For any type of evaluation, you need some kind of rubric or grading system to standardize and document your decision-making process.
In a risk of bias assessment, we typically do this by using three “levels”:
As you review each study for bias, you’ll use these labels.
|Level of risk of bias||What it means|
|Any difference between the study result and the “true” result is likely to be small.|
|Any difference between the study result and the “true” result could be large.|
|We don’t have enough information to decide either way.|
But bias is such a broad term. How do you decide if a study is low, high or unclear?
You should never label an entire study as having high, low, or unclear level of risk of bias. That’s because “bias” can take many forms. A study might be at a high risk for one type of bias, but at a low or unclear risk for another. Or, different parts of a study, e.g., outcomes, may be at different levels of risk for the same type of bias.
Let’s get specific about the types of bias you should look for.