Intermediate Outcome

What It Is

In epidemiology, an intermediate outcome is an event that occurs within the series of steps of a proposed causal pathway that leads from an initiating exposure to an ultimate outcome of interest.  In an (oversimplified) example, smoking tobacco increases activity of the sympathetic nervous system which leads to high blood pressure, which then leads to a heart attack.  An intermediate outcome on this pathway is high blood pressure.

Why It’s Important

A researcher needs to be clear about his proposed causal pathway, because it affects how the statistical analysis should be done.  If the researcher is interested in whether smoking tobacco can contribute to the development of a heart attack, he should not “control” for high blood pressure. [“Control” here means to use statistical methods to remove the influence of high blood pressure from the potential relationship between smoking and heart attack.]  If he did, his analysis might conclude that “smoking has no effect on the risk of heart attack, it’s the high blood pressure that is the cause.”  Since smoking does increase blood pressure, that conclusion is not really valid.  Even if the statistical methods still indicate that smoking increases the risk of a heart attack after controlling for high blood pressure, the researcher “over-corrected” for the influence of high blood pressure.  That is, by “over-correcting” for high blood pressure, the researcher underestimated the strength of the relationship between smoking and heart attack risk.