Household beliefs about malaria testing and treatment in Western Kenya: the role of health worker adherence to malaria test results

Table 3 Association between test status, ACT, and confidence in ACT

	Outcome: odds of taking ACT		Outcome: odds respondent believed ACT “very likely” effective in treating malaria
	OR	AOR	OR	AOR
	(1)	(2)	(3)	(4)
A. Tested positive for malaria	3.37** [2.21, 5.15]	3.41** [2.23, 5.21]	1.35 [0.76, 2.42]	1.25 [0.63, 2.50]
B. Tested negative for malaria	0.41** [0.27, 0.62]	0.45** [0.28, 0.71]	0.33** [0.17, 0.65]	0.29** [0.13, 0.63]
C. Not tested for malaria	Ref. Group	Ref. Group	Ref. Group	Ref. Group
Includes controls		X		X
Mean of outcome in reference group	0.7	0.7	0.7	0.7
P value: (A = B)	0	0	0	0
Number of observations	1062	1041	818	806

Table shows logistic regression results of the association between test status and ACT use (columns 1 and 2), and beliefs about ACT effectiveness (columns 3 and 4). Columns 3 and 4 are limited to individuals who were treated with ACT. The controls in Columns 2 and 4 include the following: the wealth of the household (defined as the first component from a principal component analysis of household characteristics and assets), the education level of the respondent (no education, some primary education, or some secondary education), the sick individual’s age and gender, and the time it takes for the household to travel to the nearest health facility. All coefficients are expressed in terms of odds ratios and 95% confidence intervals are in brackets. Standard errors are adjusted for clustering by community unit. ** P < 0.01

ISSN: 1475-2875