, and the implementation of water safety legislation, community awareness, improved supervision of children around water bodies, building lifesaving facilities and enforcement of boat construction and maintenance regulations. Using the hierarchy of controls which is a system used to minimize or eliminate hazards [64], only two studies proposed engineering controls as a way of preventing drowning [24,26]. All other studies proposed prevention strategies that require administrative controls which is the least effective way to prevent drowning [23,25,27,30,31,42,[44][45][46]48,53,54,59,60]. However, prevention interventions and methods may not be consistent between countries due to the diversity and variation in their epidemiology, demographic and cultural characteristics [8]. There is no simple solution to addressing the burden of drowning in all countries, therefore strategies would have to be designed specifically for each country, keeping in mind the cultural, economic and social structures.

Implications for Policy and Future Research
The findings of this review suggest that there is limited evidence and data on the burden of drowning in Africa. The 2017 Global Burden of Disease (GBD) using statistical models estimated that drowning contributed approximately 0.53% of the total deaths in Africa as a region. Using the GBD to obtain country specific estimates for the countries included in the review showed an average drowning mortality ranging from 1.63 per 100,000 population to 5.73 per 100,000 population [69]. However, existing data on drowning in many countries in Africa are scarce. Therefore, establishing specific databases about injuries like drowning for surveillance and data collection would aid in development of policies and prevention strategies across the different countries. Evidence from research conducted in high income countries like Australia, Canada and New Zealand suggest that robust high-quality data and better data collection system would enable the creation of targeted and effective drowning prevention interventions [70]. Developing the databases will enable cross-country comparison which allows for identification of similarities and improvements in data collection. However, establishing the databases may be challenging for some African countries, especially if drowning is not among the national health priorities. In addition, there was little exploration of the risk factors associated with drowning, highlighting a gap in the literature. Good epidemiological studies are needed to identify the risk factors and evaluate the proposed prevention strategies for drowning in Africa. Furthermore, future research should focus on the intent for drowning in Africa, which would help to inform policies and prevention interventions. A recent article on intentional drowning reported an increasing rate of intentional drowning and proposed a multidisciplinary collaboration public health and other services including mental health, education and drowning prevention organisations to prevent intentional drowning [71].

Strengths and Limitations
To our knowledge, this is the first systematic review that describes the epidemiology, risk factors and prevention strategies of drowning in Africa. However, comparing mortality data across the countries within Africa needs to be undertaken with caution given the different measures used to analyse the burden of drowning. Some studies were population-based studies, while other studies reported drowning as a part of a wider study (such as external causes of deaths or all causes of death). An example of the latter is the study by Seleye-Fubara et al. which reported that unintentional drowning accounted for approximately 80% of all drowning deaths [50]. In addition, the completeness and reliability of the data in each country varied with some studies using the national mortality surveillance statistics, while other studies relied on hospital-based data, mortuary-based data and demographic surveillance data. The variability in the sources of data may account for the variable rates of drowning reported in the review.
Other limitations include reviewing only articles published in peer-reviewed journals. We may have missed other high-quality studies that are published in non-peer reviewed journals. In addition, we excluded non-English articles, there is a possibility that we may have missed articles from Africa published in other languages. Furthermore, the majority of studies were from South Africa, it is conspicuous that there is a paucity of data from many countries in Africa. It is uncertain whether such strategies, if implemented, can be generalizable to other African countries besides South Africa.

Conclusions
There is a need to address the high rate of drowning in Africa. It is imperative that governments across the nations of Africa establish good injury surveillance systems to accurately understand the burden of drowning to inform approaches for drowning prevention. Good epidemiological studies across all African countries are needed to describe the patterns of drowning and understand risk factors. Further research is needed to investigate the risk factors and to evaluate prevention strategies.
Author Contributions: All authors contributed equally to the study concept, design, data extraction, quality assessment and writing of the paper. All authors read and approved the manuscript for submission.