- Email: [email protected]
Trends and Patterns of Playground Injuries in United States Children and Adolescents
Trends and Patterns of Playground Injuries in United States Children and Adolescents Kieran J. Phelan, MD; Jane Khoury, MSc; Heidi J. Kalkwarf, PhD; Bruce P. Lanphear, MD, MPH Objective.—To determine the prevalence, trends, and severity of injuries attributable to playground falls relative to other common unintentional mechanisms that resulted in an emergency department (ED) visit in the United States. Design and Setting.—Data from the emergency subset of the National Hospital Ambulatory Medical Care Survey collected from 1992 to 1997 for children ,20 years. Methods.—Injury rates and 95% confidence intervals (CIs) were estimated and injury severity scores were computed. Results.—There were 920 551 (95% CI: 540 803 to 1 300 299) ED visits over the 6-year study period by children and adolescents that were attributable to falls from playground equipment. The annual incidence of visits for playground injuries did not significantly decrease over the course of the study (187 000 to 98 000, P 5 .053). Injury visits for playground falls were twice as prevalent as pedestrian mechanisms, but they were less prevalent than visits for motor vehicle– and bicycle-related injuries. A larger proportion of playground falls resulted in ‘‘moderate-to-severe’’ injury than did bicycle or motor vehicle injuries. Children aged 5 to 9 years had the highest number of playground falls (P 5 .0014). Playground falls were most likely to occur at school compared to home, public, and other locations (P 5 .0016). Conclusions.—Playground injury emergency visits have not significantly declined and remain a common unintentional mechanism of injury. Injury visits for playground falls were proportionally more severe than injury visits attributable to other common unintentional mechanisms. Interventions targeting schools and 5- to 9-year-old children may have the greatest impact in reducing emergency visits for playground injuries. KEY WORDS: children; injury; National Hospital Ambulatory Medical Care Survey; playground falls; populationspecific rates; prevalence; trends; unintentional injury Ambulatory Pediatrics 2001;1:227 233
T
he number of hospital admissions and clinic visits attributable to falls from playground equipment in the United States is high.1–3 It is estimated that over 200 000 emergency department (ED) visits occur annually in children as a result of falls from playground equipment.4–6 Injuries associated with playground falls account for 15 to 20 deaths in children each year in the United States.6,7 Injury associated with playground equipment has been estimated to account for more than one third of fractures or dislocations and hospital admissions for children as a result of recreational and sports activities.1,2 Designing and targeting interventions to reduce playground falls requires an understanding of the patterns of injury, the children most likely to be affected, and the locations of occurrence. Numerous factors increase the risk of playground injuries. Male children less than 7 years of age are at highest risk for playground injury.8,36 Playgrounds located in the home or public setting account for more injuries than playgrounds in school, daycare, or
other settings.6,9 There have been increasing education and research efforts aimed at reducing playground-related injury over the past decade.4–6,10,11 It is unclear whether this increased attention to playground safety has led to a reduction in the numbers, rate, or severity of such injuries in US children. Nationally representative estimates of the prevalence and trends for playground injuries are lacking. The crude number of playground injuries is available from the US Consumer Product Safety Commission’s (CPSC) National Electronic Injury Surveillance System (NEISS), but population-specific rates and trends for injuries from playground falls are not.6,10 We conducted a study of injury visits to US EDs from 1992 through 1997 for children less than 20 years of age with injuries sustained during playground falls to determine the magnitude and severity of a potentially preventable injury mechanism. This study had 3 hypotheses: 1) playground falls decreased between 1992 and 1997; 2) the severity of such injuries decreased over the period of study; and 3) playground fall injuries were less severe than those sustained during bicycle or motor vehicle crashes. Finally, we also planned to describe the impact of child characteristics, location, and season of injury on the distribution of playground injury visits.
From the Division of General and Community Pediatrics, Department of Pediatrics, Children’s Hospital Medical Center, Cincinnati, Ohio. Presented at the PAS/APA/SPR and AAP combined meetings, Boston, Mass, May 2000. Address correspondence to Kieran J. Phelan, MD, Division of General and Community Pediatrics, Children’s Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229 (e-mail: [email protected]). Received for publication September 6, 2000; accepted March 23, 2001. AMBULATORY PEDIATRICS Copyright q 2001 by Ambulatory Pediatric Association
METHODS Data Set Data on patient visits to EDs between 1992 and 1997 obtained from the National Hospital Ambulatory Medical Care Survey (NHAMCS) were used for this analysis.12
227
Volume 1, Number 4 July-August 2001
228
Phelan et al
AMBULATORY PEDIATRICS
Weighted estimates of the average annual number of injury visits to US emergency departments for children less than 20 years of age by external cause of injury (E-Codes); NHAMCS 1992–97.
The NHAMCS, conducted by the National Center for Health Statistics (NCHS), is a nationally representative probability sample of patient visits to EDs and hospitalbased outpatient clinics of noninstitutional, non-Federal, general, and short-stay hospitals (excludes Federal, Veterans Administration, and military hospitals) in the United States. The survey uses a 4-stage complex probability sampling design. Each selected hospital participates for a randomly selected 4-week period during a given survey year. The staff of the participating hospitals abstract data at or near the time of a patient visit during the selected 4-week period. The data collected include demographics, the reason for visit, whether the visit was for an injury, the E-coded mechanism of injury, 3 diagnoses according to the ICD9-CM system (N-codes), and the place of occurrence of the injury. Place of occurrence was not collected in 1992. This database did not contain information on specific playground equipment or surfacing. Records reflected visits without individual patient identifiers. Records provided no information on the amount or time of exposure to playground environments. Emergency Department Visits for Injuries We limited the analysis to injury visits for children less than 20 years of age. Injury visits were identified by the presence of an ICD9-CM E code in either the primary, secondary, or tertiary E code fields. The collapsed, unweighted data set of injury visits (E800-999) to US EDs for children less than 20 years contained 19 391 records, representing more than 68 million ED visits for the 6-year period. Unintentional injuries were defined as those records with primary E codes in the ranges of 800–849.9 and 880– 929.9 and intentional injuries as 950–978. Playground falls were identified by the E code 884.0 (Figure 1). The NHAMCS samples patient visits rather than indi-
vidual persons. Repeated visits for the same injury event will result in inflated estimates of injuries for a given mechanism (eg, multiple ED visits for back or neck pain resulting from the same motor vehicle collision). We examined the proportion of visits for each mechanism that resulted in a repeat visit in 1992, the only survey year that included a question on ‘‘injury, follow-up visit.’’ In 1992 follow-up visits accounted for 5.4% of all injury visits and 4.8% of injury visits for playground falls. NCHS coding rules for the place of occurrence of an injury were revised every 2 years of the survey; therefore, for consistency, all place of occurrence codes were recoded to ‘‘home,’’ ‘‘school/daycare,’’ ‘‘public,’’ ‘‘other,’’ and ‘‘unknown.’’ The 1992 survey did not include place of occurrence; therefore, we coded location of injury for 1992 as missing (not ‘‘unknown’’). Injury Severity Injury severity was measured by determining the abbreviated injury scale (AIS) score and the new injury severity score (NISS).13,14 The AIS score describes the severity of each injury sustained by a patient on a scale of 1 (least severe) to 6 (most severe) in 9 defined body regions.13 The AIS was originally developed by the Association for the Advancement of Automotive Medicine to compare trauma outcomes and is applicable to various injury mechanisms.15,16 The NISS, a refinement of the Injury Severity Score (ISS),17 is the sum of squares of the AIS scores for the 3 most severe injuries in the body.14 AIS and NISS scores correlate well with mortality, with increasing scores indicating a greater probability of death.15,18 The AIS and NISS were determined from the ICD9-CM N-codes in the records using ICDMAP-90 software.19,20 The scores developed by this software result in conservative estimates of injury severity.21 This mapping
AMBULATORY PEDIATRICS
Playground Injuries in US Children
229
TABLE 1. Rate* (per 100,000) and 95% Confidence Interval (CI) Limits of Emergency Department Visits by Mechanism of Injury: United States Children ,20 Years Old, 1992–97 Mechanism Falls, rate (95% CI) Playground falls,† rate (95% CI) Bicycle, rate (95% CI) Motor vehicle, rate (95% CI)
1992
1993
1994
1995
1996
1997
4068 (2562–5574) 256 (160–352) 571 (306–835) 1444 (780–2109)
3841 (2329–5352) 232 (108–356) 464 (201–726) 1570 (875–2265)
4441 (2399–6482) 166 (34–298) 589 (269–910) 1351 (700–2001)
3655 (1892–5419) 235 (77–394) 395 (202–587) 1512 (873–2151)
3758 (2116–5399) 211 (114–308) 478 (239–717) 1543 (978–2109)
3072 (1695–4449) 127 (61–193) 519 (249–789) 1512 (868–2157)
*Adjusted for age, gender, and race. †P 5 .053, decreasing trend over time.
software could not assign a severity score to 23% of all injury visits and 13% of playground fall-related visits because of a lack of specific injury information. We used a conservative strategy and assigned severity scores of ‘‘0’’ or ‘‘1’’ (no or mild injury) to those records that did not have assignable severity scores to minimize the potential bias associated with missing data. The mapping software would not assign a score to nonspecific N-codes, such as ‘‘head injury, unspecified’’ (we assigned a score of ‘‘1’’ to these), and to N-codes outside the range 800–999 (we assigned a score of ‘‘0’’ to these). Analysis of severity scores with and without missing values did not alter the conclusions of this study. Statistical Analyses The complex sampling design of the NHAMCS requires specialized statistical procedures to develop valid variance estimates and confidence intervals for hypothesis testing. We computed weighted prevalence estimates and population specific rates with 95% confidence intervals (CIs) using SUDAAN to account for the complex sampling design.22 The NCHS Research Data Center provided information on the primary sampling unit and stratification variables needed for these computations. The lower limit of reliability for unweighted records to produce population estimates and accurate confidence intervals is 30 records per cell.23 Standardized rates were developed using the US Census data accounting for differences in the age, gender, and race distribution, as appropriate. Because of the categorical nature of the data, the Chisquare statistic was used to test for differences among proportions of injury visits according to age, race, gender, place of occurrence, season, and region of the country. Differences in proportions over time for playground injury visits for the 6-year study period and for two 3-year periods (1992–94 and 1995–97) were analyzed by the Maentel-Haenzel test for trend.24 The distribution of injury severity scores was not normal; therefore, we examined the median and percentiles of the AIS and NISS scores. We classified injury severity according to the number and proportion of AIS scores greater than 1 and NISS scores greater than 4 (Table 2). These cut-offs correspond to the 90th percentiles for emergency visits across all mechanisms of injury and constitute ‘‘moderate-to-severe’’ injuries. Other investigators
have used similar cut-offs for AIS and ISS scores to define more severe (AIS . 1)25,26 or mild injuries (ISS # 4).27 The proportion of injuries above the 90th percentiles of severity (AIS . 1 and NISS . 4) among injury mechanisms and over time were compared using the Chi-square statistic. A P value of ,.05 was considered statistically significant. RESULTS There were approximately 920 500 (95% CI, 540 803– 1 300 299) visits for playground falls out of more than 68 million visits to EDs for injuries to children less than 20 years of age during the 6-year study period. Unintentional injury visits outnumbered those attributable to intentional mechanisms by almost 20 to 1 (Figure). Fall-related injuries accounted for the greatest proportion of injury visits in children. These injuries represented 25% of all injury visits for children to US EDs from 1992 to 1997. Playground falls accounted for 5.3% of all fall-related injury visits. The average annual number of ED visits resulting from playground falls was 153 425 (95% CI, 90 134– 216 717). The number of playground fall injury visits did not decrease significantly over the 6-year study period (P 5 .053, Table 1). There was no change in the number or adjusted rate of fall-related, bicycle-related, or motor vehicle–related injuries presenting to EDs during the period of the study. There were insufficient numbers of records for pedestrian injury visits to develop reliable annual estimates or rates. The majority of unintentional injuries in children presenting to EDs in this study were minor, but a higher proportion of playground injuries were ‘‘moderate-to-severe’’ compared to injuries due to motor vehicles, bicycles, or all falls. The medians of the AIS and NISS distributions for all unintentional injury categories examined were 1. The distribution of AIS scores for all unintentional mechanisms (except bicycle injury visits) examined in this study were identical; the 10th and 90th percentiles were 1 and 3, respectively. The 10th and 90th percentiles of AIS scores for bicycle-related injuries were 1 and 2, respectively. The 10th and 90th percentiles of NISS scores for all fall and playground fall injuries were 1 and 10, for motor vehicle–related injuries, 1 and 9, and for bicyclerelated injuries, 1 and 8. Playground falls had the highest proportion of injuries classified as ‘‘moderate-to-severe’’
230
Phelan et al
AMBULATORY PEDIATRICS
TABLE 2. The Number (in Thousands) and Proportion (%) of Moderate-to-Severe Injuries (Abbreviated Injury Scale [AIS] . 1 and New Injury Severity Score [NISS] . 4) by Mechanism of Injury 1992–94 Mechanism
AIS . 1 Number (%)
All falls Playground fall Bicycle Motor vehicle
1994 (21.8)* 213 (44.0) 270 (22.5) 290 (9.0)
1995–97 NISS . 4 Number (%) 500 39 95 165
(4.5) (8.1) (7.9) (5.1)
AIS . 1 Number (%) 1723 158 198 381
NISS . 4 Number (%)
(21.5)* (36.2) (18.6) (10.9)
525 51 62 221
(6.6) (11.6) (5.8) (6.3)
*P , .001 among mechanisms.
based on the AIS scores (.1) compared with the other unintentional mechanisms (P , .001, Table 2). This categorization resulted in playground falls having a similar total number of ‘‘moderate-to-severe’’ injuries requiring an ED visit compared with motor vehicle– and bicyclerelated injuries, despite the fact that there were a greater total number of visits for motor vehicle or bicycle crashes. The type of tissue injury and its anatomic location will determine the ultimate severity of an injury.13,14,28 To assess the severity of injury further, we examined region of the body injured and the type of injury sustained. The most frequently injured regions of the body resulting from playground falls were the extremities followed by the head and neck, the thorax, and the abdomen or pelvis (Table 3). The most frequent type of injury was a fracture or dislocation followed by a contusion or crush (Table 4). There were insufficient numbers of injury subtypes to develop accurate annual estimates or rates. Motor vehicle injury visits resulted in greater numbers of hospital admissions or ED transfers than did playground fall visits. We estimated 393 383 admissions or ED transfers due to motor vehicle–related injuries compared to 40 158 for playground falls during the 6-year study period. Approximately 5.9% of visits for motor vehicle–related injuries were admitted, compared to 4.4% of visits for playground falls. There were insufficient numbers of ‘‘dead on arrival,’’ intubation, or resuscitation procedures in emergency visits for any mechanism to develop reliable estimates for such measures of injury severity. Because alterations in playground environments might have occurred over the period of study, we examined the severity of injuries due to playground falls for 2 periods within the 6-year study: 1992 through 1994 and 1995 TABLE 3. Number (in Thousands), Rate* (per 100,000), and 95% Confidence Interval (CI) Limits of Injuries From Playground Falls According to Body Region
Head/neck Thorax/abdomen/pelvis Extremities
Number (95% CI)
Rate (95% CI)
277 (154–339) 78† (36–121) 527 (291–763)
61 (34–89) 17 (8–27) 117 (64–169)
*Adjusted for age, gender, and race. †Number of records below limit of reliability for accurate estimates.
through 1997. There was no significant difference in the number or proportion of moderate-to-severe injuries from playground falls between the 2 periods (P 5 0.29). Because of insufficient numbers of moderate-to-severe injuries, it was not possible to evaluate injury severity by individual year. Children in the 5–9-year-old age group had significantly higher standardized rates of ED visits for playground falls compared to the 1–4- and 10–14-year-old age groups (P 5 .001, Table 5). Children 5–9 years of age were 3 times as likely to have an ED visit for a playground fall as were children aged 10–14 years. There were insufficient numbers of records to develop reliable estimates for 15–19year-old children. The rate of playground injury visits was not higher in boys compared to girls nor in white compared to black children. There were insufficient numbers of injury visits to develop estimates for other racial/ethnic groups and for individual age-by-gender or age-by-race categories. The standardized rates of ED visits for playground falls did not vary by region of the country (P 5 .57, Table 5). The rates of visits for playground falls did vary significantly by season (P , .001, Table 5), with higher rates in the spring (April–June) and summer (June–September) compared to the winter (January–March). The most frequent location of a playground fall was the school (including daycare). Playground injuries at schools accounted for 40.4% of such injuries, the home for 25.3% percent, and public parks or recreation areas for 9.4%.
TABLE 4. Number (in Thousands), Rate* (per 100,000), and 95% Confidence Interval (CI) Limits of Injuries From Playground Falls According to Type of Injury
Fracture/dislocation Sprain/strain Internal Open wound/superficial Contusion/crush
Number (95% CI)
Rate (95% CI)
367 (164–571) 85† (39–130) 82† (38–126) 184 (96–273) 199 (112–285)
81 (36–126) 19 (9–29) 18 (8–28) 41 (21–61) 44 (25–63)
*Adjusted for age, gender, and race. †Number of records below the limit of reliability for accurate estimates.
AMBULATORY PEDIATRICS
Playground falls accounted for a greater proportion (16.9%) of all falls at school compared to less than 6% in other locations (P 5 .002, Table 5). The school remained the predominant location of injury for playground falls, independent of the region of the United States or season of the year. DISCUSSION We estimate that approximately 150 000 visits to EDs for playground injuries occur each year in the United States among children and adolescents. We found a decreasing trend in emergency visits for injuries from playground falls, but it did not reach statistical significance (P 5 .053, Table 1). The proportion of playground falls resulting in a severe injury was significantly greater than the proportion for falls resulting from other unintentional mechanisms (P , .001, Table 2); however, the proportion of severe playground injuries did not significantly change over time. The unexpected severity of playground falls in this study could be attributable to a number of factors. We speculate that parents may be more comfortable assessing the severity of playground injuries compared with other types of injuries. Therefore, parents and other care providers, such as paramedics, may be more likely to bring a child injured in a motor vehicle crash to an ED for evaluation than one injured from a playground fall. If this is the case, then playground injuries are underrepresented relative to other types of injury in the ED setting. Despite the fact that more children die from motor vehicle–related injuries each year in the United States than from playground falls,7,29 visits for playground fall injuries resulted in a similar proportion of hospital admissions in this survey. The rate and severity of injuries from playground falls did not significantly decline over the period of study. One reason for the stable rates of playground injury in the United States may be increased exposure. More playgrounds may have been built, and existing ones may have experienced increased usage. Despite the construction of safer playgrounds, these effects would lead to a relative increase in falls, injuries, and ED visits. We were unable to measure exposure to playground environments from the NHAMCS data set. There has been increasing education, research, and several controlled trials aimed at reducing playground-related injury over the past decade.4–6,11,30 Despite these efforts, we found no significant decline in the rate or severity of such injuries over the 6-year period of study. Preventing playground injuries is a complex child public health issue requiring coordination of school, park district, and parental efforts.30,31 The estimates for playground fall visits developed from these NHAMCS data contrast with slightly higher estimates (150 000 vs 210 000) generated from the CPSC NEISS.6,10 The CPSC NEISS–generated estimates for playground falls are within the 95% CIs developed with these NHAMCS analyses.10 These discrepancies may be due to differences in sampling methodologies between the 2 surveys.
Playground Injuries in US Children
231
TABLE 5. Number (in Thousands), Rate* (per 100,000), and 95% Confidence Interval (CI) Limits of Emergency Department Visits for Playground Injuries According to Child Characteristics, Location, and Season: United States Children 1992–97
Age (years)† 1–4 5–9 10–14 15–19 Gender Female Male Race Black White Region Northeast South Midwest West Season§ January–March April–June July–September October–December Place†\ Home School/daycare Public Other Unknown
Number (95% CI)
Rate (95% CI)
% of All Falls
237 (112–361) 512 (289–735) 164 (91–236) ‡
251 (118-384) 451 (255–647) 146 (81–210) ‡
4.3
383 (232–534) 537 (274–800)
174 (106–243) 230 (117–344)
5.2
117 (46–188) 641 (387–896)
165 (64–267) 211 (126–295)
4.7
142 (43–241) 276 (85–468) 299 (55–543) 203 (8–399)
166 (46–286) 174 (52–295) 282 (54–510) 196 (8–384)
4.2
112 (45–180) 314 (182–446) 343 (202–484) 151 (59–243)
25 (10–40) 70 (40–99) 76 (45–107) 34 (13–54)
2.9
186 (83–288) 296 (153–440) 69 (21–117) 110 (34–185) 73 (18–128)
41 (18–64) 66 (34–97) 15 (5–26) 24 (8–41) 16 (4–28)
2.4
12.1 4.3 ‡
5.5
5.4
5.3 5.7 6.2
6.9 7.2 3.9
16.9 5.9 7.4 3.7
*Adjusted for age, gender, and race, where appropriate. †Significantly different among age group categories, P , .05. ‡Insufficient numbers to generate reliable estimates or standard errors. §Significantly different among seasonal categories, P , .001. \1992 Survey data not included.
We were unable to estimate the total number of visits to non–emergency care facilities for playground fall injuries using the NHAMCS database. The National Health Interview Survey includes data on all injury events collected from household interviews for a 3-month recall period. The National Health Interview Survey estimates that almost 500 000 episodes of injury from playground equip-
232
Phelan et al
ment occur each year in the United States that result in a phone call or visit to a health care professional.3 This implies that only 30.0% of all playground injury episodes result in a ED evaluation. We examined a number of risk factors for playground injuries. Children 5–9 years of age were at highest risk for playground injuries compared to other age groups. Similarly, in New Zealand, playground injuries peaked in children between 5 and 9 years of age; the risk of injury in 5–9-year-olds was threefold higher than the risk in 10– 14-year-olds and sevenfold higher than the risk in 0–4year-olds.32 Certain groups of children may experience more exposure to playground environments than others. Indeed, in this study there were only 2 unweighted records in the 15–19-year-old age group for visits for playground fall injuries. We found in the present study that the prevalence of playground falls did not differ by gender. In contrast, others have found that males are approximately 2 times more likely to experience playground injuries than are females.1,32 Playground injuries occur in a variety of settings, including daycare facilities 5,8,33–35 school grounds,2 public parks,36 and private residences. This is the first national survey to identify the school (including daycare) as the primary location of playground injuries. This likely reflects the playground setting with the highest exposure for 5–9-year-old children. Children 5–9 years of age spend a great proportion of their time in school settings. Many primary schools provide a playground as a stimulating recess or physical education activity. It would follow that 5–9-year-old children would be more likely to be exposed to such playground environments. Also, caregivers in school settings may be more likely than a parent or guardian to take a child injured from a playground fall to an emergency facility. Daycare centers were included in our school category, and playground falls have been shown to be a leading mechanism of injury in daycare centers in other studies.2,8,33,34 One study of the US CPSC NEISS database reported in 1994 that public playgrounds were the most common place of occurrence for injury.10 However, that study included school and daycare playgrounds in their public playground definition. The rate of ED visits for playground injuries varied by season. The number and rates of visits were significantly higher in the spring or summer compared to the winter. Mack et al6 also found a seasonal variation with peaks in May and September. This seasonal variation corresponds closely with the end and start, respectively, of the school year in the primary grades. Playgrounds serve to provide a stimulating, protected environment where children can exercise and enhance motor and social skills.37,38 Increased access to safe and secure playground environments has been utilized to reduce injuries due to pedestrian, motor vehicles, and street violence.39–41 Despite the obvious benefits of playgrounds, these environments should be engineered to provide safety from falls and other injuries associated with their design. Certain characteristics of playground equipment and ground surfacing have been identified as risk factors for
AMBULATORY PEDIATRICS
injury 25,32,42,43 Falls from heights, such as those occurring from swings, climbers, and slides, in particular, have been associated with playground-related injury.6,44 Falls from heights greater than 1.5 m, for example, were associated with a 10-fold greater risk for injuries requiring medical attention compared with falls from heights of ,1.5 m.25,32 The type and severity of injury are also related to the impact-absorbing properties of the playground surface.32,43 In one study, falls to concrete or tarmac resulted in a fivefold increased risk of injury compared with falls to rubberized surfaces.42 Injuries due to falls from playground equipment in the United States result in an unacceptably high number of ED visits. Indeed, the results of these analyses indicate that the severity of playground injuries has been underappreciated; these injuries in children are proportionally more severe than other more common unintentional injury mechanisms evaluated in US EDs. Targeting interventions to school and daycare settings will have the largest impact on reducing emergency visits and injuries associated with playground falls. These interventions could reduce the risk and increase the benefits associated with childhood exposure to playground environments. ACKNOWLEDGMENTS This research was supported by a National Research Service Award (HRSA PE10027) and an APA Special Projects Grant (APA 31-160-608).
REFERENCES 1. Bijur PE, Trumble A, Harel Y, et al. Sports and recreation injuries in US children and adolescents. Arch Pediatr Adolesc Med. 1995;149:1009–1016. 2. Di Scala C, Gallagher SS, Schneps SE. Causes and outcomes of pediatric injuries occurring at school. J School Health. 1997; 67:384–389. 3. Warner M, Barnes B, Fingerhut L. Injury and Poisoning Episodes and Conditions; National Health Interview Survey, 1997. Hyattsville, Md: Centers for Disease Control/National Center for Health Statistics; 2000. 4. Playground-related injuries in preschool-aged children—United States, 1983–1987. Morbid Mortal Wkly Rep. 1988;37:629– 632. 5. Playground safety—United States, 1998–1999. Morbid Mortal Wkly Rep. 1999;48:329–332. 6. Mack M, Thompson D, Hudson S. Playground injuries in the 90’s. Playground Recreation. 1998:81–94. 7. Drago DA, Winston FK, Baker SP. Clothing drawstring entrapment in playground slides and school buses. Contributing factors and potential interventions. Arch Pediatr Adolesc Med. 1997;151:72–77. 8. Sacks JJ, Smith JD, Kaplan KM, et al. The epidemiology of injuries in Atlanta day-care centers. JAMA. 1989;262:1641– 1645. 9. Anonymous. Playground injuries. NY State J Med. 1988;88: 667–668. 10. Mack MG, Hudson S, Thompson D. A descriptive analysis of children’s playground injuries in the United States 1990–4. Inj Prev. 1997;3:100–103. 11. Sacks JJ, Brantley MD, Holmgreen P, Rochat RW. Evaluation of an intervention to reduce playground hazards in Atlanta child-care centers. Am J Public Health. 1992;82:429–431. 12. McCaig LF McLemore, T. Plan and operation of the National Hospital Ambulatory Medical Care Survey. National Center for Health Statistics. Vital Health Stat. 1994;1:1–79.
AMBULATORY PEDIATRICS 13. The Abbreviated Injury Scale. Update 98. Des Plaines, Ill: Association for the Advancement of Automotive Medicine; 1990. 14. Osler T, Baker SP, Long W. A modification of the injury severity score that both improves accuracy and simplifies scoring. J Trauma Inj Infect Crit Care. 1997;43:922–926. 15. Sacco W, MacKenzie E, Champion H, et al. Comparison of alternative methods for assessing injury severity based on anatomic descriptors. J Trauma Inj Infect Crit Care. 1999;47:441– 447. 16. Marganitt B, MacKenzie EJ, Deshpande J, et al. Hospitalizations for traumatic injuries among children in Maryland—trends in incidence and severity 1979 through 1988. Pediatrics. 1992; 89:608–613. 17. Baker S, O’Neill B, Haddon W, Long WB. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. Trauma. 1974;14:187–196. 18. Wan G, Neffsmith M. The impact of demographics, injury severity, and trauma type on the likelihood of survival in child and adolescent trauma patients. J Trauma. 1996;40:412–416. 19. Tri-Analytics I. ICDMAP-90. Bel Air, Md; 1998. 20. MacKenzie E, Steinwachs D, Shankar B. Classifying severity of trauma based on hospital discharge diagnoses: validation of an ICD-9CM to AIS-85 conversion table. Med Care. 1989;27: 412–422. 21. MacKenzie E, Morris J, Smith G, Fahey M. Acute hospital costs of trauma in the United States: implications for regionalized systems of care. J Trauma. 1990;30:1096–1103. 22. Shah B, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 7.5. Research Triangle Park, NC: Research Triangle Institute; 1997. 23. Burt C, Fingerhut L. Injury Visits to Hospital Emergency Departments: United States, 1992–95. Hyattsville, Md: National Center for Health Statistics. Vital Health Statistics; 1998. 24. Fleiss J. Statistical Methods for Rates and Proportions. 2nd ed. New York: John Wiley & Sons; 1981. 25. Macarthur C, Hu X, Wesson D, Parkin P. Risk factors for severe injuries associated with falls from playground equipment. Accid Anal Prev. 2000;32:377–382. 26. Durbin D, Winston F, Applegate S, et al. Development and validation of the injury severity assessment survey/parent report. Arch Pediatr Adolesc Med. 1999;153:404–408. 27. Hoff W, Tinkoff GH, Lucke JF, Lehr S. Impact of minimal injuries on a level I trauma center. J Trauma. 1992;33:408–412. 28. Robertson L. Injury Epidemiology. 2nd ed. Oxford: Oxford University Press; 1998.
Playground Injuries in US Children
233
29. Rivara F, Grossman DC. Prevention of traumatic deaths to children in the United States: how far have we come and where do we need to go? Pediatrics. 1996;97:791–797. 30. Sibert J, Mott A, Rolfe K, et al. Preventing injuries in public playgrounds through partnership between health services and local authority: community intervention study. BMJ. 1999;318: 1595. 31. Laraque D, Barlow B, Durkin M, Heagarty M. Injury prevention in an urban setting: challenges and successes. Bull NY Acad Med. 1995;72:16–30. 32. Chalmers DJ, Marshall SW, Langley JD, et al. Height and surfacing as risk factors for injury in falls from playground equipment: a case-control study. Inj Prev. 1996;2:98–104. 33. Briss PA, Sacks JJ, Addiss DG, et al. A nationwide study of the risk of injury associated with day care center attendance. Pediatrics. 1994;93:364–368. 34. Sacks JJ, Holt KW, Holmgreen P, et al. Playground hazards in Atlanta child care centers. Am J Public Health. 1990;80:986– 988. 35. Alkon A, Genevro JL, Tschann JM, et al. The epidemiology of injuries in 4 child care centers. Arch Pediatr Adolesc Med. 1999;153:1248–1254. 36. Mott A, Evans R, Rolfe K, et al. Patterns of injuries to children on public playgrounds. Arch Dis Child. 1994;71:328–330. 37. Coppens N, Gentry L. Video analysis of playground injury-risk situations. Res Nursing Health. 1991;14:129–136. 38. Pellegrini AD, Davis PD. Relations between children’s playground and classroom behaviour. Br J Educ Psychol. 1993;63: 88–95. 39. Laraque D, Barlow B, Davidson L, Welborn C. The Central Harlem playground injury prevention project: a model for change. Am J Public Health. 1994;84:1691–1692. 40. Davidson L, Durkin M, Kuhn L, et al. The impact of the Safe Kids/Healthy Neighborhoods Injury Prevention Program in Harlem, 1988 through 1991. Am J Public Health. 1994;84:580– 586. 41. von Kries R, Kohne C, Bohm O, von Voss H. Road injuries in school age children: relation to environmental factors amenable to interventions. Inj Prev. 1998;4:103–105. 42. Mott A, Rolfe K, James R, et al. Safety of surfaces and equipment for children in playgrounds. Lancet. 1997;349:1874–1876. 43. Mowat DL, Wang F, Pickett W, Brison RJ. A case-control study of risk factors for playground injuries among children in Kingston and area. Inj Prev. 1998;4:39–43. 44. Waltzman ML, Shannon M, Bowen AP, Bailey MC. Monkeybar injuries: complications of play. Pediatrics. 1999;103:e58.
T
he number of hospital admissions and clinic visits attributable to falls from playground equipment in the United States is high.1–3 It is estimated that over 200 000 emergency department (ED) visits occur annually in children as a result of falls from playground equipment.4–6 Injuries associated with playground falls account for 15 to 20 deaths in children each year in the United States.6,7 Injury associated with playground equipment has been estimated to account for more than one third of fractures or dislocations and hospital admissions for children as a result of recreational and sports activities.1,2 Designing and targeting interventions to reduce playground falls requires an understanding of the patterns of injury, the children most likely to be affected, and the locations of occurrence. Numerous factors increase the risk of playground injuries. Male children less than 7 years of age are at highest risk for playground injury.8,36 Playgrounds located in the home or public setting account for more injuries than playgrounds in school, daycare, or
other settings.6,9 There have been increasing education and research efforts aimed at reducing playground-related injury over the past decade.4–6,10,11 It is unclear whether this increased attention to playground safety has led to a reduction in the numbers, rate, or severity of such injuries in US children. Nationally representative estimates of the prevalence and trends for playground injuries are lacking. The crude number of playground injuries is available from the US Consumer Product Safety Commission’s (CPSC) National Electronic Injury Surveillance System (NEISS), but population-specific rates and trends for injuries from playground falls are not.6,10 We conducted a study of injury visits to US EDs from 1992 through 1997 for children less than 20 years of age with injuries sustained during playground falls to determine the magnitude and severity of a potentially preventable injury mechanism. This study had 3 hypotheses: 1) playground falls decreased between 1992 and 1997; 2) the severity of such injuries decreased over the period of study; and 3) playground fall injuries were less severe than those sustained during bicycle or motor vehicle crashes. Finally, we also planned to describe the impact of child characteristics, location, and season of injury on the distribution of playground injury visits.
From the Division of General and Community Pediatrics, Department of Pediatrics, Children’s Hospital Medical Center, Cincinnati, Ohio. Presented at the PAS/APA/SPR and AAP combined meetings, Boston, Mass, May 2000. Address correspondence to Kieran J. Phelan, MD, Division of General and Community Pediatrics, Children’s Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229 (e-mail: [email protected]). Received for publication September 6, 2000; accepted March 23, 2001. AMBULATORY PEDIATRICS Copyright q 2001 by Ambulatory Pediatric Association
METHODS Data Set Data on patient visits to EDs between 1992 and 1997 obtained from the National Hospital Ambulatory Medical Care Survey (NHAMCS) were used for this analysis.12
227
Volume 1, Number 4 July-August 2001
228
Phelan et al
AMBULATORY PEDIATRICS
Weighted estimates of the average annual number of injury visits to US emergency departments for children less than 20 years of age by external cause of injury (E-Codes); NHAMCS 1992–97.
The NHAMCS, conducted by the National Center for Health Statistics (NCHS), is a nationally representative probability sample of patient visits to EDs and hospitalbased outpatient clinics of noninstitutional, non-Federal, general, and short-stay hospitals (excludes Federal, Veterans Administration, and military hospitals) in the United States. The survey uses a 4-stage complex probability sampling design. Each selected hospital participates for a randomly selected 4-week period during a given survey year. The staff of the participating hospitals abstract data at or near the time of a patient visit during the selected 4-week period. The data collected include demographics, the reason for visit, whether the visit was for an injury, the E-coded mechanism of injury, 3 diagnoses according to the ICD9-CM system (N-codes), and the place of occurrence of the injury. Place of occurrence was not collected in 1992. This database did not contain information on specific playground equipment or surfacing. Records reflected visits without individual patient identifiers. Records provided no information on the amount or time of exposure to playground environments. Emergency Department Visits for Injuries We limited the analysis to injury visits for children less than 20 years of age. Injury visits were identified by the presence of an ICD9-CM E code in either the primary, secondary, or tertiary E code fields. The collapsed, unweighted data set of injury visits (E800-999) to US EDs for children less than 20 years contained 19 391 records, representing more than 68 million ED visits for the 6-year period. Unintentional injuries were defined as those records with primary E codes in the ranges of 800–849.9 and 880– 929.9 and intentional injuries as 950–978. Playground falls were identified by the E code 884.0 (Figure 1). The NHAMCS samples patient visits rather than indi-
vidual persons. Repeated visits for the same injury event will result in inflated estimates of injuries for a given mechanism (eg, multiple ED visits for back or neck pain resulting from the same motor vehicle collision). We examined the proportion of visits for each mechanism that resulted in a repeat visit in 1992, the only survey year that included a question on ‘‘injury, follow-up visit.’’ In 1992 follow-up visits accounted for 5.4% of all injury visits and 4.8% of injury visits for playground falls. NCHS coding rules for the place of occurrence of an injury were revised every 2 years of the survey; therefore, for consistency, all place of occurrence codes were recoded to ‘‘home,’’ ‘‘school/daycare,’’ ‘‘public,’’ ‘‘other,’’ and ‘‘unknown.’’ The 1992 survey did not include place of occurrence; therefore, we coded location of injury for 1992 as missing (not ‘‘unknown’’). Injury Severity Injury severity was measured by determining the abbreviated injury scale (AIS) score and the new injury severity score (NISS).13,14 The AIS score describes the severity of each injury sustained by a patient on a scale of 1 (least severe) to 6 (most severe) in 9 defined body regions.13 The AIS was originally developed by the Association for the Advancement of Automotive Medicine to compare trauma outcomes and is applicable to various injury mechanisms.15,16 The NISS, a refinement of the Injury Severity Score (ISS),17 is the sum of squares of the AIS scores for the 3 most severe injuries in the body.14 AIS and NISS scores correlate well with mortality, with increasing scores indicating a greater probability of death.15,18 The AIS and NISS were determined from the ICD9-CM N-codes in the records using ICDMAP-90 software.19,20 The scores developed by this software result in conservative estimates of injury severity.21 This mapping
AMBULATORY PEDIATRICS
Playground Injuries in US Children
229
TABLE 1. Rate* (per 100,000) and 95% Confidence Interval (CI) Limits of Emergency Department Visits by Mechanism of Injury: United States Children ,20 Years Old, 1992–97 Mechanism Falls, rate (95% CI) Playground falls,† rate (95% CI) Bicycle, rate (95% CI) Motor vehicle, rate (95% CI)
1992
1993
1994
1995
1996
1997
4068 (2562–5574) 256 (160–352) 571 (306–835) 1444 (780–2109)
3841 (2329–5352) 232 (108–356) 464 (201–726) 1570 (875–2265)
4441 (2399–6482) 166 (34–298) 589 (269–910) 1351 (700–2001)
3655 (1892–5419) 235 (77–394) 395 (202–587) 1512 (873–2151)
3758 (2116–5399) 211 (114–308) 478 (239–717) 1543 (978–2109)
3072 (1695–4449) 127 (61–193) 519 (249–789) 1512 (868–2157)
*Adjusted for age, gender, and race. †P 5 .053, decreasing trend over time.
software could not assign a severity score to 23% of all injury visits and 13% of playground fall-related visits because of a lack of specific injury information. We used a conservative strategy and assigned severity scores of ‘‘0’’ or ‘‘1’’ (no or mild injury) to those records that did not have assignable severity scores to minimize the potential bias associated with missing data. The mapping software would not assign a score to nonspecific N-codes, such as ‘‘head injury, unspecified’’ (we assigned a score of ‘‘1’’ to these), and to N-codes outside the range 800–999 (we assigned a score of ‘‘0’’ to these). Analysis of severity scores with and without missing values did not alter the conclusions of this study. Statistical Analyses The complex sampling design of the NHAMCS requires specialized statistical procedures to develop valid variance estimates and confidence intervals for hypothesis testing. We computed weighted prevalence estimates and population specific rates with 95% confidence intervals (CIs) using SUDAAN to account for the complex sampling design.22 The NCHS Research Data Center provided information on the primary sampling unit and stratification variables needed for these computations. The lower limit of reliability for unweighted records to produce population estimates and accurate confidence intervals is 30 records per cell.23 Standardized rates were developed using the US Census data accounting for differences in the age, gender, and race distribution, as appropriate. Because of the categorical nature of the data, the Chisquare statistic was used to test for differences among proportions of injury visits according to age, race, gender, place of occurrence, season, and region of the country. Differences in proportions over time for playground injury visits for the 6-year study period and for two 3-year periods (1992–94 and 1995–97) were analyzed by the Maentel-Haenzel test for trend.24 The distribution of injury severity scores was not normal; therefore, we examined the median and percentiles of the AIS and NISS scores. We classified injury severity according to the number and proportion of AIS scores greater than 1 and NISS scores greater than 4 (Table 2). These cut-offs correspond to the 90th percentiles for emergency visits across all mechanisms of injury and constitute ‘‘moderate-to-severe’’ injuries. Other investigators
have used similar cut-offs for AIS and ISS scores to define more severe (AIS . 1)25,26 or mild injuries (ISS # 4).27 The proportion of injuries above the 90th percentiles of severity (AIS . 1 and NISS . 4) among injury mechanisms and over time were compared using the Chi-square statistic. A P value of ,.05 was considered statistically significant. RESULTS There were approximately 920 500 (95% CI, 540 803– 1 300 299) visits for playground falls out of more than 68 million visits to EDs for injuries to children less than 20 years of age during the 6-year study period. Unintentional injury visits outnumbered those attributable to intentional mechanisms by almost 20 to 1 (Figure). Fall-related injuries accounted for the greatest proportion of injury visits in children. These injuries represented 25% of all injury visits for children to US EDs from 1992 to 1997. Playground falls accounted for 5.3% of all fall-related injury visits. The average annual number of ED visits resulting from playground falls was 153 425 (95% CI, 90 134– 216 717). The number of playground fall injury visits did not decrease significantly over the 6-year study period (P 5 .053, Table 1). There was no change in the number or adjusted rate of fall-related, bicycle-related, or motor vehicle–related injuries presenting to EDs during the period of the study. There were insufficient numbers of records for pedestrian injury visits to develop reliable annual estimates or rates. The majority of unintentional injuries in children presenting to EDs in this study were minor, but a higher proportion of playground injuries were ‘‘moderate-to-severe’’ compared to injuries due to motor vehicles, bicycles, or all falls. The medians of the AIS and NISS distributions for all unintentional injury categories examined were 1. The distribution of AIS scores for all unintentional mechanisms (except bicycle injury visits) examined in this study were identical; the 10th and 90th percentiles were 1 and 3, respectively. The 10th and 90th percentiles of AIS scores for bicycle-related injuries were 1 and 2, respectively. The 10th and 90th percentiles of NISS scores for all fall and playground fall injuries were 1 and 10, for motor vehicle–related injuries, 1 and 9, and for bicyclerelated injuries, 1 and 8. Playground falls had the highest proportion of injuries classified as ‘‘moderate-to-severe’’
230
Phelan et al
AMBULATORY PEDIATRICS
TABLE 2. The Number (in Thousands) and Proportion (%) of Moderate-to-Severe Injuries (Abbreviated Injury Scale [AIS] . 1 and New Injury Severity Score [NISS] . 4) by Mechanism of Injury 1992–94 Mechanism
AIS . 1 Number (%)
All falls Playground fall Bicycle Motor vehicle
1994 (21.8)* 213 (44.0) 270 (22.5) 290 (9.0)
1995–97 NISS . 4 Number (%) 500 39 95 165
(4.5) (8.1) (7.9) (5.1)
AIS . 1 Number (%) 1723 158 198 381
NISS . 4 Number (%)
(21.5)* (36.2) (18.6) (10.9)
525 51 62 221
(6.6) (11.6) (5.8) (6.3)
*P , .001 among mechanisms.
based on the AIS scores (.1) compared with the other unintentional mechanisms (P , .001, Table 2). This categorization resulted in playground falls having a similar total number of ‘‘moderate-to-severe’’ injuries requiring an ED visit compared with motor vehicle– and bicyclerelated injuries, despite the fact that there were a greater total number of visits for motor vehicle or bicycle crashes. The type of tissue injury and its anatomic location will determine the ultimate severity of an injury.13,14,28 To assess the severity of injury further, we examined region of the body injured and the type of injury sustained. The most frequently injured regions of the body resulting from playground falls were the extremities followed by the head and neck, the thorax, and the abdomen or pelvis (Table 3). The most frequent type of injury was a fracture or dislocation followed by a contusion or crush (Table 4). There were insufficient numbers of injury subtypes to develop accurate annual estimates or rates. Motor vehicle injury visits resulted in greater numbers of hospital admissions or ED transfers than did playground fall visits. We estimated 393 383 admissions or ED transfers due to motor vehicle–related injuries compared to 40 158 for playground falls during the 6-year study period. Approximately 5.9% of visits for motor vehicle–related injuries were admitted, compared to 4.4% of visits for playground falls. There were insufficient numbers of ‘‘dead on arrival,’’ intubation, or resuscitation procedures in emergency visits for any mechanism to develop reliable estimates for such measures of injury severity. Because alterations in playground environments might have occurred over the period of study, we examined the severity of injuries due to playground falls for 2 periods within the 6-year study: 1992 through 1994 and 1995 TABLE 3. Number (in Thousands), Rate* (per 100,000), and 95% Confidence Interval (CI) Limits of Injuries From Playground Falls According to Body Region
Head/neck Thorax/abdomen/pelvis Extremities
Number (95% CI)
Rate (95% CI)
277 (154–339) 78† (36–121) 527 (291–763)
61 (34–89) 17 (8–27) 117 (64–169)
*Adjusted for age, gender, and race. †Number of records below limit of reliability for accurate estimates.
through 1997. There was no significant difference in the number or proportion of moderate-to-severe injuries from playground falls between the 2 periods (P 5 0.29). Because of insufficient numbers of moderate-to-severe injuries, it was not possible to evaluate injury severity by individual year. Children in the 5–9-year-old age group had significantly higher standardized rates of ED visits for playground falls compared to the 1–4- and 10–14-year-old age groups (P 5 .001, Table 5). Children 5–9 years of age were 3 times as likely to have an ED visit for a playground fall as were children aged 10–14 years. There were insufficient numbers of records to develop reliable estimates for 15–19year-old children. The rate of playground injury visits was not higher in boys compared to girls nor in white compared to black children. There were insufficient numbers of injury visits to develop estimates for other racial/ethnic groups and for individual age-by-gender or age-by-race categories. The standardized rates of ED visits for playground falls did not vary by region of the country (P 5 .57, Table 5). The rates of visits for playground falls did vary significantly by season (P , .001, Table 5), with higher rates in the spring (April–June) and summer (June–September) compared to the winter (January–March). The most frequent location of a playground fall was the school (including daycare). Playground injuries at schools accounted for 40.4% of such injuries, the home for 25.3% percent, and public parks or recreation areas for 9.4%.
TABLE 4. Number (in Thousands), Rate* (per 100,000), and 95% Confidence Interval (CI) Limits of Injuries From Playground Falls According to Type of Injury
Fracture/dislocation Sprain/strain Internal Open wound/superficial Contusion/crush
Number (95% CI)
Rate (95% CI)
367 (164–571) 85† (39–130) 82† (38–126) 184 (96–273) 199 (112–285)
81 (36–126) 19 (9–29) 18 (8–28) 41 (21–61) 44 (25–63)
*Adjusted for age, gender, and race. †Number of records below the limit of reliability for accurate estimates.
AMBULATORY PEDIATRICS
Playground falls accounted for a greater proportion (16.9%) of all falls at school compared to less than 6% in other locations (P 5 .002, Table 5). The school remained the predominant location of injury for playground falls, independent of the region of the United States or season of the year. DISCUSSION We estimate that approximately 150 000 visits to EDs for playground injuries occur each year in the United States among children and adolescents. We found a decreasing trend in emergency visits for injuries from playground falls, but it did not reach statistical significance (P 5 .053, Table 1). The proportion of playground falls resulting in a severe injury was significantly greater than the proportion for falls resulting from other unintentional mechanisms (P , .001, Table 2); however, the proportion of severe playground injuries did not significantly change over time. The unexpected severity of playground falls in this study could be attributable to a number of factors. We speculate that parents may be more comfortable assessing the severity of playground injuries compared with other types of injuries. Therefore, parents and other care providers, such as paramedics, may be more likely to bring a child injured in a motor vehicle crash to an ED for evaluation than one injured from a playground fall. If this is the case, then playground injuries are underrepresented relative to other types of injury in the ED setting. Despite the fact that more children die from motor vehicle–related injuries each year in the United States than from playground falls,7,29 visits for playground fall injuries resulted in a similar proportion of hospital admissions in this survey. The rate and severity of injuries from playground falls did not significantly decline over the period of study. One reason for the stable rates of playground injury in the United States may be increased exposure. More playgrounds may have been built, and existing ones may have experienced increased usage. Despite the construction of safer playgrounds, these effects would lead to a relative increase in falls, injuries, and ED visits. We were unable to measure exposure to playground environments from the NHAMCS data set. There has been increasing education, research, and several controlled trials aimed at reducing playground-related injury over the past decade.4–6,11,30 Despite these efforts, we found no significant decline in the rate or severity of such injuries over the 6-year period of study. Preventing playground injuries is a complex child public health issue requiring coordination of school, park district, and parental efforts.30,31 The estimates for playground fall visits developed from these NHAMCS data contrast with slightly higher estimates (150 000 vs 210 000) generated from the CPSC NEISS.6,10 The CPSC NEISS–generated estimates for playground falls are within the 95% CIs developed with these NHAMCS analyses.10 These discrepancies may be due to differences in sampling methodologies between the 2 surveys.
Playground Injuries in US Children
231
TABLE 5. Number (in Thousands), Rate* (per 100,000), and 95% Confidence Interval (CI) Limits of Emergency Department Visits for Playground Injuries According to Child Characteristics, Location, and Season: United States Children 1992–97
Age (years)† 1–4 5–9 10–14 15–19 Gender Female Male Race Black White Region Northeast South Midwest West Season§ January–March April–June July–September October–December Place†\ Home School/daycare Public Other Unknown
Number (95% CI)
Rate (95% CI)
% of All Falls
237 (112–361) 512 (289–735) 164 (91–236) ‡
251 (118-384) 451 (255–647) 146 (81–210) ‡
4.3
383 (232–534) 537 (274–800)
174 (106–243) 230 (117–344)
5.2
117 (46–188) 641 (387–896)
165 (64–267) 211 (126–295)
4.7
142 (43–241) 276 (85–468) 299 (55–543) 203 (8–399)
166 (46–286) 174 (52–295) 282 (54–510) 196 (8–384)
4.2
112 (45–180) 314 (182–446) 343 (202–484) 151 (59–243)
25 (10–40) 70 (40–99) 76 (45–107) 34 (13–54)
2.9
186 (83–288) 296 (153–440) 69 (21–117) 110 (34–185) 73 (18–128)
41 (18–64) 66 (34–97) 15 (5–26) 24 (8–41) 16 (4–28)
2.4
12.1 4.3 ‡
5.5
5.4
5.3 5.7 6.2
6.9 7.2 3.9
16.9 5.9 7.4 3.7
*Adjusted for age, gender, and race, where appropriate. †Significantly different among age group categories, P , .05. ‡Insufficient numbers to generate reliable estimates or standard errors. §Significantly different among seasonal categories, P , .001. \1992 Survey data not included.
We were unable to estimate the total number of visits to non–emergency care facilities for playground fall injuries using the NHAMCS database. The National Health Interview Survey includes data on all injury events collected from household interviews for a 3-month recall period. The National Health Interview Survey estimates that almost 500 000 episodes of injury from playground equip-
232
Phelan et al
ment occur each year in the United States that result in a phone call or visit to a health care professional.3 This implies that only 30.0% of all playground injury episodes result in a ED evaluation. We examined a number of risk factors for playground injuries. Children 5–9 years of age were at highest risk for playground injuries compared to other age groups. Similarly, in New Zealand, playground injuries peaked in children between 5 and 9 years of age; the risk of injury in 5–9-year-olds was threefold higher than the risk in 10– 14-year-olds and sevenfold higher than the risk in 0–4year-olds.32 Certain groups of children may experience more exposure to playground environments than others. Indeed, in this study there were only 2 unweighted records in the 15–19-year-old age group for visits for playground fall injuries. We found in the present study that the prevalence of playground falls did not differ by gender. In contrast, others have found that males are approximately 2 times more likely to experience playground injuries than are females.1,32 Playground injuries occur in a variety of settings, including daycare facilities 5,8,33–35 school grounds,2 public parks,36 and private residences. This is the first national survey to identify the school (including daycare) as the primary location of playground injuries. This likely reflects the playground setting with the highest exposure for 5–9-year-old children. Children 5–9 years of age spend a great proportion of their time in school settings. Many primary schools provide a playground as a stimulating recess or physical education activity. It would follow that 5–9-year-old children would be more likely to be exposed to such playground environments. Also, caregivers in school settings may be more likely than a parent or guardian to take a child injured from a playground fall to an emergency facility. Daycare centers were included in our school category, and playground falls have been shown to be a leading mechanism of injury in daycare centers in other studies.2,8,33,34 One study of the US CPSC NEISS database reported in 1994 that public playgrounds were the most common place of occurrence for injury.10 However, that study included school and daycare playgrounds in their public playground definition. The rate of ED visits for playground injuries varied by season. The number and rates of visits were significantly higher in the spring or summer compared to the winter. Mack et al6 also found a seasonal variation with peaks in May and September. This seasonal variation corresponds closely with the end and start, respectively, of the school year in the primary grades. Playgrounds serve to provide a stimulating, protected environment where children can exercise and enhance motor and social skills.37,38 Increased access to safe and secure playground environments has been utilized to reduce injuries due to pedestrian, motor vehicles, and street violence.39–41 Despite the obvious benefits of playgrounds, these environments should be engineered to provide safety from falls and other injuries associated with their design. Certain characteristics of playground equipment and ground surfacing have been identified as risk factors for
AMBULATORY PEDIATRICS
injury 25,32,42,43 Falls from heights, such as those occurring from swings, climbers, and slides, in particular, have been associated with playground-related injury.6,44 Falls from heights greater than 1.5 m, for example, were associated with a 10-fold greater risk for injuries requiring medical attention compared with falls from heights of ,1.5 m.25,32 The type and severity of injury are also related to the impact-absorbing properties of the playground surface.32,43 In one study, falls to concrete or tarmac resulted in a fivefold increased risk of injury compared with falls to rubberized surfaces.42 Injuries due to falls from playground equipment in the United States result in an unacceptably high number of ED visits. Indeed, the results of these analyses indicate that the severity of playground injuries has been underappreciated; these injuries in children are proportionally more severe than other more common unintentional injury mechanisms evaluated in US EDs. Targeting interventions to school and daycare settings will have the largest impact on reducing emergency visits and injuries associated with playground falls. These interventions could reduce the risk and increase the benefits associated with childhood exposure to playground environments. ACKNOWLEDGMENTS This research was supported by a National Research Service Award (HRSA PE10027) and an APA Special Projects Grant (APA 31-160-608).
REFERENCES 1. Bijur PE, Trumble A, Harel Y, et al. Sports and recreation injuries in US children and adolescents. Arch Pediatr Adolesc Med. 1995;149:1009–1016. 2. Di Scala C, Gallagher SS, Schneps SE. Causes and outcomes of pediatric injuries occurring at school. J School Health. 1997; 67:384–389. 3. Warner M, Barnes B, Fingerhut L. Injury and Poisoning Episodes and Conditions; National Health Interview Survey, 1997. Hyattsville, Md: Centers for Disease Control/National Center for Health Statistics; 2000. 4. Playground-related injuries in preschool-aged children—United States, 1983–1987. Morbid Mortal Wkly Rep. 1988;37:629– 632. 5. Playground safety—United States, 1998–1999. Morbid Mortal Wkly Rep. 1999;48:329–332. 6. Mack M, Thompson D, Hudson S. Playground injuries in the 90’s. Playground Recreation. 1998:81–94. 7. Drago DA, Winston FK, Baker SP. Clothing drawstring entrapment in playground slides and school buses. Contributing factors and potential interventions. Arch Pediatr Adolesc Med. 1997;151:72–77. 8. Sacks JJ, Smith JD, Kaplan KM, et al. The epidemiology of injuries in Atlanta day-care centers. JAMA. 1989;262:1641– 1645. 9. Anonymous. Playground injuries. NY State J Med. 1988;88: 667–668. 10. Mack MG, Hudson S, Thompson D. A descriptive analysis of children’s playground injuries in the United States 1990–4. Inj Prev. 1997;3:100–103. 11. Sacks JJ, Brantley MD, Holmgreen P, Rochat RW. Evaluation of an intervention to reduce playground hazards in Atlanta child-care centers. Am J Public Health. 1992;82:429–431. 12. McCaig LF McLemore, T. Plan and operation of the National Hospital Ambulatory Medical Care Survey. National Center for Health Statistics. Vital Health Stat. 1994;1:1–79.
AMBULATORY PEDIATRICS 13. The Abbreviated Injury Scale. Update 98. Des Plaines, Ill: Association for the Advancement of Automotive Medicine; 1990. 14. Osler T, Baker SP, Long W. A modification of the injury severity score that both improves accuracy and simplifies scoring. J Trauma Inj Infect Crit Care. 1997;43:922–926. 15. Sacco W, MacKenzie E, Champion H, et al. Comparison of alternative methods for assessing injury severity based on anatomic descriptors. J Trauma Inj Infect Crit Care. 1999;47:441– 447. 16. Marganitt B, MacKenzie EJ, Deshpande J, et al. Hospitalizations for traumatic injuries among children in Maryland—trends in incidence and severity 1979 through 1988. Pediatrics. 1992; 89:608–613. 17. Baker S, O’Neill B, Haddon W, Long WB. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. Trauma. 1974;14:187–196. 18. Wan G, Neffsmith M. The impact of demographics, injury severity, and trauma type on the likelihood of survival in child and adolescent trauma patients. J Trauma. 1996;40:412–416. 19. Tri-Analytics I. ICDMAP-90. Bel Air, Md; 1998. 20. MacKenzie E, Steinwachs D, Shankar B. Classifying severity of trauma based on hospital discharge diagnoses: validation of an ICD-9CM to AIS-85 conversion table. Med Care. 1989;27: 412–422. 21. MacKenzie E, Morris J, Smith G, Fahey M. Acute hospital costs of trauma in the United States: implications for regionalized systems of care. J Trauma. 1990;30:1096–1103. 22. Shah B, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 7.5. Research Triangle Park, NC: Research Triangle Institute; 1997. 23. Burt C, Fingerhut L. Injury Visits to Hospital Emergency Departments: United States, 1992–95. Hyattsville, Md: National Center for Health Statistics. Vital Health Statistics; 1998. 24. Fleiss J. Statistical Methods for Rates and Proportions. 2nd ed. New York: John Wiley & Sons; 1981. 25. Macarthur C, Hu X, Wesson D, Parkin P. Risk factors for severe injuries associated with falls from playground equipment. Accid Anal Prev. 2000;32:377–382. 26. Durbin D, Winston F, Applegate S, et al. Development and validation of the injury severity assessment survey/parent report. Arch Pediatr Adolesc Med. 1999;153:404–408. 27. Hoff W, Tinkoff GH, Lucke JF, Lehr S. Impact of minimal injuries on a level I trauma center. J Trauma. 1992;33:408–412. 28. Robertson L. Injury Epidemiology. 2nd ed. Oxford: Oxford University Press; 1998.
Playground Injuries in US Children
233
29. Rivara F, Grossman DC. Prevention of traumatic deaths to children in the United States: how far have we come and where do we need to go? Pediatrics. 1996;97:791–797. 30. Sibert J, Mott A, Rolfe K, et al. Preventing injuries in public playgrounds through partnership between health services and local authority: community intervention study. BMJ. 1999;318: 1595. 31. Laraque D, Barlow B, Durkin M, Heagarty M. Injury prevention in an urban setting: challenges and successes. Bull NY Acad Med. 1995;72:16–30. 32. Chalmers DJ, Marshall SW, Langley JD, et al. Height and surfacing as risk factors for injury in falls from playground equipment: a case-control study. Inj Prev. 1996;2:98–104. 33. Briss PA, Sacks JJ, Addiss DG, et al. A nationwide study of the risk of injury associated with day care center attendance. Pediatrics. 1994;93:364–368. 34. Sacks JJ, Holt KW, Holmgreen P, et al. Playground hazards in Atlanta child care centers. Am J Public Health. 1990;80:986– 988. 35. Alkon A, Genevro JL, Tschann JM, et al. The epidemiology of injuries in 4 child care centers. Arch Pediatr Adolesc Med. 1999;153:1248–1254. 36. Mott A, Evans R, Rolfe K, et al. Patterns of injuries to children on public playgrounds. Arch Dis Child. 1994;71:328–330. 37. Coppens N, Gentry L. Video analysis of playground injury-risk situations. Res Nursing Health. 1991;14:129–136. 38. Pellegrini AD, Davis PD. Relations between children’s playground and classroom behaviour. Br J Educ Psychol. 1993;63: 88–95. 39. Laraque D, Barlow B, Davidson L, Welborn C. The Central Harlem playground injury prevention project: a model for change. Am J Public Health. 1994;84:1691–1692. 40. Davidson L, Durkin M, Kuhn L, et al. The impact of the Safe Kids/Healthy Neighborhoods Injury Prevention Program in Harlem, 1988 through 1991. Am J Public Health. 1994;84:580– 586. 41. von Kries R, Kohne C, Bohm O, von Voss H. Road injuries in school age children: relation to environmental factors amenable to interventions. Inj Prev. 1998;4:103–105. 42. Mott A, Rolfe K, James R, et al. Safety of surfaces and equipment for children in playgrounds. Lancet. 1997;349:1874–1876. 43. Mowat DL, Wang F, Pickett W, Brison RJ. A case-control study of risk factors for playground injuries among children in Kingston and area. Inj Prev. 1998;4:39–43. 44. Waltzman ML, Shannon M, Bowen AP, Bailey MC. Monkeybar injuries: complications of play. Pediatrics. 1999;103:e58.