Taking missing data into account in everyday practice.

Abstract

Ideally, health professionals would have all the necessary information at their disposal for making healthcare decisions, but in reality they lack a great deal of clinically relevant data. Some of these missing data simply cannot, or cannot yet, exist. In other cases, the data are missing because the healthcare professional was unable to find them even though they are available, because they have not been generated even though they could have been, because they have not been published, or because they have been concealed. In most cases, when data on medi- cal interventions are suspected or known to be missing, this can be taken into account by assuming that the supposed benefits of the interven- tion are overestimated, while its harms are underestimated. When several drugs appear to have a similar harm-benefit balance, it is better to choose those for which fewer data are missing, or those which have the longest clinical use.