Diffusion of subsidized ACTs in accredited drug shops in Tanzania: determinants of stocking and characteristics of early and late adopters

Table 5 Results of optimal multivariate model based on a backward selection procedure

	OR (95% CI)	p
(Intercept)	0.14 (0.09, 0.24)	<.0001
Total number of customers on previous day	1.02 (1.01, 1.02)	<.0002
Percentage of customers appearing for malaria concerns (per 10% increase)	1.19 (0.52, 2.70)	<.0003
Number of ADDOs within 3km radius	1.04 (1.02, 1.07)	0.001
Number of RCH clinics within 3 km	0.89 (0.78, 1.01)	0.065
Round 1	Ref
Round 2	2.72 (1.61, 4.61)	<.0001
Round 3	4.99 (2.96, 8.43)	<.0001
Round 4	22.55 (13.07, 38.90)	<.0001
Round 5	28.41 (16.40, 49.21)	<.0001
Round 5	41.73 (23.78, 73.24)	<.0001
Round 7	49.06 (27.58, 87.26)	<.0001
Mtwara	Ref
Rukwa	0.15 (0.11, 0.20)	<.0001

All variables presented previously were included in the model, and then removed one by one in order of highest p-value until an optimal set of significant variables (with minimum AIC) was reached.

ISSN: 1472-6963