Quality improvement (QI) analytic methodology is rarely encountered in the emergency medicine literature. We sought to comparatively apply QI design and analysis techniques to an existing data set, and discuss these techniques as an alternative to standard research methodology for evaluating a change in a process of care.

We used data from a previously published randomized controlled trial on triage-nurse initiated radiography using the Ottawa ankle rules (OAR). QI analytic tools were applied to the data set from this study and evaluated comparatively against the original standard research methodology.

The original study concluded that triage nurse-initiated radiographs led to a statistically significant decrease in mean emergency department length of stay. Using QI analytic methodology, we applied control charts and interpreted the results using established methods that preserved the time sequence of the data. This analysis found a compelling signal of a positive treatment effect that would have been identified after the enrolment of 58% of the original study sample, and in the 6th month of this 11-month study.

Our comparative analysis demonstrates some of the potential benefits of QI analytic methodology. We found that had this approach been used in the original study, insights regarding the benefits of nurse-initiated radiography using the OAR would have been achieved earlier, and thus potentially at a lower cost. In situations where the overarching aim is to accelerate implementation of practice improvement to benefit future patients, we believe that increased consideration should be given to the use of QI analytic methodology.

To determine the effect of triage nurse initiated radiographs using the Ottawa Ankle Rules (OAR) on emergency department (ED) throughput. We hypothesized OAR use would reduce median ED length of stay (LOS) by 25 minutes or more.

A randomized controlled trial was conducted at a tertiary centre ED with an annual census of over 90,000 patients. Adult patients presenting within 10 days of isolated blunt ankle trauma were eligible. Participants were randomly assigned to standard triage or OAR application by 15 explicitly trained triage nurses. Our primary outcome was ED LOS. Secondary outcomes included triage nurses' and patients’ satisfaction. A power calculation indicated 142 patients were required. The Mann-Whitney U test was used to compare the medians between the two groups.

Of 176 patients with blunt ankle injury screened, 146 were enrolled (83.0%); baseline characteristics in the two groups were similar. The median/mean ED LOS in the control and OAR groups were 128/143 minutes and 108/115 minutes respectively (median difference 20 minutes; p=0.003). Agreement in OAR use between emergency physicians and nurses was moderate (kappa 0.46/0.77 for foot/ankle rule components), and satisfaction of both nurses and participants was high.

Triage nurse initiated radiography using OAR leads to a statistically significant decrease of 20 minutes in the median ED LOS at a tertiary care centre. The overall impact of implementing such a process is likely site-specific, and the decision to do so should involve consideration of the local context.