On the bias caused by regression toward the mean in studying the relation between change and initial value

Abstract

In studies of treatment effects in clinical trials and longitudinal followup investigations, it is not unusual to inquire as to whether a relationship exists between the change of a variable and its initial value, for example pocket depth Such studies must be carried out with great care, since the results are biassed by the regression towards the mean (RTM) effect. Examples are presented, the magnitude of the RTM effect is estimated and means of dealing with the RTM problem are discussed. RTM also appears as a selection phenomenon, emphasizing the need to include control groups in order to make possible adjustments for the bias caused by RTM.