Common errors in the design of orthopaedic trials: Has anything changed?

The adoption of evidence-based orthopaedics has shifted the focus from expert base opinions and anecdotal evidence to a focus on integrating the best available clinical research. This shift has led to an increased focus on randomized controlled trials (RCTs) within the field. Although RCTs are considered the highest level of evidence, methodologic errors can introduce bias and limit the validity of the results. Early trials were hampered by lack of blinding, inadequate sample sizes and other design flaws. The objective of this review was to examine the current literature to determine if the design and execution of RCTs has improved.

The awareness of the importance of sample size increased over time with substantially more trials reporting sample size calculations. However, many contemporary RCTs are still underpowered and fail to reach their calculated sample size. Given the challenges of surgically based RCTs, the majority of historical trials lacked blinding, increasing the risk of bias. There is evidence that there has been a concerted effort to increase the blinding in RCTs, particularly in outcome assessors. A more recent development in the design of surgical trials is the introduction of expertise-based trial designs in which patients are randomized to a surgeon with expertise in a particular intervention. These trials minimize the bias that can arise from differential expertise bias and have the potential to improve the validity and feasibility of RCTs. Finally, there has been an increased focus on the reporting of patient reported outcomes (PROs) in orthopaedic RCTs. Alongside this movement has been the development of minimal important differences (MIDs) to define the changes that are relevant and meaningful to patients. Both PROs and MIDs should be taken into consideration when calculating the sample size and study power in clinical trials.

Although marked improvements have been made in the design and implementation of trials, there is still considerable room for improvement. Adequately blinded and powered studies evaluating clinically important outcomes and differences should be key considerations in trial design moving forward.