Modelled estimates are an integral and essential part of compiling international statistics. Many countries, often but not exclusively, low and middle income are unable to supply estimates for a range of statistical indicators. These data gaps are often filled using modelled estimates. While this solves one problem, it can introduce others. First and foremost, modelled estimates may not take into account the particularities or idiosyncrasies of a particular country and thus may not give an accurate representation of an indicator. Secondly, modelled estimates can be harder to explain, and defend, especially if the estimate is unpalatable. Finally, from a broader communication perspective, modelled estimates can paper of the cracks, disguising just how weak some statistical systems are. Using examples from WHO, this presentation will discuss the use and shortcomings of modelled estimates for decision making.
