Expert consensus on contamination in educational trials elicited by a Delphi exercise

Area of Science:

• Educational Research
• Clinical Trial Design
• Intervention Science

Background:

• Contamination in clinical trials occurs when control groups receive interventions, biasing results towards no effect and reducing statistical power.
• This issue is particularly relevant in educational intervention trials due to the nature of knowledge and experience transfer.
• Standard methods to reduce contamination, like cluster randomization, can introduce their own biases or reduce statistical power.

Purpose of the Study:

• To identify factors influencing contamination in educational intervention trials.
• To rank methods for avoiding contamination during the planning stages of such trials.
• To provide evidence-based guidance for selecting appropriate trial designs.

Main Methods:

• A Delphi exercise was employed to gather expert opinions on contamination risks.
• Experts were asked to identify factors promoting or inhibiting contamination.
• Experts ranked strategies for contamination avoidance in educational trials.

Main Results:

• Trials in settings with high social, geographical, or professional overlap among participants face the greatest contamination risk.
• Complex interventions aimed at behavioral change are less likely to cause contamination than simple interventions or those focused on knowledge acquisition.
• Interventions with greater impact on behavioral change inherently minimize contamination.

Conclusions:

• The likelihood of contamination in educational studies depends on intervention complexity and target behavior.
• Complex interventions, while potentially more impactful, may offer an advantage in minimizing contamination.
• Findings assist researchers in choosing between trial designs, such as cluster randomized trials, based on contamination risk assessment.