Evidence-based medicine has recently gained significant prominence in the field of orthopaedic surgery as well as in other areas of medicine. Numerous manuscripts and textbooks in this field have been published and a detailed review of all evidence-based medicine principles is beyond the scope of this chapter. One of the key aspects of evidence-based medicine is the introduction of a hierarchical rating system for the level of evidence whereby the level of evidence is grading the quality of the overall study design. In this context, a higher level of evidence suggests a lower risk of bias. Most rating systems for the level of evidence of therapeutic studies (i.e. the majority of orthopaedic trauma outcome studies) use a five level scale including level 1 (randomized clinical trial), level 2 (prospective cohort study or poor quality randomized clinical trial), level 3 (case control study), level 4 (case series), and level 5 (expert opinion) [1]. Most major orthopaedic journals have adapted this five-level hierarchical rating system and grade the published articles accordingly. When assessing the clinical impact of outcome studies, the hierarchical grading system for the level of evidence plays an important role.

While this rating system provides the reader with important information on potential bias, the level of evidence should also be used cautiously. First, the level of evidence only provides an overall assessment of the study design and further critical assessment of the study methods and study results is necessary. Second, randomized clinical trials are not always possible for each clinical scenario in particular in the orthopaedic trauma population. For instance, the Lower Extremity Assessment Project (LEAP) was designed to evaluate the outcomes of mangled lower extremity injuries to assess lower limb amputation versus salvage [2, 3]. This well-designed study was performed in a non-randomized fashion as randomizing patients with mangled lower extremities into limb salvage versus amputation would not appear feasible [2, 3]. Finally, it must be emphasized that no single study can provide a definitive answer to a study question. Clinical treatment algorithms in orthopaedic trauma should be based on a composite assessment of the entire literature and should consider all levels of evidence from level 1 (randomized clinical trial) to level 5 (expert opinion).

The outcome measure is another important variable when assessing the validity of an orthopaedic trauma outcome study. In the orthopaedic trauma literature, numerous outcome scoring systems have been used [4]. In general, outcome measures can be divided into clinician-based and patient-reported outcome measures. Standardized outcome measures may focus on general health, body region-specific function, or disease-specific function. As of today, no general recommendations exist as to which outcome measures should be used in orthopaedic trauma outcome studies. Well-designed outcome studies provide outcome data on the patient’s general health in addition to a body region- or disease-specific questionnaire. When using more than one outcome measure, it is crucial to identify the main outcome measure of the study. The main outcome measure should be according to the main hypothesis that is being tested in the study. Another important consideration is whether the used outcome measure has been validated in prior investigations. An outcome instrument is considered valid if it truly measures what it is supposed to measure. In this context, it is important to emphasize that validation of an outcome measure is not an “all or nothing” concept and validity has several components (e.g. face validity, criterion validity, construct validity, content validity, etc.). A detailed discussion of outcome measure validation procedures is beyond the scope of this chapter. In general, the validity of an outcome measure is typically established by comparison between the tested outcome measure and an established outcome instrument. For instance, the short musculoskeletal function assessment (SMFA) questionnaire has been established for the use of trauma patients, and validation studies used the Medical Outcomes Study 36-Item Short Form (SF-36), a well-established and validated outcome measure, as a reference to test the validity of the SMFA [5].

As of today, the SF-36 can be considered to be one of the most commonly employed outcome instruments in orthopaedic trauma surgery as well as in orthopaedic surgery in general [6]. The SF-36 is a patient-reported outcome measure and measures the overall patient’s quality of life. It consists of 36 items and the questionnaire can be completed within 5–10 min. It has been validated and used in numerous studies within the field of orthopaedic surgery as well as in other fields of medicine [7–14]. The SF-36 also allows the comparison of outcomes with normative population data from age- and sex-matched controls. Moreover, the SF-36 has been translated and validated in multiple languages and international normative data have been recorded. Despite these favourable characteristics, the SF-36 also has some limitations that can affect the interpretation of outcome data. First, the items of the SF-36 tend to focus more on lower extremity function than on upper extremity function [15]. This emphasizes the importance of including a body region-specific questionnaire along with a general health questionnaire when performing clinical outcome research in orthopaedic trauma. Moreover, the SF-36 does not incorporate certain basic quality of life domains, such as sexual function or sleep. In some instances, this may lead to the scenario that improvements as well as diminishments in these areas may go undetected. For example, patients undergoing treatment of pelvic fractures may frequently be impaired by sexual dysfunction and recording of SF-36 data may be limited by “ceiling effects” as well as “floor effects”.

Besides the assessment of clinical limitations, healthcare utilization and treatment costs represent important outcome measures when evaluating the efficiency of orthopaedic trauma care. In this context, it is important to emphasize that most importantly treatment should be chosen based on the best patient’s interest and according to highest level of standard of care. However, in the treatment of patients with musculoskeletal injuries, the orthopaedic trauma surgeon is also mandated to make fiscally sound decisions since cost-efficient treatment is above all in the best interest of the injured patient. The question of cost effectiveness becomes specifically important in areas of complex surgeries and when the effectiveness of treatment remains uncertain. In the orthopaedic trauma literature, a frequently discussed topic is the cost effectiveness of limb salvage versus amputation in patients with mangled lower extremities. A more detailed discussion on the clinical aspects of this topic will be provided in the Sects. 29.5. As of today, the question of limb salvage versus amputation remains controversial and the multiple medical, social, and economical aspects need to be considered when discussing treatment plans with the injured patient. In brief, patients with mangled lower extremities face the situation that attempted limb salvage may offer them the undoubted benefits of keeping their lower limb. In order to achieve this favourable result, these patients may undergo several reconstructive surgical procedures and repeat hospitalizations with the remaining risk of requiring an amputation at a later time point. In contrast, primary amputation may offer the potential benefits of quicker discharge from the hospital, earlier ambulation after prosthesis fitting is completed, and earlier return to work. For these reasons, the medical outcomes of limb salvage versus amputation need to be assessed carefully and this sensitive issue needs to be discussed thoroughly with the patients and their families. Importantly, the financial aspects of treatment have to be included in this discussion as it is clearly in the patient’s best interest to be educated about the costs that will incur from medical treatment, hospitalizations, time away from work, as well as lifetime costs from ongoing prostheses needs. Recent investigations have focused on cost utility analyses of amputation versus limb salvage in patients with mangled lower extremities emphasizing the importance of costs as a critical outcome measure [16].

A critical assessment of study data also requires a careful evaluation of the clinical follow-up that was obtained in the study presented. In an outcome study reporting on patients’ recoveries after treatment of extremity injuries, the presented study data may be flawed if subjects who received treatment are not included in the data analysis due to lack of follow-up data. This lack of outcome data may both overestimate as well as underestimate the benefit of treatment effects depending on the outcome of patients not returning for follow-up. Hypothetically, “best case scenarios” and “worst case scenarios” could occur. Thus, patients who did not recover well from their injuries may be upset about their outcome and chose to receive follow-up treatment at a different institution (“I am upset. I am not going back”). If a large number of these patients do not get enrolled in the outcome analysis, the recorded outcomes may be better than the actual real outcomes that have been achieved with the treatment rendered. Vice versa, patients who achieved an excellent recovery potentially may decide to skip their follow-up since they may not feel the necessity to seek any further evaluations (“I feel fine. Why bother?”). If a large number of these patients do not get enrolled in the outcome analysis, the recorded outcomes may be worse than the actual real outcomes of the treatment rendered. For these reasons, any remarkable loss of follow-up carries the risk of skewing the study data and a critical assessment of the study data needs to include the assessment of the loss of follow-up.

Current guidelines of major orthopaedic journals request that any randomized controlled trial with more than 20 % loss of follow-up should be downgraded from an evidence level 1 to an evidence level 2 study [www.​jbjs.​org]. However, these recommendations are based on traditional postulations and it remains unclear how much loss of follow-up can be considered as acceptable. Recent statistical models using trauma databases have pointed out that even less than 20 % loss of follow-up may frequently yield in a significant change of study results [17]. For these reasons, authors of orthopaedic trauma outcome studies should not only report their loss of follow-up, but should also report which specific attempts were made to minimize loss of follow-up and should record the data available on those patients who did not comply with their final follow-up examinations.

Besides the loss of follow-up, orthopaedic trauma outcome studies need to be assessed for their length of follow-up. Patients with extremity injuries go through different phases in their rehabilitation process. Along the different phases of the recovery process, different outcome variables can be recorded. The immediate postoperative period provides information on early complications, such as surgical site infections, mortality rates, thromboembolic events, and length of hospital stay. Within the first few months after surgery further information, such as fracture healing and return to work, can be recorded. Furthermore, long-term outcome studies provide valuable information on the functional recovery, health-related quality of life, as well as the incidence of posttraumatic arthritis and the need for late reconstructive procedures. Many guidelines have recommended 2-year outcome evaluations for patients with extremity injuries. However, these postulations have recently been challenged. In patients with mangled lower extremities, comparisons between 1-year follow-up data and 2-year follow-up data have shown that 1-year follow-up data provide sufficient information to test the pertinent study hypotheses while creating significantly less costs than 2-year follow-up evaluations [18]. Current recommendations from the major orthopaedic trauma journal request 6 months of follow-up for pure soft tissue injuries, 1-year patients follow-up for fracture care, and 2-year follow-up data for treatment of arthritic conditions [http://​journals.​lww.​com/​jorthotrauma]. These issues emphasize the significant implications of the length of follow-up as an important variable for assessing the validity of outcome data. The length of follow-up provides valuable information as to which specific outcome measures can be addressed in outcome studies on patients with extremity injuries.