# Table 7 Results of the seemingly unrelated regression of CHE: Haiti, 2012, 2013, household Level (end of document)

2012 2013 Difference (2013–2012) Test (chi2)
Odds Ratio (OR) Standard Error (Std. Err) Odds Ratio (OR) Standard Error (Std. Er)
Quintile (Poorest)
Poorer 0.77 0.164 0.59* 0.12 0.79
Middle 0.83 0.183 0.42*** 0.10 4.60*
Richer 0.72 0.181 0.30*** 0.07 6.32*
Richest *0.54 0.156 0.18*** 0.06 6.01*
Having children < 4 y (yes = 1; otherwise = 0) 1.09 0.183 0.91 0.16 0.54
Literate (1 = literate; 0 = otherwise) 1.35 0.232 1.42* 0.25 0.04
Having older household > 65 y (yes = 1; otherwise = 0) 1.47* 0.257 2.04*** 0.35 1.78
Gender (1 = women; 0 = men) 0.98 0.152 0.81 0.12 0.74
Household size 1.03 0.036 1.19*** 0.04 9.30**
Region (North)
South 1.54 0.38 1.25 0.31 0.34
Transversal 1.31 0.33 1.22 0.32 0.04
West 1.12 0.30 0.75 0.21 1.12
Metropolitan 0.95 0.24 1.19 0.31 0.37
Urban (1 = living in urban area; 0 = rural area) 1.04 0.22 1.19 0.25 0.19
Health system variables (2013)
Health Insurance (yes = 1; otherwise = 0)    2.53* 1.19
Public facilities (yes = 1; otherwise = 0)    3.83*** 0.85
Private facilities (yes = 1; otherwise = 0)    10.45*** 2.47
CHW (yes = 1; otherwise = 0)    0.29* 0.20
Traditional Healer (yes = 1; otherwise = 0)    1.91 1.26
Other and ancillary services (yes = 1; otherwise = 0)    1.08 0.53
Constant 0.07 0.02 0.04 0.01
1. Note: Each model had 2282 observationsa. * p < 0.05; ** p < 0.01, *** p < 0.001
2. aThe 2012 ECVMAS I had a sample size of 4,930 households which were representative at the department and national levels [15]. The 2013 ECVMAS II was a panel survey with a sample size of 2,282 households. These 2,282 households are the same households included in ECVMAS I’s larger sample of 4,930. The SUR model utilized these same 2,282 households from the 2012 and 2013 surveys to run the analysis