We conducted a cross-sectional study of playground equipment-related extremity fractures presenting to US EDs in children aged ≤14 years using the National Electronic Injury Surveillance System (NEISS) from January 1, 2006 through December 31, 2016. The NEISS is operated by the Consumer Product Safety Commission and collects data on injuries that present to US hospital-based EDs (US Consumer Product Safety Commission). The NEISS samples approximately 100 hospitals to represent a stratified probability sample of the over 5000 US EDs. The EDs included have a minimum of 6 beds, are 24-h EDs, and range in hospital size and type. Coordinators in each NEISS site collect and record specific variables and create a narrative for each patient seen in the ED for an injury during the initial visit. The NEISS variables include primary diagnosis, body part affected, type of injury, disposition, and location of event. The NEISS reports the primary visit diagnosis and primary body part injured.
Cases of playground equipment-related extremity injury were identified using the following inclusion criteria. First, all cases of playground-related injuries in children aged ≤14 years associated with playground equipment were identified using the six NEISS playground equipment product codes. Subsequently, we selected patients with extremity injuries using the NEISS body part codes. To determine the completeness of our search strategy and identify possible missed cases of playground equipment-related extremity fractures, we queried the narratives of all patients with extremity fractures in the NEISS database who did not have associated playground product codes. We queried the narratives for key words related to playground location and equipment (including playground, slide, monkey bar, seesaw, jungle gym, teeter-totter). By identifying cases via equipment codes, we excluded injuries unrelated to playground equipment, such as falls from standing or sports related injuries. Shoulder injuries were excluded, as the mechanism of these fractures is often different from that of other arm injuries.
We examined body part injured, setting of injury, equipment type, and disposition. Age was categorized based on developmental age groups and body part injured was grouped into lower and upper extremity injuries. Upper extremity injuries included the NEISS codes for upper and lower arm, elbow, wrist, finger, or hand (the NEISS body part codes 80, 33, 32, 34, 82, 92, respectively). Lower extremity injuries included the NEISS codes for foot, knee, ankle, upper and lower leg, and toe (the NEISS body part codes 83, 35, 37, 81, 36, 93, respectively). We categorized setting of injury as place of recreation, home, school or daycare, or other public property. An injury was determined to occur on playground equipment if it was associated with one of the NEISS product codes related to playground equipment. These equipment categories include slides or sliding boards, seesaws or teeterboards, monkey bars or climbing gyms (including playground gyms or other climbing apparatus), swings or swing sets, other playground equipment, and unspecified playground equipment (the NEISS codes 1242, 1243, 1244, 3246, 3219, 3273 respectively). To differentiate monkey bar injuries from other climbing gym injuries, we searched the narratives of all cases of extremity fractures associated with the NEISS product code for monkey bars or climbing gym for the terms “monkeybars” or “monkey bars”.
We combined NEISS disposition codes to categorize disposition from the ED as discharged (treated and released), admitted (treated and admitted, treated and transferred to another hospital, or held for observation) or other (left against medical advice, left without being seen, died, or unknown). Severe extremity fractures were defined as fractures requiring hospital admission.
We used the NEISS sample weighting based on the inverse probability of selection. Sample weights were analyzed via complex survey procedure to obtain weighted average annual numbers of national ED visits and 95% confidence intervals (CIs). US Census Bureau population estimates for children ≤14 years were used to determine national rates of injury. We performed descriptive analyses using SPSS 24.0 (IBM corporation Armonk, NY) and trend analysis using Joinpoint (Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute, Bethesda, MD) linear weighted regression. For all analyses, an overall unweighted count of < 20 cases or a weighted estimate < 1200 were considered possibly unstable and noted when reported. A logistic regression was performed, accounting for the complex weighted survey design, to determine the association of equipment type with extremity fracture, adjusting for age as a known confounder.