Table 2.

Logistic regression to estimate propensity scores

Dependent variable: non-EG households
Independent variables	Coef.	[95% Conf.	Interval]	Sig.
Total household expenditure	−0.0017	−0.0026	−0.0086	Yes
Distance to commune health center	0.0111	−0.0246	0.0468	No
Distance to district hospital	0.0502	0.0365	0.0639	Yes
Household economic status	−0.0471	−0.1000	0.0056	No
Number of household member	−0.0218	−0.0708	0.0272	No
Number of males in household	0.0389	−0.0353	0.1132	No
Number of children under 6 years in household	0.0603	−0.0555	0.1762	No
Number of elder people over 60 years in household	0.0579	−0.0761	0.1920	No
Number of sick episodes	0.0294	0.0046	0.0543	Yes
Number of sick person	−0.0406	−0.1149	0.0338	No
Number of restricted days caused by sickness	−0.0029	−0.0051	−0.0006	Yes
Number of children under 6 years who were sick	−0.0153	−0.1447	0.1141	No
Number of elder people > 60 years who were sick	−0.0564	−0.2014	0.0886	No
Number of males who were sick	−0.0205	−0.1102	0.0691	No
Age of household head in years	−0.0095	−0.0138	−0.0052	Yes
Sex of household head	−0.0656	−0.2210	0.0899	No
Education of household head	0.0373	0.0230	0.0516	Yes
Religion of household head	−0.4345	−0.6806	−0.1883	Yes
Ethnic of households head	0.1283	−0.0900	0.3467	No
Married status of household head	−0.0427	−0.1009	0.0149	No
Occupation of household head	−0.0063	−0.0155	0.0029	No
Constant	1.6697	1.1617	2.1777	Yes

N=11,260; LR Chi² (21) = 223.10; Prob> Chi² =0.0000.