Table 7.

Ecological models for tick-borne pathogens in host blood and ticks from infested dogs

Covariate	A. Platys: Ta, Na, Ni, Gh		H. canis: all countries		E. canis: Ta, Na, Ke, Ni, Gh		C. burnetti (tick only): Ke, Ug, Ni
	Blood	Tick	Blood	Tick	Blood	Tick
Age (months)					− 0.38 ± 0.16*
Body condition				− 0.47 ± 0.17**
Tick loads
R. sanguineus	0.42 ± 0.12***			0.33 ± 0.11**	0.25 ± 0.10**
Rhipicephalus sp.
H. leachi					− 0.48 ± 0.25*		0.44 ± 0.15**
Haemaphysalis spp.				0.42 ± 0.18*			0.47 ± 0.15**
Deworminga
< 1 month	− 1.55 ± 0.44***		− 1.26 ± 0.31***		− 0.74 ± 0.36***
1–6 months	− 1.58 ± 0.38***		− 1.25 ± 0.32***		− 0.36 ± 0.29***
> 6 months	− 1.10 ± 0.43**		− 1.00 ± 0.38**		− 0.68 ± 0.40**
Pathogen in blood tissueb
Yes–No		4.00 ± 0.30***		1.41 ± 0.24***		4.35 ± 0.33***	3.02 ± 1.12**

Parameter estimates (± empirical standard error) from the logistic regressions (GEEs) that model the pathogen prevalence (levels: 0, 1) in host blood and the ticks. Only countries for which at least one area had a prevalence > 10% were included. The main assumption here is that pathogen in the blood is driven by vector presence (proxy: ticks found on dogs) and the dog’s physiology; therefore, extrinsic characteristics that correct for tick presence (urban vs rural; housing conditions; dogs around) are not included. We assume macro-geographic variation in pathogen (wildlife) reservoirs at the country level; therefore ‘country’ remained in all of the models.

Sex of dog, dogs in the environment and ectoparasiticide treatment did not significantly explain any of the variation, and therefore are not shown in the table

***P < 0.001, **P < 0.01, *P < 0.05 

^aContrasts with group of dogs that have never been treated with a deworming drug

^bOnly included in the analyses on pathogens in feeding ticks