How the SFR was developed
The SFR was created by orthopedic surgeons to fill the gap in knowledge relating to the treatment of fractures. The process of defining the variables to be included started in 2007. The number of variables in any register that aims to include all fractures has to be limited. The main outcome measure, as in most other registers studying surgical interventions, was chosen to be reoperation rates. Registration of the reoperations subdivided by reasons for the reoperation will cover most complications as deep infection, mal-union, non-union etc. Each chosen variable must add valuable information. Otherwise it has to be discarded because of the risk of non-compliance if the process of data entry becomes too time-consuming. The structure of the register was finalized in 2009, when the new competence center in Gothenburg offered its support to the founders. After a year of close collaboration between system developers, project managers and orthopedic surgeons, a beta version was launched on 1 January 2011 by the Department of Orthopedics and Trauma at Sahlgrenska University Hospital. Fractures of the tibia and humerus were entered during the trial period. The long bones, shoulder, pelvis and foot were included in April 2012, followed by the hand in October 2012 and the spine in February 2015.
The SFR is run by a national board that has members representing various parts of the country, orthopedic departments, specialties and academic disciplines. The board is supervised by a director who is responsible for maintaining and developing the register. The Swedish Orthopedic Trauma Society, a section of the Swedish Orthopedic Association, is the professional organization that provides support. Funding comes from the Western Healthcare Region and the Swedish Association of Local Authorities and Regions.
Inclusion and exclusion criteria
The SFR collects data on all extremity, pelvic and spine fractures in adults who have been diagnosed or treated at affiliated departments. Data entry requires the patient to have a permanent Swedish personal identity number, be 16 years of age or older and have a fracture diagnosed on the basis of radiographs, Computed Tomography (CT scan) or Magnetic Resonance Imaging (MRI). Pediatric fractures are in the process of being added. Swetrau and other Swedish national quality registers focus on major trauma, whereas various other orthopedic registers specialize in non-skeletal injuries [12, 13].
Technical description
The fully web based SFR is built on the new Stratum platform, designed specifically for health quality registers. Apart from the PROM questionnaires, no paper is used in the data entry process. The system provides users with input choices based on previously entered data, thereby speeding up the process and minimizing the risk of error, and permits the consecutive entry of new injuries, treatment and follow-up, including PROM. The vision from the outset has been to ensure user friendliness and an intuitive interface.
The data entry process
The data entry process consists of four different color-coded steps (Figs. 1, 2, 3, 4 and 5). The first three steps are performed by the physician, while the fourth step contains PROM. Data entry is performed by the attending physician, normally a specialist or resident in orthopedics and trauma, or by others who are on call at accident and emergency departments. They log in with a personal service identification card and a Personal Identification Number (PIN) code. The Swedish Personal Data Act mandates the two-step process. The patient’s personal identity number, an eight-digit date of birth and a unique four-digit control code, is then entered. The number is verified online with the Swedish Population Register and a new file is created if the number is correct.
The diagnosing physician enters the date and the cause of the injury and information about the fracture(s), generally classified in accordance with Arbeitsgemeinschaft für Osteosynthesefragen (AO)/Orthopaedic Trauma Association (OTA) [14, 15]. Classification is based on the available radiological information. A detailed description of the classification process appears later in this article. Treatment is entered once it has been performed. If the fracture is treated non-surgically, the physician on call enters this. If the fracture is treated surgically, the surgeon enters the data. If surgery is performed secondary to non-surgical treatment that has failed, the entire sequence of events is recorded. Scheduled secondary procedures are distinguished from reoperations, which are entered, along with the indications for which they were performed. The illustrations in Figs. 1, 2, 3, 4 and 5 show the data entry process. Data are stored under the responsibility of the County council and on computer servers run by the University of Gothenburg.
Classifying the fracture
A fracture must be analyzed and described before it can be correctly treated. A classification system is essential to the success of a fracture register. Maurice E. Müller argues that classification is useful only if it considers the severity of the bone lesion and serves as a basis for treatment and for evaluating the results [16]. A classification system suitable for a fracture register should ideally be comprehensive, widely recognized, extensively employed, user friendly and valid. No current classification system meets all these criteria. However, AO/OTA is the best available option and we have chosen to use it whenever feasible and meaningful [14, 15]. It has been adapted to the demands of the register and to the online features that are particularly useful for pelvic, acetabular and forearm fractures. For example, pelvic fractures can be assigned an ICD code based on the individual fracture components indicated on a pelvic overview in the first step, with the AO/OTA code (instability pattern) generated in the next step. Similarly, proximal radius and ulna fractures are classified for each bone separately, ultimately linking them together. Acetabular fractures are classified in accordance with both AO/OTA and Letournel [17]. Hip fractures are classified in accordance with AO/OTA, but the descriptions associated with the diagrams refer to the Garden classification of cervical hip fractures [18]. In the same way, proximal humerus fractures are assigned an AO/OTA class, but the description parallels Neer’s terminology [19]. Clavicle fractures are classified in accordance with Robinson and scapula fractures are classified in accordance with Euler and Rüedi and Ideberg, whereas foot fractures are assigned a modified OTA code [20–22].
Patient-reported outcome measures
Shortly after primary treatment at the hospital, two questionnaires are sent to the patient: the Euroqol 5 dimensions 3 level (EQ-5D-3L) and the Short Musculoskeletal Function Assessment (SMFA) [23–25]. The advantage of the EQ-5D is its widespread use by other registers and health-economic analyses, while, on the other hand, the SMFA permits a greater understanding of the patient’s musculoskeletal function. The questionnaires relate to health status and level of functioning the week before the fracture and are answered using recall technique. Responding patients are sent identical questionnaires one year later to assess their recovery. The completed questionnaires are electronically scanned centrally at the SFR. Their calculated indices are displayed in a fourth (grey) panel, without the responses to individual questions. Data are presented as a bother index and five different dysfunction indices. A few additional questions, such as smoking habits, have been added.
The implementation process
Virtually all fractures in Sweden are treated at 55 publicly funded hospitals. All orthopedic and trauma departments at Swedish hospitals have been offered the opportunity to participate. The participating departments have received at least one visit by the register staff to aid the start-up. It normally takes 3–4 months for logistical issues to be ironed out and data entry to begin. As of 31 December 2014, 26 hospitals enter fracture data on a regular basis.
Completeness
The register has a search function to identify incomplete data relating to the injury, fracture, treatment or PROM. Each department is free to incorporate methods of its own in order to ensure the most complete data possible. At the Sahlgrenska University Hospital structured searches in the digital medical records are being made. Each week the medical records are scanned for ICD-codes related to fractures. These search results are matched for entries in the SFR and the fractures that have not been registered in the SFR are secondarily registered. In this way all patients who have a fracture diagnosis in the medical chart are registered in the SFR. Some small departments that treat only a few fractures a day can simply check their daily radiograph demonstration.
Retrieval of data from the register
A national health quality register has a better chance of succeeding and fulfilling its purpose if it is able to provide relevant real-time information to its users. Since the SFR relies on surgeons to find the time to enter data, the surgeon should preferably recognize the usefulness of the information in order to be fully motivated. Moreover, the medical directors need to appreciate the value of the information to promote the department’s participation. With this in mind, we have developed a dozen search functions that enable easy retrieval of data from the register. By combining different parameters, the user is able to obtain immediate, up-to-date statistics for both the particular department and the register as a whole. The data are presented in such a way that it is easy to save tables and diagrams. Another useful feature is the ability to create a search list for a particular type of fracture, gender, age group or treatment during a specified period at the user’s department. The register returns a list of personal identity numbers much more readily than most hospital databases are capable of doing. This function greatly facilitates internal quality control and targeted follow-up.
Ethics
The Swedish Fracture Register is approved by the Swedish Data Inspection Board and operates in accordance with Swedish legislation i.e. the Swedish Personal Data Act and the Swedish Patient Data Act. All patients are informed that the registration takes place and that they have the right to decline, however according to Swedish legislation (the Swedish Patient Data Act) national quality registers do not need signed consent from the individual registered patient. The research conducted in the SFR was approved by the Central Ethical Review Board, Gothenburg. For research purposes, approval by the Regional Ethical Review Board has to be obtained prior to use of register data. If a single center intend to use data registered at their own center for local quality improvement, no such approval is needed. Centers are still encouraged to seek ethical approval as this will facilitate publishing of the results in peer reviewed scientific journals.