The paper by Hu et al. on the spatial incidence of dengue infections in Queensland, Australia [Reference Hu1], has a serious, indeed fatal, flaw. There has not been any recognized locally acquired dengue in Queensland south of the city of Townsville (in the north of the state) since the mid-1950s when dengue transmission occurred in Rockhampton [Reference Kay2, Reference Hanna and Ritchie3]. Therefore, the maps (Figures 3–5) in the paper that indicate dengue incidence, inferring local transmission of dengue south of Townsville, are not only quite incorrect but also misleading. Unfortunately this misleading inference has already been cited in an Australian Government report on the likely impact of climate change in Australia [Reference Hughes and McMichael4], and consequently that report should be revised.

The error has occurred because Hu et al. have not carefully checked the data supplied to them by Queensland Health. A travel history is obtained for each notification of dengue in Queensland so that a determination on the place of acquisition can be made. The dataset supplied by Queensland Health contained numbers of dengue cases ‘acquired locally and overseas’ [Reference Hu1]. However, Hu et al. ignored this information and failed to either separate locally acquired cases from imported cases or to determine where the locally acquired cases were actually acquired, and included all cases to estimate local incidences based on residential postcodes. This failing has led to the quite erroneous conclusion that ‘dengue transmission had already begun to extend south-easternwards in Queensland between 1993 and 2004’ [Reference Hu1]; it hadn't, and indeed still has not.