Context. Drowning is the second leading cause of unintentional injury death among children ages 1 to 19 years. Details about the specific site of submersion are important for implementation of primary prevention efforts, but are not routinely available from national data.
Objectives. To provide national data about the specific sites of drowning among US children and to examine site-specific drowning rates by age, race, and gender.
Setting. United States, 1995.
Design. Information was abstracted from death certificates for unintentional drownings among children <20 years of age. The site of drowning was coded from free text on death certificates. Proportional distributions of the site of drowning by age, region, gender, and race were examined. Site-specific drowning rates were calculated by age, race, and gender. Denominators were based on US Census Bureau, June 1995 estimates of the US population.
Results. For 1995, death certificates were obtained for 1420 unintentional drownings among children <20 years of age, 98% of the number of drowning deaths reported by the National Center for Health Statistics for the same year. Site of drowning was specified on 1308 (92%) death certificates. Fifty-five percent of infant drownings were in bathtubs. Among children between the ages of 1 to 4 years, 56% of drownings were in artificial pools and 26% were in other bodies of freshwater. Among older children, 63% of drownings were in natural bodies of freshwater. Site-specific drowning rates varied by race. Importantly, after the age of 5 years, the risk of drowning in a swimming pool was greater among black males compared with white males with rate ratios of 15.1 (95% confidence interval: 6.7–38.5) among 10- to 14-year-olds and 12.8 (95% confidence interval: 6.5–26.9) among 15- to 19-year-olds.
Conclusion. The variety of sites in which children drown, even within specific age groups, emphasizes the need for a multifaceted approach to prevention. Reductions in the relatively high rates of drowning among black adolescent males will require targeted interventions to prevent swimming pool drownings among this group.
Drowning is an important and preventable cause of death in childhood. In 1998, >1500 US children under 20 years of age drowned.1 Most of these deaths (93%) were unintentional and were not related to boating. Groups at increased risk of drowning include toddlers and adolescent males, with particularly high rates among black adolescent males.2–4 Factors that place black males at increased risk of drowning are not well understood, attributable at least in part to lack of information about the circumstances surrounding submersion events.
Information about the bodies of water in which children drown might help explain racial disparities in drowning rates, as well as aid in the design of appropriately targeted preventive strategies. However, there are no national studies in the United States describing sites of childhood drownings. This is primarily attributable to the lack of specificity of drowning codes in the International Classification of Diseases-Ninth Revision (ICD-9), the system used to code deaths that occurred before 1999.5 The bathtub is the only body of water for which there is a specific ICD-9 code. In 1998, bathtub drownings comprised >50% of unintentional drownings among infants; however, after the age of 1 year, <10% of drownings occurred in bathtubs.1
A number of regional studies have examined the most common sites of drowning by age. Studies in California, Texas, and Washington found that younger children were most likely to drown in swimming pools, whereas older children were most likely to drown in natural bodies of freshwater such as rivers and lakes.6–10 In contrast, most drownings in Minnesota among children <5 years old seemed to occur in natural bodies of water, whereas in Pinnelas county, a coastal community in Florida, after the age of 11 years drownings were most likely to occur in salt water.11,,12
The purpose of the current study was to provide national data about the most common sites of drowning among US children and to examine site-specific drowning rates by age, gender, and race.
The free text on hard copies of death certificates often provides details about injury events and, for drownings, may include information about the body of water in which the submersion occurred. However, these data are not routinely coded or available in national data sets. As detailed below, information abstracted from hard copies of death certificates was used to describe the epidemiology of childhood drownings.
As part of ongoing injury surveillance activities, the Consumer Product Safety Commission (CPSC) has arrangements with each of the 50 states and 2 major health jurisdictions (New York City and Washington, DC) for the purchase of specific death certificates where the death is likely to involve a consumer product. For drownings, this includes only those submersions that occur at a home, public building, residential institution, or place of sports/recreation. If staff at the CPSC determine that the death is product-related, detailed information, including the complete narrative describing the event, is entered into the CPSC death certificate data file. If no consumer product is involved, only the external cause of injury code, date, and state of death are entered into the database.
To obtain detailed data on all childhood drownings, the above procedures were modified. State contracts were modified so that death certificates for all nonboat-related childhood drownings were sent to the CPSC. At the CPSC, procedures were also modified so that detailed data were abstracted from death certificates for all drownings, both product-related and nonproduct-related, among children and adolescents <20 years old. Although data were collected for drownings that occurred between 1993 and 1995, data were most complete for 1995; thus, we restricted the analysis to that year.
Information from death certificates was abstracted and keyed by trained coders. Variables abstracted from the death certificate included age on the date of death, race and gender of the decedent, date of the event, date of death, city and state of death, general place of injury (eg, home, farm, street, school, etc), external cause of injury code, and the complete narrative describing how and where the injury occurred. If additional information about the site of drowning appeared on the death certificate in fields outside of the narrative field, this information was also entered. The age at the time of submersion was estimated from the age at death, the date of the event, and the date of death. Only 16 deaths occurred ≥30 days after submersion.
Classification of Site of Drowning
The site of drowning was coded by 1 of the investigators (R.A.B.) and was based on documentation in the narrative, place of injury, or other fields on the death certificate, as described above. Specific drowning sites (eg, lagoon, pond, swimming pool, etc) were further grouped into 4 broad categories: artificial pools, other natural or manmade bodies of freshwater (referred to hereafter as freshwater), domestic sites (eg, bathtubs and buckets), and salt water. For 31 deaths, the site of drowning was determined by examining both the narrative field and the city of death. For example, 14 deaths reportedly occurred at a “beach.” By looking at the location of the city in relation to the closest known body of water, it was concluded that 3 of these deaths were in rivers, 3 in lakes, and 8 in oceans.
To estimate the completeness of ascertainment, CPSC data were compared with data reported by the National Center for Health Statistics (NCHS). For 1995, 1447 unintentional, nonboat-related drownings among children <20 years of age were reported by NCHS (ICD-9 codes E910.0 through E910.9).13 For the same year, 1420 death certificates for unintentional childhood drownings were submitted to the CPSC. The 1420 deaths in the CPSC dataset were comparable to the 1447 deaths reported by NCHS with respect to age, gender, race, and state of death (P > .05).
Using CPSC data, proportional distributions of the site of drowning by age, race, gender, and region of the country were examined. Differences in proportional distributions were evaluated using the χ2 statistic. Overall and site-specific drowning rates were calculated by age, gender, and race and are reported as deaths per 100 000 person-years. Numerators were based on the CPSC data, and denominators were based on US Census Bureau, June 1995 estimates of the US population.14 Confidence intervals (CIs) around rate ratios (RR) were based on a Poisson distribution and calculated using the SAS Version 8.1 statistical software package (SAS, Inc, Cary NC).15
Death certificates were reviewed for 1420 unintentional drownings. Demographic characteristics of decedents are shown in Table 1. Thirty-seven percent of drowning victims were between 1 and 4 years of age, and 29% were between 15 and 19 years of age. Seventy-four percent of victims were male. The race was listed as white for 72%, black for 22%, other races for 4%, and was not stated for 1%.
Site of Drowning
Among the 1420 drowning deaths for which death certificates were available, 669 (47%) occurred in freshwater, the most common sites being rivers/creeks, lakes, and ponds, and 457 (32%) were in artificial pools (Table 2). Domestic sites, such as bathtubs, accounted for 125 (9%) deaths, and 57 (4%) drownings were in salt water. The place of drowning was not specified on 112 (8%) death certificates. Those with no site specified were comparable to those with a specified site with respect to age, gender, and race (P > .05).
Among infants, the majority (78%) of drownings occurred in the home (Table 3), with 71% of these in bathtubs and 16% in buckets (data not shown). Children between the ages of 1 and 4 years were most likely to drown in artificial pools (58% and 51% of drownings among 1- to 2- and 3- to 4-year-olds, respectively). Freshwater submersions also accounted for a sizable proportion of drownings in this age group (22% and 36% of drownings among 1- to 2- and 3- to 4-year-olds, respectively). After the age of 5 years, children were most likely to drown in freshwater: 54%, 61%, and 69% of drownings among 5- to 9-, 10- to 14-, and 15- to 19-year-olds, respectively.
In the West, Midwest, South, and New England, freshwater drownings were most common followed by submersions in artificial pools and domestic sites. In the Mid-Atlantic, freshwater drownings were also most common followed by submersions in artificial pools. However, the number of drownings in salt water exceeded the number in domestic settings. Age-specific patterns within each region were similar to national age-specific patterns, with infants being most likely to drown in domestic sites, 1- to 4-year-olds in swimming pools, and children >5 years old in other freshwater sites (data not shown).
Among males, 52% of drownings were in freshwater sites and 30% in artificial pools. Among females, the proportion of drownings in freshwater and artificial pools were 33% and 38%, respectively. Among whites, submersions were most common in freshwater sites (50%) followed by artificial pools (31%). Among blacks, the order was reversed with a higher proportion of drownings in artificial pools (38%) than in freshwater sites (35%).
Site-Specific Drowning Rates by Age, Race, and Gender
To determine if known racial disparities in drowning rates were because of an increased risk of drowning in specific locations, we compared site-specific drowning rates for blacks versus whites (Table 4). Among 1- to 4-year-olds, drowning rates were highest in swimming pools for all groups; males, females, blacks, and whites. Among males in this age group, the risks of drowning in both swimming pools and freshwater sites were slightly lower among blacks than whites, whereas among females, drowning rates were comparable for both races.
After the age of 5 years, drowning rates were significantly higher among black males compared with white males. Drowning rates in freshwater were comparable for black and white males except among 10- to 14-year-olds where rates were higher among blacks (RR: 2.7; 95% CI: 1.8–4.1). Conversely, rates of drowning in swimming pools were much higher among black males than among white males with rate ratios of 4.4 (95% CI: 2.4–8.0) among 5- to 9-year-olds, 15.1 (95% CI: 6.7–38.5) among 10- to 14-year-olds, and 12.8 (95% CI: 6.5–26.9) among 15- to 19-year-olds. Among females, drowning rates were low after the age of 5 years. Although the number of drownings in swimming pools was small (18 white females and 16 black females, ages 5 to 19 years) black females were also at increased risk of drowning in a swimming pool as compared with white females of the same age (Table 4).
Between the ages of 5 and 19 years, 41 white males and 67 black males drowned in swimming pools. Among white males, 18 (44%) drownings were in private residential pools, 4 (10%) in hotel or motel pools, 4 (10%) in public pools (1 each in a wave pool, water park, club, and an unspecified public pool), 2 (5%) in pools located at apartment complexes, and 13 (32%) were unspecified. Among black males, 14 (21%) drownings were in private residential pools, 10 (15%) in hotel or motel pools, 14 (21%) in public pools (5 in city pools, 3 in unspecified public pools, and 1 each in a boys/girls club, local pool, community pool, high school, water park, and swim club), 7 (10%) in pools located at apartment complexes, and 22 (33%) were unspecified (P = .09).
This study provides the first national data describing the bodies of water in which children drown. Infants were most likely to drown in bathtubs, young children in swimming pools, and older children and adolescents in natural bodies of freshwater. However, ∼25% of drownings in children ages 1 to 4 years occurred in natural freshwater settings. After the age of 10 years, the risk of drowning in a swimming pool was >10 times greater among black males compared with white males. Saltwater drownings were uncommon in all age groups
Although the site of submersion was included on 92% of death certificates, other important details about the event generally were not documented. For example, bathtub seats were not mentioned in any of the bathtub drownings, yet other sources show they were involved in at least 9 drowning deaths in 1995, the year of the study.16Similarly, for drownings in swimming pools information was lacking regarding fencing, and blood alcohol levels were not available for drownings among adolescents. A second limitation was the analysis of deaths occurring in only 1 year, 1995. Beginning with deaths occurring in 1999, injury-related deaths are coded using the International Classification of Diseases-Tenth Revision, which includes specific codes for drownings in bathtubs, swimming pools, and natural water.17 This expanded coding scheme will facilitate monitoring of drowning patterns in the future. Finally, the specific site of submersion was known for 1308 deaths, which represents 90% of the number of deaths reported to NCHS. Thus, site-specific rates reported herein probably underestimate actual site-specific rates by about 10%.
Relation to Previous Studies
Consistent with previous studies, we found that infants were most likely to drown in bathtubs.3,,7 Buckets also played a role in this age group, and it is disheartening to report that 10 years after the publication of 2 articles describing the risks of large buckets, the number of bucket-related drownings per year remains unchanged.18,,19 Swimming pools presented the greatest risk for children between the ages of 1 and 4 years, and older children were most likely to drown in freshwater. These findings are also largely consistent with regional studies.6–10
In the adolescent age group, the increased risk of drowning among black males has been well described; however, factors underlying this increased risk are unknown.20 In our study, drowning rates in swimming pools were 12 to 15 times greater among black males compared with white males ages 10 to 19 years, such that much of the increased risk of drowning among black males could be explained by the increased risk of drowning in a swimming pool. Ellis et al21 also reported increased rates of nonfatal swimming pool submersions among black males. Based on data from the current study, it seems that this elevated risk may be attributable to an increased risk of drowning in a pool that was accessible to the public rather than a private residential pool. It is possible that the public pools in which black adolescent males swim are inherently less safe, ie, with more crowded conditions and/or poor supervision. Alternatively, the increased risk of drowning among black teens could be explained by a difference in swimming ability, which in turn may be attributable to fewer opportunities to participate in swimming lessons.22,,23 Little data are available about swimming ability in the general population. A study published in 1977 reported results of a survey of 9420 students in grades 1 through 8; 60% of black males reported an inability to swim in deep water versus 25% of white males.22 A more recent 1994 survey of swimming ability among adults had similar results with 62% of blacks reporting limited swimming ability versus 32% of whites.23
Implications for Preventive Counseling
In a 1993 policy statement that addressed the prevention of childhood drownings, the American Academy of Pediatrics recommended multiple topics for preventive counseling including: constant supervision of infants and young children when they are in the bathtub or around other bodies of water; installation of isolation fencing that separates the pool from the house for homes with residential pools; use of personal floatation devices when riding on a boat or playing near a river, lake, or ocean; the importance of teaching children never to swim alone or without adult supervision; the dangers of alcohol and drug consumption during aquatic activities; and the need for parents and teens to learn cardiopulmonary resuscitation.24 Our data support this multifaceted approach, given the variety of sites in which children drown, even within specific age groups. The American Academy of Pediatrics also recommends that after the age of 5 years, all children learn how to swim.24,,25 Because a relatively high proportion of black children between the ages of 5 and 19 years drown in swimming pools, sites that are inherently less dangerous than natural bodies of water, our data suggest that this recommendation may be particularly relevant for this group.
This research was funded through an interagency agreement between the National Institute of Child Health and Human Development and the Consumer Product Safety Commission. Dr Gordon Smith was supported by the National Institute of Alcohol Abuse and Alcoholism (R29AA07700) and the University of Auckland Injury Prevention Research Center.
We thank Dr Mark Klebanoff, Dr Peter Scheidt, and Tom Schroeder for their thoughtful reviews of the manuscript.
- Received June 15, 2000.
- Accepted November 6, 2000.
Reprint requests to (R.A.B.) Division of Epidemiology, Statistics, and Prevention Research, National Institute of Child Health and Human Development, 6100 Executive Blvd, Room 7B03, Bethesda, MD 20892. E-mail:
- ICD-9 =
- International Classification of Diseases, Ninth Revision •
- CPSC =
- Consumer Product Safety Commission •
- NCHS =
- National Center for Health Statistics •
- CI =
- confidence interval •
- RR =
- rate ratio
- ↵Center for Disease Control and Prevention. Compressed Mortality File, CDC WONDER Website. Available at: http://wonder.cdc.gov Accessed October 20, 2000
- ↵Baker SP, O'Neill B, Ginsburg MJ, Li G. The Injury Fact Book. New York, NY: Oxford University Press; 1992:174–183
- ↵National Center for Health Statistics, Fingerhut L. Trends and current status in childhood mortality, United States, 1980–85. Vital Health Stat 3. 1989. No. 26
- ↵World Health Organization. World Health Classification: Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death Ninth Revision. Geneva, Switzerland: World Health Organization; 1977
- Quan L,
- Gore EJ,
- Wentz K,
- Allen J,
- Novack AH
- ↵Warneke CL, Cooper SP. Child and adolescent drownings in Harris County, Texas, 1983–1990. Am J Public Health 1994;84;593–598
- Nichter MA,
- Everett PB
- ↵National Center for Health Statistics. Mortality Detail, 1995 Data, Public Use Data File Documentation. Hyattsville, MD: US Dept of Health and Human Services; 1997
- ↵SAS Institute, Inc. SAS/STAT software: changes and enhancements through release 6.11. Cary, NC: SAS Institute, Inc; 1996
- ↵Rauchschwalbe R, Brenner R, Smith G. The role of bathtub seats and rings in infant drowning deaths. Pediatrics. 1997;100(4). URL: http://www.pediatrics.org/cgi/content/full/100/4/e1
- ↵World Health Organization. ICD-10: International Statistical Classification of Diseases and Related Health Problems,Tenth Revision. Geneva, Switzerland: World Health Organization; 1992
- Mann NC,
- Weller SC,
- Rauchschwalbe R
- ↵Smith G, Brenner R. The changing risks of drowning for adolescents in the US and effective control strategies. Adolescent Medicine: State-of-the-Art-Reviews. 1995;6:153–169
- Schuman SH,
- Rowe JR,
- Glazer HM,
- Redding JS
- American Academy of Pediatrics, Committee on Injury and Poison Prevention
- American Academy of Pediatrics, Committee on Injury and Poison Prevention
- Copyright © 2001 American Academy of Pediatrics