In recent years, dementia has been considered a public health priority and become a topic of major political interest. Recent reviews and studies have reported with varying degrees of alarm an impending and existing "dementia epidemic" with increasing predicted trends in prevalence and enormous numbers of people with dementia particularly in low- and middle-income countries (LMICs). However, robust evidence from dementia research in high-income countries suggests stable or decreased prevalence over the last decades. Current evidence is not sufficient to suggest increasing trends of prevalence in LMICs once variation in methodological factors and study populations are taken into account. Changes in diagnostic methods over the last decades substantially influence the identification of dementia cases with systematic difference between the resulting individual prevalence studies. Potential geographical variations at the country level might indicate potential risk factors at population levels or systematic difference in clinical application of dementia diagnosis. Although it is important and necessary to use information from dementia research for evidence-based policymaking, over-interpretation of results without carefully considering underlying factors could exaggerate the findings and influence policy planning in ways which do not serve current and future population best. Planning of dementia policy needs to take full cognisance of the provenance of the data being used and be integrated with policies which optimise health across the lifecourse.