Dog demography and population estimates for rabies control in Bali, Indonesia

Abstract

Rabies is a fatal zoonotic disease with global significance. At the end of 2008, rabies emerged in the Island of Bali, Indonesia, drawing international attention. As the disease became wide-spread, the government focused on island-wide mass vaccination of dogs and improving public awareness, however the local dog population is not well documented. The Center for Indonesian Veterinary Analytical Studies (CIVAS), a local non-government organization in Indonesia, and the International Livestock Research Institute (ILRI) initiated a project to explore the link between the dog population and local communities in Bali with focus on the impact of this relationship in the spread of rabies. As part of that project, the objective of the study is to (1) characterize the demographics and rabies vaccination of owned and free-roaming dogs and (2) estimate the abundance and identify factors associated with the distribution of the dogs in Bali. The study was conducted on two dog subpopulations, owned and free-roaming dogs, in 310 banjars in Denpasar city, Gianyar district, and Karangasem district in Bali. Banjar is a subvillage structure in Bali. The sampling design was a two-stage sampling with villages as the primary sampling unit and banjars as the secondary sampling unit. Data were collected between March 2011 and March 2012. Survey of owned dogs was carried out through door to door interview of owners and photographic mark recapture was used to collect data on free-roaming dogs. Dogs were predominantly owned and the effect of unowned dogs towards the total population was minimal. Demographically, the sex ratio was male-biased and juveniles make up 15-20% of the population. Free-roaming dogs were dominated by adults and a higher proportion of males. There were differences in the demographics of dogs in urban and non-urban areas which should be considered when planning and implementing control programs. Overall vaccination coverage was high (>70%), however juveniles and females have a higher likelihood of not being vaccinated. The endurance of vaccination collars should be improved to better represent the true vaccination coverage in free-roaming dogs as there is high confidence that most free-roaming dogs were actually owned dogs. Recent culling was associated with increased proportions of juveniles and a 40% higher risk of dogs not being vaccinated. The observation of free-roaming dogs should always account for detection probability as only 20% of dogs in this study were seen at any given time. Failure to account for detection probability will result in severe underestimation of the population abundance. The human population, presence of a forest and recent culling accounted for 28% of variation in the number of owned dogs in banjars. Accordingly, the number of owned dogs and presence of rice paddies accounted for 61% of variation in the number of free-roaming dogs in banjars. Finally, the overall and median human to dog ratios were the least biased ratios available for estimating the overall dog population, however it is a crude tool with poor precision. The overall ratio on average slightly underestimates the total population, while the median ratio slightly overestimates the total population. For rabies control purposes, overestimation is preferred over underestimation and the median human to dog ratio is recommended.