[Empirical evaluation of quality indicators for severely injured patients in the TraumaRegister DGU®].

BACKGROUND AND

objectiveA systematic assessment of the quality of medical treatment by using key indicators has been required in Germany for many years. These quality indicators (QI) have to satisfy many requirements. Besides an expert review an empirical data-based evaluation is also necessary. The TraumaRegister DGU® (TR-DGU) has reported QI in the annual reports from the beginning. The objective of this study was to validate 40 QI for the treatment of severely injured patients reviewed by experts using data from the TR-DGU. MATERIAL AND

methodsThe association of the 40 QI with hospital mortality was verified using healthcare data from the TR-DGU from a 5‑year period (2012-2016). Of these 26 QI consider events while the remaining 14 QI are key indicators, such as time spent in the trauma room. To compensate differences in injury severity, adjusted mortality rates were calculated using the revised injury severity classification (RISC) II score. For this two different approaches were chosen: the hospital-based approach classifies all hospitals into three categories and analyzes the grade of fulfilment of the indicator. The indicator-based approach considers the adjusted mortality depending on the grade of fulfilment of the indicator.

resultsThe analysis was based on 111,656 cases documented in the TR-DGU (mean age 50 years; 70 % male). The data analysis showed an obvious correlation with mortality for half of the QI, including only three procedural times. A clear correlation in both approaches was shown for two QI: prehospital capnometry in intubated patients and sonography used for patients without whole body computed tomography (CT) scans. Of the 20 QI with a positive result 15 were also positively rated by the experts. Of the 14 QI reported annually since 2017 in the TR-DGU report, 8 (57%) showed a clear correlation with mortality.

conclusionThere is no doubt regarding the necessity of scientifically assessing QI. Approximately half of the evaluated QI showed an empirical association with mortality. Interventions and events showed better results than measurements of procedural times; however, many QI may require a refined definition. The interpretation of the results is still challenging due to differences in the patient groups. Secondary endpoints, such as hospital length of stay and quality of life after trauma were not considered here.