The most important findings in the present study were that evidence-based measures for preventing SSI during anesthetic care were not sufficiently implemented Furthermore, differences in the quality of care appear to exist between patients undergoing TJA and patients undergoing FS. These differences cannot be justified, especially since we know that fracture patients are more susceptible to infection. Orthopedic trauma patients suffer from preoperative soft-tissue and skeletal damage, along with co-morbidities and minimal opportunities for preoperative optimization, which have been shown to be major risk factors for this group of patients . A trend towards higher ASA classification scores, which are per se associated with a higher risk of SSI, was also demonstrated in our study . An overall risk assessment of the trauma patient should lead to meticulously applied risk reduction measures during anesthetic care. Among hip and knee arthroplasty surgeons, there is a strong tradition of research on how SSI can be prevented [22–24] and the quality of care is thoroughly monitored . The national PRISS project could also be seen as a reflection of this interest. However, in the area of orthopedic trauma surgery, there are more limited data on preventive measures and risk factors , along with reports on relatively high infection rates, 4.2% , 5.2% , 6.9% . This high SSI rate could at least partly be explained by differences in the quality of care in relation to infection control observed in this study between TJA and FS.
We found that more favorable conditions were created for TJA patients during surgery. They were all operated on in operating rooms equipped with laminar airflow systems, designed to reduce the number of colony forming units (CFU) to well below 5/m3. Fracture patients, on the other hand, had their procedures performed in displacement-ventilated ORs (91.4%). A recent study carried out in the same displacement –ventilated ORs and based on 116 active air samples demonstrated that the mean CFU/m3 values exceeded the recommended levels for orthopedic surgery, < 10 CFU/m3 (m = 15.9, SD 13.4 CI 13.1-18.7) . One of the basic prerequisites for safe surgery in orthopedics is optimal air quality [23, 29, 30]. The dispersal of particles from the individuals present in the OR is considered to be the most important source of airborne contamination and, for this reason, the non-scrubbed staff can reduce airborne contamination by observing the correct clothing regimen and by wearing surgical hoods that cover all their hair [31–33]. In 14 of 66 procedures, it was observed that OR staff had hair hanging outside the surgical hood, a fact that can adversely affect air quality and, as a result, patient safety.
Systematic reviews strongly support the importance of the optimal timing of antibiotic prophylaxis in relation to TJA, as well as fracture surgery, stating that, for every 13 patients who are treated, one wound infection would be prevented [34, 35]. In the present study, only 47% of the patients received prophylaxis within the recommended time span. Similar results (45-57%) have been reported by Stefansdottir et al. . In eight cases, other types of prophylaxis then Cloxacillin were administrated. This raises the question on if it is manageable in clinical practice to have different guidelines depending on type of prophylaxis and their half-life. One interesting observation in the present study was that none of the patients in the TJA group had a major violation of the recommended timing, i.e. received prophylaxis after incision or the application of a tourniquet, whereas 10 of 29 patients undergoing fracture surgery had their antibiotics after the start of surgery. The timely administration of prophylactic antibiotics is of the utmost importance, as a study of 1992 patients undergoing total hip arthroplasty showed that those who received prophylaxis after incision had the highest odds of developing an SSI . The WHO checklist did, in fact, function as an important reminder, but, as we discovered, the checklist per se is not a guarantee of safety; it is instead the way we react to mistakes or lapses that finally matters.
Clear evidence has been presented of the relationship between SSI and mild hypothermia and accordingly the protective effect of normothermia during surgery [6, 7]. The clinical setting in our hospital, with fairly cold ambient air (19-21°C) in combination with the patient’s impaired thermoregulatory system caused by regional or general anesthesia , supports the use of an active patient warming system. Even mild perioperative hypothermia has been shown to produce a series of adverse effects in patients undergoing surgery. It is associated with an increased risk of blood loss and blood transfusion , as well as a risk of increased cardiac morbidity , altered drug metabolism  and prolonged hospitalization . Questions have been raised whether these warming systems could actually be vectors of infection, but studies have shown that this is not the case [42, 43].
Urinary tract infection (UTI) is the most common healthcare-associated infection and a frequently observed complication after major joint surgery . In hospital settings, almost all these infections develop as a result of urinary tract catheterizations . It has been demonstrated that catheter-related UTI contributes to an increased length of stay, costs, morbidity and excessive antimicrobial drug use . However, the management of the UTC and length of time it is used, influences the development of a UTI. We found that the use of UTC increased with increasing ASA-classification score, which is not surprising as this reflects the patients’ health status. In patients with an ASA score of 3 or 4, the use of UTC is not only justified but also most frequently necessary. Even so, on the basis of our results, we draw the conclusion that more could be done to avoid its use in healthy patients, when the estimated length of surgery does not exceed 2.5 h. However, the most worrying finding was the poor compliance with the practice of using an aseptic insertion technique. In 10 of 11 directly observed insertions of UTC, the OR staff did not perform hand disinfection before the insertion and, in 6 of 11 cases, they did not even do so after the insertion. These results are linked to poor adherence (11.9%) to hand disinfection guidelines, resulting in bacterial transmissions observed in the OR. Recent studies in the UK  and the USA [47, 48] presenting similar results indicates that this is an international problem that needs to be resolved. The reasons behind low adherence to the different clinical guidelines in this study is in line with consisting findings of the gap between evidence and practice in health care . Producing standard protocols and guidelines will not per se result in enhanced patient safety  Hence; the complexity of implementing guidelines and behavior change should not be underestimated as adoption of a guideline depends on many different factors. Obstacles for successful implementation could be found on individual, structural and cultural levels. In addition, we also need to take in account the many different and competing demands health care professionals meet in every day practice . By extracting knowledge from the implementation science, it is possible that we could gain deeper insight in how to select the appropriate strategies for implementation of guidelines in the surgical environment.
Observational studies could be susceptible to bias . Human perceptual errors could affect the information that is obtained, together with behavioral distortion due to the presence of an observer. Several measures were taken to address potential bias. Firstly, the observational form was pre-tested and modified, secondly, the observer had no prior connection with the ward under observation and, thirdly, the observer underwent self-training sessions to maximize accuracy. The staff was also blinded to exactly what was being observed. Concealed observations to reduce reactivity were not feasible and were also considered to be a possible source of distrust between the OR staff and the observer.
One limitation of this study was that comparisons between groups were not included in the initial study protocol, resulting in an estimated statistical power of 75%.