The National Electronic Injury Surveillance System (NEISS) database contains the most comprehensive surveillance of injuries treated in the emergency department [10]. The NEISS database is produced by the Consumer Product Safety Commission (CPSC). It includes data from 100 hospitals in the USA selected and weighted to represent the average US population to allow estimates of specific injury types presented to all US emergency departments. A full description of the sample design is published on the CPSC website [6]. The NEISS database has been a reliable source of data for many epidemiological studies [14, 15, 17].

The CPCS NEISS database is a probability sample of 100 emergency departments in the USA. These hospitals are selected and weighted to represent the average US population by accounting for geographic location, department size, and children’s hospitals. Each year, a sampling frame based on all emergency department visits in the USA is used to determine statistical weights for the current NEISS hospital sample that provide estimates of the entire US population [6]. A NEISS query for a specific injury provides both the number of cases presenting to the 100 emergency departments in the probability sample as well as the estimate for the entire US population. Based on these estimates, we calculated incidence rates using estimates of the US population from the US Census Bureau, Population division [16].

At participating emergency departments, a case record is made of each injury that includes treatment date, age, sex, race, diagnostic category, body region (e.g., wrist, shoulder), patient disposition (e.g., treated and released, treated and admitted, etc.), location of injury, and two descriptive narrative fields. These data are directly keyed in computer systems by staff at each participating hospitals [7, 11].

In this study, the NEISS database was queried for all injuries in the upper extremity between 01 January 2009 and 31 December 2009. The NEISS database uses a body part diagram. The upper extremity includes the following regions: finger, wrist, lower arm, elbow, upper arm, and shoulder. In the NEISS estimates query builder, three parameters are entered: treatment dates, product codes, and other parameters. Treatment date was between 01 January 2009 and 31 December 2009. The product code section (each diagnosis is coded) was not filled in, and the section with other parameters was queried for each body part of the upper extremity separately in the Body Part section and then put in a common database. The specific diagnostic categories used in the NEISS database are amputation, contusion/abrasion, crush, dislocation, presence of a foreign body, fracture, hematoma, laceration, nerve damage, puncture, strain or sprain, hemorrhage, avulsion, dermatitis, burns, and other/not stated. Burns can be divided into electrical, scald, chemical, radiation, thermal, and burns not specified, but we considered them as a single group because they were relatively uncommon.

We analyzed the data in several ways. First, we calculated the percentage of total upper extremity injuries that occurred in each body region. Next, for each body region, we calculated the distribution of injury types as a percentage of the total number of injuries for that region. Next, for each injury type, we calculated the distribution among the body regions as a percentage of the total number of that type of injury. Next, we calculated the distribution of the circumstances/location where the injury occurred as a percentage of the total number of injuries and as a percentage of the total number of injuries for each body region. Finally, we calculated incidence rates for each specific type of injury in a specific body part. For the ten most common injuries, we used the Spearman correlation test to see whether age correlated with incidence. To calculate age-specific incidences, we used 5 year age groups to age 105 years.

By anatomic site, the majority were finger injuries (38.4%), followed by shoulder (16.8%), lower arm (between the elbow and the wrist) (15.3%), wrist (15.2%), elbow (10.5%), and upper arm (between the shoulder and the elbow) injuries (3.7%) (Table 1). Specific diagnoses for each body part are also detailed in Table 1.

The most common type of injury of the upper extremity that brings a patient to an emergency department in the USA is a fracture (29.2%). Other frequent diagnoses in the upper extremity are lacerations (23.3%), strains/sprains (16.2%), contusions/abrasions (14.0%), and dislocations (4.6%) (Table 2).

The majority of upper extremity injuries occur at home (45.4%), while 16.2% are sports-related. Other common sites are school (6.6%), other public property (4.1%), and street (2.5%). However, a substantial amount of locations were not recorded (25.1%) (Table 3). In the most common injury types, we found that most fractures occur at home (38.6%). Most lacerations also occur at home (67.1%), and most contusions and abrasions occur at home as well (44.5%).

According to incidence rates, the five most common specific injury (body part and type of injury combined) that brings a person to the emergency room is a laceration of the finger (221 per 100,000 persons per year; or 1 in 452 of us each year), wrist fractures (72 per 100,000 persons per year), finger fractures (68 per 100,000 persons per year), shoulder strains/sprains (68 per 100,000 persons per year), and lower arm fractures (64 per 100,000 persons per year) (Table 4). Incidence rates over the past 9 years of the five most common injuries all show significant change. The following injuries showed decrease: finger lacerations (B=−6.6; p=0.000; R²=0.94), wrist fractures (B=−1.3; p=0.002; R²=0.72), finger fractures (B=−1.7; p=0.000; R²=0.88), and lower arm fractures (B=−1.1; p=0.047; R²=0.41). Shoulder strain/sprain showed an increase (B=2.5; p=0.000; R²=0.86).

Larsen and colleagues found that 1 out of 55 Dutch and 1 out of 28 Danish people presented to emergency departments with hand injuries between 1997 and 1998 [5]. A Dutch study reported that 19% of all fractures presenting to the emergency department of a large university hospital in The Netherlands were fractures of the hand [9]. Prior reported incidences of finger fractures include 380 per 100,000 in the UK [2]. These results are similar to our findings.

Singer and colleagues reviewed 90 million lacerations presenting to an emergency department in the USA between 1992 and 2002 and found that 35% involved the upper extremity [12]; Singer and colleagues [13] looked at fracture incidence in 15,000 adults in Edinburgh between 1992 and 1993 and found patterns consistent with lifestyle and osteoporosis risk.

The NEISS database has been a useful tool for identifying injuries that present to US emergency departments. Zacchilli and colleagues used the NEISS database to look at shoulder dislocations in the USA between 2002 and 2006. They calculated an estimated incidence rate of 23.9 per 100.000 persons. This is approximately twice the previously reported value [17]. The Center for Disease and Control Prevention studied nonfatal traumatic brain injuries from sports and recreational activities and found that most injuries (both for males and females) occurred between the ages of 10–14 years [1]. They advocate increased awareness and development of prevention strategies.

This study should be interpreted in light of its shortcomings. The diagnostic categories are broad, which may decrease the likelihood of inaccuracy, but there is no way to know, for example, if a diagnosis of fracture was sometimes used for a suspected fracture. The NEISS database also uses estimates for incidence rates; this may influence accuracy as well. Further more, we studied the rates of injury presenting to an emergency department, which would underestimate the true incidence of injuries since some patients do not present to the emergency department. Finally, the NEISS database does not account for return visits to the emergency department for the same injury.