Abstract
Over the last decade, there has been a growing body of level I evidence supporting non-operative management (focused on early range of motion and weight bearing) of acute Achilles tendon ruptures. Despite this emerging evidence, there have been very few studies evaluating its uptake. Our primary objective was to determine whether the findings from a landmark Canadian trial assessing the optimal management strategy for acute Achilles tendon ruptures influenced the practice patterns of orthopaedic surgeons in Ontario, Canada over a 12-year time period. As a second objective we examined whether patient and provider predictors of surgical repair utilisation differed before and after dissemination of the landmark trial results.
Using provincial health administrative databases, we identified Ontario residents 18 years of age and older with an acute Achilles tendon rupture from April 2002 to March 2014. The proportion of surgically repaired ruptures was calculated for each calendar quarter and year. A time-series analysis using an interventional autoregressive integrated moving average (ARIMA) model was used to determine whether changes in the proportion of surgically repaired ruptures were chronologically related to the dissemination of results from a landmark Canadian trial by Willits et al. (first quarter, 2009). Spline regression was then used to independently identify critical time-points of change in the surgical repair rate to confirm our findings. A multivariate logistic regression model was used to assess for differences in patient and provider predictors of surgical repair utilisation before and after the landmark trial.
From the second quarter of 2002 to the first quarter of 2010 the surgical repair rate remained constant at ∼21%, however, by the first quarter of 2014 it fell to 6.5%. A statistically significant decrease in the rate of surgical repair (P<0.001) was observed after the results from a landmark Canadian trial were presented at a major North American conference (February 2009). Both teaching and non-teaching hospitals demonstrated a decline in the surgical repair rate over the study period, however, only the decrease seen at non-teaching hospitals was found to be significantly associated with the dissemination of landmark trial results (P<0.001). All other predictors of surgical repair utilisation remained unchanged in the before-and-after analysis with the exception of patients 30 years of age and younger having a higher odds of undergoing surgical repair after the trial when compared to those 51 years of age and older.
The current study demonstrates that large, well-designed randomised trials, such as the one conducted by Willits et al. can significantly change the practice patterns of orthopaedic surgeons. Moreover, the decline in surgical repair rate observed at both teaching and non-teaching hospitals suggests both academic and non-academic surgeons readily incorporate high quality evidence in to their practice.
