The distinction between causal, predictive, and descriptive research – there is still room for improvement

It has been proposed that medical research questions can be categorised into three classes: causal, predictive, and descriptive. This distinction was proposed to encourage researchers to think clearly about how study design, analysis, interpretation, and clinical implications should differ according to the type of research question being investigated. This article highlights four common mistakes that remain in observational research regarding the classification of research questions as causal, predictive, or descriptive, and provides suggestions about how they may be rectified. The four common mistakes are (1) Adjustment for "confounders" in predictive and descriptive research, (2) Interpreting "effects" in prediction models, (3) The use of non-specific terminology that does not indicate which class of research question is being investigated, and (4) Prioritising parsimony over confounder adjustment in causal models.