What is confounding by indication?

What is confounding by indication?

• A. When treatment choice is related to the underlying condition, confounding the treatment effect.
• B. When geographic location influences outcomes.
• C. When sample size confounds the p-value.
• D. When age confounds the exposure measurement.

Answer: (A)

Confounding by indication happens when the reason a patient is given a treatment is tied to their underlying health status or prognosis. Because doctors prescribe therapies based on how sick someone is or how likely they are to benefit, the groups receiving different treatments already differ in baseline risk. When you compare outcomes between these groups, the apparent treatment effect can reflect those preexisting differences rather than the true effect of the treatment itself. This is especially a problem in observational studies where treatment isn’t randomly assigned. For example, sicker patients might be more likely to receive a new or aggressive therapy; if those patients have worse outcomes, it could look like the therapy is harmful even if it isn’t. Randomized trials help prevent this by balancing indications across groups, and in nonrandomized studies, methods like propensity scoring or careful adjustment can mitigate the bias. Geographic location, sample size, and age represent other issues but do not describe treatment decisions tied to a patient’s indication in the same way.

Confounding by indication happens when the reason a patient is given a treatment is tied to their underlying health status or prognosis. Because doctors prescribe therapies based on how sick someone is or how likely they are to benefit, the groups receiving different treatments already differ in baseline risk. When you compare outcomes between these groups, the apparent treatment effect can reflect those preexisting differences rather than the true effect of the treatment itself. This is especially a problem in observational studies where treatment isn’t randomly assigned. For example, sicker patients might be more likely to receive a new or aggressive therapy; if those patients have worse outcomes, it could look like the therapy is harmful even if it isn’t. Randomized trials help prevent this by balancing indications across groups, and in nonrandomized studies, methods like propensity scoring or careful adjustment can mitigate the bias. Geographic location, sample size, and age represent other issues but do not describe treatment decisions tied to a patient’s indication in the same way.