Patients included in this study were those admitted in a level 1 trauma center from January 2007 until December 2011. The organisation of trauma care in the Netherlands is based on the American model of trauma regionalization. The Netherlands is divided in 11 separate trauma regions, each region contains a level one trauma center .
In this study prospective data from the Dutch National Trauma Database (DNTD) for the area Central Netherlands were used. The DNTD contains documentation on all trauma patients that are treated at the emergency department and subsequently admitted. Data in the DNTD were collected in a standardized manner and include detailed information on demographics, trauma event and mechanism, primary trauma survey, initial treatment and injuries. Injuries were diagnosed at primary survey, subsequent surgery or during admission. Thoracic and pelvic x-ray imaging were performed for all trauma patients and when indicated supplemented with ultrasound and computed tomography (CT). The database accuracy is constantly evaluated by two database managers.
All injuries were coded using Abbreviated Injury Scale (AIS) location codes allocated to one of the six body regions (head and neck, face, chest, abdomen, extremities and external) to calculate the Injury of Severity Score (ISS) . Patients with a clavicle fracture were selected using AIS location codes. The ISS provides an overall score for patients with multiple injuries and is used to determine injury severity; 0 corresponds with no injury, the maximum score of 75 corresponds with injury leading to death . Patients with an ISS ≥ 16, obtained from ≥2 AIS regions and physiological alterations due to the injuries are considered severely injured and were included in our analysis .
For these patients, age, gender, trauma mechanism, injured side, additional injuries, department of admission (Intensive care Unit, Medium Care Unit, Operation Room) and discharge facility were collected from the DNTD. In all patients trauma mechanism was analysed and determined if it was a high energy trauma. The ATLS definition for high energy trauma was used . Furthermore death associated with the trauma was obtained from the electronic patient documentation (EPD).
To evaluate the clavicle fractures we used the imaging studies performed. These radiological tests allowed for clear images of the fracture and of possible dislocation in anterior-posterior or cranial-caudal direction. Fractures were classified by the researchers (JL, SF and MH) using the Robinson classification. This classification divides the clavicle in a medial fifth (type 1), a diaphyseal part (type 2) and a lateral fifth (type 3). This is further divided by three other variables; intra-articular extent, degree of comminution, and degree of displacement .
Mean numbers were noted with standard deviation (SD), median numbers were noted with interquartile range (IQR). Statistical analysis was performed using the χ
2 test for categorical variables and t-test and one-way-ANOVA for continuous variables. Binary logistic regression was used for the calculation of the dependent variables in additional injuries. A p-value of ≤0.05 was considered significant. Data were analyzed with SPSS Version 20.0, Chicago, IL, USA.