Background
Methods
Inclusion | Exclusion |
---|---|
Clinical treatment and outcomes of female adults treated in the emergency department (ED) and hospitalised for drowning | Studies where adults and children could not be separated in the data |
Epidemiology (age and proportion by gender and sex) and risk factors for unintentional female fatal and non-fatal drowning (were included if they added value to our understanding) | Case reports, government reports, conference abstracts, literature reviews, editorials, policy statements, letters |
Adults (18 years or older) | Children (less than 18 years of age) |
Study locations: Australia, New Zealand, United States, United Kingdom and Canada | Countries outside the chosen study locations |
Published in English in peer-reviewed journal | Published in a language other than English and/or not peer-reviewed |
Date: January 2003 to April 2019 | Date: Studies published before January 2003 |
Author | Country | Study aim | Study design/methodology | Study sample | Results | Risk factor | Evidence hierarchy (34)/GRADE [35] | Limitations |
---|---|---|---|---|---|---|---|---|
Peden et al. (2016) [14] | Australia | Unintentional drowning deaths in rivers, creeks and streams | Design: Cross-sectional study Data collection: Australian National Coronial Information System, July 2002 to June 2012. | Drowning deaths: Females n=151/770 (19.6%) | Females 2.27 times more likely to drown in rivers in non-aquatic transport, and 4.45 times more likely to drown being swept away by floodwaters. | Activity: non-aquatic transport, being swept away | Level IV-Low level evidence GRADE-Very low | Coroner cases with open findings may overestimate number of deaths from river drownings. Some variables had limited information due to open finding. |
Gulliver & Begg (2005) [41] | New Zealand | To describe water-related behaviour and non-fatal drowning incidents among young adults | Design: Cross-sectional study Data collection: Longitudinal study, Dunedin Multidisciplinary Health and Development Study, 1993 to 1994. | Non-fatal drowning incidents: Females n=52/141 (37%) | Females reported lower rates of water confidence, risk exposure and experience of a non-fatal drowning incident. Females were higher risk of drowning within two hours of consuming alcohol and engaging in watercraft activity. | Water confidence, exposure, alcohol and watercraft | Level IV-Low level evidence GRADE-Very low | Participants were asked about alcohol use and engagement in water activity, although the amount of alcohol consumption was not recorded. |
Henderson & Wilson (2006) [42] | United Kingdom | To examine hospital admissions from a water-related incident. | Design: Cross-sectional study Data collection: HES England using ICD-10 coding, 1997 to 2004. | Females n=1787/6793 (26.3%) | 1 fatal drowning=3 hospital presentations. Numbers of females increased in the ICD-10 codes for: drowning and submersion in bathtub and following fall into bathtub, and victim of flood. | Submersion in bathtub, fall into bathtub and victim of flood | Level IV-Low level evidence GRADE-Very low | Personal and environmental factors prior to drowning incident were not captured in this study. |
Hudson et al. (2006) [43] | United States | To examine factors associated with injuries occurring in drowning incidents among hospitalised patients in Alaska | Design: Cross-sectional study Data collection: State of Alaska Department of Public Health Alaska Trauma Registry, 1991 to 2000. | Immersion only: females n=25/89 (28%) Associated injuries: Females n=17/87 (20%) | Females associated with increased hospitalisations from a drowning incident (p=0.02). Females were found to be at higher risk of an associated injury than males. | High risk of injury from drowning | Level IV-Low level evidence GRADE-Very low | The higher risk of associated injury from immersion-related event among females may be due to Alaskan males being more experienced in working and recreationally interacting with the water. |
Morgan et al. (2009) [44] | Australia | To explore self-reported water exposure, activity, protective factors and drowning risk at surf beaches by gender. | Design: Prospective cross-sectional study Data collection: Self-reported survey from December 2003 to February 2004 | Females n=210/406 (51.7%) | Females visited surf beaches less than males, but length of stay was similar. Number of males and females engaged in wave swimming were similar (females n=80, males n=88). | Exposure, risk-taking behaviour | Level IV-Low level evidence GRADE-Very low | Sampling method may have been biased, as a person may have visited the surf beach more than once during data collection (16 days). Small sample size in sub-group for surfing (females n=14, males n=79). |
Nasrullah & Muazzam (2011) [45] | United States | To describe demographics and changes in unintentional drowning mortality | Design: Cross-sectional study Data collection: Centers for Disease Control and Prevention Web-based Injury Statistics Query and Reporting System mortality data. 1999 to 2006 | Females n=5846/27514 (21.2%) | Increase of 9.5% of drowning fatalities for females, male drowning fatalities decreased by 0.7%. Female drowning fatalities increased by 7% in 2004 to 2005 | Fatal drowning | Level IV-Low level evidence GRADE-Very low | There may be errors or missing data on death certificates that could lead to an incorrect classification of injury. |
Peden et al. (2018) [46] | Australia | To explore river use and alcohol consumption and attitudes towards drowning risk | Design: Prospective Cross-sectional study Data Collection: Convenience sample, survey conducted in four river locations across Australia | Females n=353/684 (51.6%) | Females visiting rivers in similar number to males but females engaging in non-aquatic activities. Higher number of females with positive blood concentration than males | Exposure, alcohol, risk-taking behaviour | Level IV-Low level evidence GRADE-Vey low | Possibility of recall bias due to self-reporting survey. Random convenience sampling used, and refusal rate was not documented. |
Peden et al. (2018) [47] | Australia | To compare fatal and non-fatal drowning databases in Australia to identify key ratios, differences and inform drowning prevention strategies. | Design: Retrospective Cross-sectional study Data collection: Royal Life Saving Society Australia National Fatal Drowning Database and Australian Institute of Health and Welfare National Hospital Morbidity Database, July 2002 to June 2015. | Fatalities (all ages): Females n=525/2272 (23.1%) Hospital separations (all ages): Females n=2088/6158 (33.9%) | Unintentional fatal drowning: Adult females n=353/1737 (20.3%) Hospitalisations for drowning: Adult females n=687/2585 (26.6%) | Non-fatal drowning | Level IV-Low level evidence GRADE-Very low | An error could have occurred during in the prediction of fatal drownings, a correction factor was applied to the number of fatal drowning incidents and to predict the number of non-fatal drowning incidents. ICD coding for drowning data was limited to the codes W65-74 as primary cause of hospitalisation for non-fatal drowning incidents. |
Clemens et al. (2016) [48] | Canada | To describe the characteristics of drowning fatalities by age | Design: Retrospective descriptive analysis Data collection: Coroner’s reports, hospital data, police reports and death certificates, January 2008 to December 2012. | Females n=424/2391 (17.7%) | Drowning incidents: adult females: 20-34 years: n=73/592 (12.3%), 35-64 years: n=179/1033 (17.3%), 65 + years: n=93/419 (22.2%) Female adult unintentional drowning deaths*: 20-34 years: (0.42), 35-64 years: (0.5), 65 + years: (0.69) | Age | Level IV-Low level evidence GRADE-Very low | Risk of information bias due to data recorded by individual data collectors. Some data missing |
Author | Country | Study aim | Study Design/Methodology | Study sample | Results | Limitations | |
---|---|---|---|---|---|---|---|
El Sibai et al. (2018) [4] | United States | To describe the characteristics and predictors of poor outcome among ED presentations for drowning | Design: Retrospective cross-sectional study Data Collection: NEDS dataset, 2013 | Females n = 4283/12529 (34.2%) | Multivariate analysis results report significant positive predictor of poor outcomes among males. Over half of patients were discharged home from ED. | Level IV-Low level evidence GRADE-Very low | NEDS data represents 20% of US hospital based ED’s. Errors reported in injury coding for submersion therefore some drownings missed. |
Lee et al. (2006) [36] | United States | To describe the epidemiology and outcomes of serious paediatric submersion, and identify factors associated with increased mortality and morbidity. | Design: Retrospective case-series Data collection: Massachusetts electronic death database from Department of Public Health Registry of Vital Records and Statistics, and Massachusetts Hospital discharge Database from 1994 to 2000. Ages 0–19 years | Females n = 89/267 (33%) | Males were 2.52 times more likely to have a poor outcome than females (mortality and morbidity) | Level IV-Low level evidence GRADE-Very low | Limitations in data collected from databases such as quality, duplicates or missed cases. Data only contained patients admitted to hospital wards as ED state-wide data was not available. |
Quan et al. (2014) [37] | United States, United Kingdom, Australia | To assess the association between water temperature and duration of submersion in the outcome of drowning. | Design: Case-control study Data collection: Western Washington database, January 1974 to June 1996. Data collection was limited to unintentional drowning that occurred in open waters (lakes, rivers and oceans) | Females n = 161/1094 (15%) | Good outcome (survived with or without limited neurological deficit or injury): Females n = 69/276 (24.9%) Good outcome associated with age less than 15 years, female and immersion duration of less than 6 min in water greater than 16 degrees Celsius. | Level III-3-Low level evidence GRADE-Very low | Retrospective data collection limitations with hand searching, although authors believe high case ascertainment from this method. Water temperatures were estimated, as not routinely collected at the time of the drowning incident. |
Reynolds et al. (2017) [38] | United States | To estimate long-term mortality and identify prognostic factors in drowning victims. | Design: Cohort study Data collection: Western Washing Drowning Registry on non-fatal drowning incidents from January 1974 to July 1996. | Females n = 247/776 (31.8%) | Long-term mortality: univariate analysis for Male sex –non-significant (Cox proportional hazard modelling). | Level III-3-Low level evidence GRADE-Very low | Difficult to predict long-term survival from drowning event due to accuracy of coding data, although this association suggested. |
Reynolds et al. (2019) [39] | United States | To estimate long-term survival after cardiac arrest from drowning. | Design: Cohort Study Data collection: Western Washington Drowning Registry on non-fatal drowning incidents. January 1974 to July 1996. | Females n = 109/407 (26.8%) (Survived to hospital admission) | Long-term survival: Females n = 18/54 (33.3%) | Level III-3-Low level evidence GRADE-Very low | Missing data, only variables with less than 25% missing data were included in multivariate analysis. |