Author: H Janaka de Silva
Position: Professor of Medicine
Institution: Faculty of Medicine, University of Kelaniya, Ragama, Sri Lanka
E-mail: hjdes@sltnet.lk
Additional Authors: A Kasturiratne, A R Wickremasinghe, N de Silva, N K Gunawardena, A Pathmeswaran, R Premaratna, L Savioloi, D G Lalloo
Submitted Date: December 24, 2008
Published Date: December 31, 2008
This comment was originally posted as a “Reader Response” on the publication date indicated above. All Reader Responses are now available as comments.

We thank Dr Whyte for her interest and acknowledge that our approach of giving low and high estimates based on available data for a country may lead to using values for the estimation that are outliers. We evaluated all available data based on predetermined criteria which included a hierarchy depending on the scientific merit of the publication. Preference was given to published community based national data. Data from other sources were considered depending on the scientific methods used, the size of the study, the area in which the study was conducted etc. In this hierarchy, data that were considered less representative were used only in the absence of superior quality data. As explained in the manuscript, we originally planned to model snakebite incidence and mortality using available snakebite data and ecological, socio-economic and healthcare related variables as predictors. This was not possible because of the small number of high quality estimates and this same problem would have limited the use of weighted averages. Having assessed the quality of the data, we felt that our approach ensured that we could estimate a minimum burden of envenoming and give some indication of the likely upper limit. The methodology that we used was certainly not influenced by attempts to reduce our workload.

It is inevitable that the constraints of the data will lead to considerable variation, such as the high estimate for East Asia being 47.7 times that of the low estimate. China was the only country that had a high and low estimate in that region, and given its size and geographical diversity, variations in rates of snake bites are bound to occur. The large population of China substantially influences the high estimate and examples such as this simply reflects the lack of scientifically valid epidemiological studies of snakebite.

Most of the literature on snake bites only reports envenomings, although this is not always clear from published papers. Dr Whyte suggests a discrepancy in our reporting of total snakebites. However, as stated in our methodology, we used the limited number of studies that reported both venomous and non-venomous bites to estimate that the total number of snake bites is 2-3 times the number of envenomings for countries where there was no data. We acknowledge the fact that more data is required for a more precise estimate.

We were also intrigued by the reduction in snakebites in Africa and India that we observed compared to previous estimates. We maintain that without clear explanation of how the high African estimates were reached, it is very difficult to be sure why ours are so different; we also believe that we have used the best quality data currently available for the Indian estimates. We acknowledge that there may be flaws in our approach, but these have been necessitated by the paucity of high quality data. The main point remains that we need many more high quality epidemiological studies of snakebite and it will be difficult to refine our approach and calculate precise estimates of the burden of snakebite until such data are available.

A Kasturiratne, A R Wickremasinghe, N de Silva, N K Gunawardena, A Pathmeswaran, R Premaratna, L Savioli*, D G Lalloo**, H J de Silva

Faculty of Medicine, University of Kelaniya, P. O. Box 6, Ragama, Sri Lanka, *Department of Control of Neglected Tropical Diseases, World Health Organization, Geneva, Switzerland, and **Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, UK