Population-based studies on trauma care: models and measurements of adverse outcome

Abstract

In this thesis we investigated whether population-based studies with routinely collected data are eligible to assess (adverse) outcome after trauma. We used the Dutch trauma registry which was designed in order to get insight into the magnitude of trauma victims in the Netherlands and to measure, evaluate and improve the outcome of the quality of the trauma care. We evaluate outcome prediction models, the effect of several interventions and trauma systems, and the economic aspects of adverse outcome in trauma care.

Several international prediction models, based on the Trauma and Injury Severity Score (TRISS), can be used to calculate the probability of survival after trauma with adjustment for patient characteristics. In Chapter 2 the performance of these prediction models were tested in a single, large cohort. The results described in this chapter demonstrated that there was no single ‘best’ model. The discriminative power decreases in groups of patients with severe traumatic brain injury and those of older age with an isolated hip fracture. In Chapter 3 the results showed that the trauma care in the Dutch county Noord-Brabant is comparable with international norms. In Chapter 3 the admission policy for trauma patients in Noord-Brabant was evaluated also. Because there was no funding for centralisation of trauma care, the different chain partners in Noord-Brabant agreed to continue after the assignment of the trauma centre the existing policy of transporting trauma patients to the nearest hospital. After stabilisation, patients are further transferred to the trauma centre when necessary. The hypothesis that there is no difference in risk of hospital death between major trauma patients transferred from another hospital to the trauma centre and those directly admitted to the trauma centre was confirmed. In Chapter 4 we analysed the effect of the helicopter emergency medical services (HEMS) on mortality. The patients with severe traumatic brain injury had a non-significant association with a higher risk of dying when they were treated by the HEMS at the accident scene. Our study suggests that the higher mortality risk for patients with severe traumatic brain injury is caused by this increased prehospital time. For patients without severe brain injury the increased time did not have to seem a negative effect. Short distances and a high number of well equipped hospitals in the county could be an explanation for the results in Chapter 3 and 4.

The results of Chapter 5 show that diagnosis-related complications increase the costs of hospital care, with adjustment for age and injury severity, with a total monetary amount of € 5 420. Chapter 6 of is the basic step in the development of a model to predict with patient characteristics the probabilities of absence of various types of complications. Separate models for diagnosis- and institution-related complications were developed. Using these formulas the expected admissions without complications can be calculated and compared with those observed in order to measure and compare the quality of hospital care given to different trauma populations.