Dengue disease surveillance has been notoriously difficult in the past, with differences between reported and estimated cases being very large; since the quality of available data is generally poor, case estimates vary widely [1], [2]. While it is clear that the incidence of dengue is on the increase worldwide, it is difficult to determine how much of this increase should be attributed to improvements in surveillance systems versus increased transmission and disease burden. Further complicating the interpretation of surveillance data are (1) differences in laboratory confirmation rates, (2) dynamics in treatment-seeking behaviour, and (3) changes in case definitions and classifications. Laboratory confirmation of cases is often limited [3] because of cost and the requirement for technical expertise. Although there are an increasing number of rapid diagnostic tests available, their sensitivity and specificity can vary dramatically from those reported by the manufacturer and across endemic settings [4], [5]. There is a need for diagnostic tests that are cheaper and easier to use than those that are currently available. However, even given accurate diagnostics, studies of health-seeking behaviour dynamics are urgently needed to interpret case counts. One factor influencing health-seeking behaviour is disease severity; while the majority of dengue infections are thought to be asymptomatic [2], they are nonetheless important because of their contribution to transmission. Case definitions and case classifications for dengue have varied [6], complicating the interpretation of surveillance data collected before and after the change; although the new case definitions and classifications [1] may have improved patient management, their impact on the study of the biology of dengue disease remains controversial. In summary, only few dengue cases are diagnosed and even fewer are confirmed with a diagnostic test—for these reasons the true burden of disease must rely on estimates rather than counts of reported dengue cases.