Table 2.

Logistic regression model* of factors affecting the decision to request a malaria slide.

	Antimalarial with no slide request (n = 132)	Slide request (n = 214)	Initial adjusted OR	P	Final adjusted OR	P
Age 15 years or over, n (%)	47(36%)	139(65%)	4.03	<0.01	3.86	<0.01
Male (%)	58(44%)	78(37%)	1.01	0.16
Distance from village/km†: median (IQR)‡	5(5–25) kms	10(5–25) kms	1.02	0.78
Hours to travel to clinic: median (IQR)	1(1–2) hrs	2(1–2) hrs	1.95	<0.01	1.77	<0.01
Socio-economic score§: mean	0.97	1.05	1.16	0.26
Number of days ill: median (IQR)	3(2–4) days	4(3–7) days	0.97	0.15
Antimalarial used in previous 48 hrs: n (%)	3(2.3%)	8(3.8%)	0.90	0.89
Fever in previous 48 hrs: n (%)	85(64%)	149(70%)	2.23	0.03	1.72	0.03
'Body pain' in previous 48 hrs: n (%)	92(70%)	168(79%)	1.06	0.84
Cough in previous 48 hrs: n (%)	71(54%)	97(46%)	0.93	0.79
Diarrhoea or vomiting in prev. 48 hrs: n (%)	55(43%)	82(38%)	0.87	0.60
Other symptom in previous 48 hrs: n (%)	33(25%)	46(22%)	0.66	0.18

*Logistic regression model with dependant variable = slide requested (1) or not requested (0), independent variables in model: aged 15 years or over, socio-economic score, distance to village, hours of travel to clinic, number of days ill, use of an antimalarial in 48 hrs before attendance, a report in the previous 48 hours of: fever, body pain, cough, diarrhoea or vomiting, or 'other' symptom (not specified). Variables were eliminated in a backwards fashion if p >= 0.1. All binary variables used the negative response as the baseline odds (i.e. OR = 1).

^†Estimated distance between hospital outpatient clinic and village of residence.

^‡IQR = Interquartile range

^§Socio-economic score was derived from principal components analysis of reported possession in the household of the following: car, motorbike, radio, mobile phone, television, or refrigerator.