dc.description.abstract | This thesis describes a study of the epidemiology and control of rabies in Machakos District
of Kenya. The main objectives were: I) to design and implement a community-based, active
surveillance programme to estimate rabies incidence; 2) to collect dog ecology data required
for planning rabies control; 3) to model the current transmission patterns and predict future
changes under different control scenarios; and 4) to make recommendations on rabies
control in the district based on this information.
Study areas were selected by a stratified random sample with 6 of the 10 administrative
divisions and 1 sublocation per division selected. Five of the six sublocations were largely
rural with low to moderate human densities (47.3-190.2 people krn'), while the other had
a much higher human population density (737.4 people krn"). The sizes of the sublocations
ranged from 8 to 123 km2 and covered a total of379 krrr', which was 2.7% of the district's
land area and an estimated 2.5% of its human population. In each study area, two major
studies were carried out, active surveillance for rabies cases using locally recruited rabies
workers and a survey of randomly sampled households for dog ecology and rabies
vaccination information. The active surveillance programme was conducted by communitybased
rabies workers. Rabies diagnosis was by the fluorescent antibody test (FAT) carried
out at the Central Veterinary Laboratory, Kabete.
During the one-year active surveillance for rabies, a total of369 rabies suspects, 277 animal
and 91 humans, were reported. Of the 7..77animal rabies suspects, 194 were primary and 83
secondary cases. Dogs accounted for the majority of the primary (92% or 179/194) and
secondary (80% or 66/83) animal rabies suspects. Of the 66 dogs secondarily exposed, 63
(95%) were bitten by other dogs. Using dog population estimates for the study sublocations,
this converted to a dog-to-dog exposure rate of 1,230/1 00,000 dogs. Of the 83 exposed
animals, 55% (46/83) were alive at the end of the follow-up period and a further 30%
(15/83) experienced an event directly related to the exposure including: death due to bite
wounds (8), sold (3), killed by their owners (9), and death due to confirmed rabies (5). Of
the 92 human cases of animal-bites, 89 (97%) were bitten by dogs. Using the human
population estimates for the study sublocations, this converted to an annual incidence of
human cases of animal-bites of 1341100,000 people, equivalent to 3,680 human cases of
animal-bites per year for the entire district.
Out of 141 potential animal specimens, 156 were collected for diagnosis. Uncollected
specimens were dogs that had either escaped (33) or had been killed and buried (52) (often
secretly by their owners). A further 10% (16/156) of the collected specimens decomposed
due to freezer failure. Of the specimens submitted, 130 were diagnosed and 10 samples were
lost due to handling errors at the diagnostic laboratory. Dogs accounted for 81% (44/54) of
the confirmed rabid animals for an annual confirmed dog rabies incidence of 900
cases/1 00,000 dogs. One human death due to clinical rabies from a dog bite, was recorded
providing a point estimate of 25 deaths/million persons. Active surveillance uncovered
approximately 40 times more rabies cases than the traditional passive reporting system.
The dog ecology study of 150 households over one year revealed that almost all dogs were
owned hut were not carefully controlled and scavenged for food. The sample of 320 dogs
in these households had a high turnover rate, with half the dogs replaced each year. Overall
life expectancy was estimated to be 2.8-2.9 years; males had a life expectancy of3.5 years
and females 2.4 years. Only 39% of female dogs survived to one year of age. This high
mortality was balanced by high fecundity (1.3 females produced per female per year) so that
the population was estimated to grow by 9 percent per annum (+/- 5%). No female dogs
were spayed and only 15.] % of males were castrated. 28.9% of dogs over 3 months of age
were reportedly vaccinated against rabies. This was close to the 29% of dogs tested with
detectable rabies antibodies .
. A deterministic model of dog rabies was developed with parameters estimated from the
Machakos dog population to describe and predict rabies transmission. Important model
predictions included: 1)there was a threshold dog density (KT) of approximately 5 dogs krn"
below which rabies did not persist; 2) rabies could act as a severe density-dependent
constraint limiting the dog population growth: 3) the predicted equilibrium prevalence of
rabies was low (0.6%); and 4) where rabies persisted, there were two possible patterns -
rabies oscillated with decreasing amplitude if the carrying capacity (K) was not much larger
than KT' but there was no damping ifK was much larger than KT. Model predictions with
respect to vaccination were: 1) that the current vaccination rate (24%) had virtually no effect
on rabies transmission; 2) that for annual vaccination, a vaccination coverage of85% would
be required; and 3) ifsemi-annual vaccination was adopted, a vaccination coverage of45%
would be adequate.
Based on study findings the following recommendations for improved rabies control in
Machakos District are made. First, for control programmes to be effective, public
participation is essential and peoples' needs must be considered. Of particular importance
is linking the supply of post-exposure treatment to rabies surveillance. As far as dog
vaccination coverage is concerned, if annual vaccination is retained, coverage must be
drastically increased by actively involving the community. Otherwise, semi-annual
vaccination could be considered but will be more expensive. Community-based field trials
(particularly involving schoolboys) are the logical next step to investigate the potential
efficacy of different dog vaccination options. Strategies are likely to be different for high
. and low dog-density communities. | en |