Table 5 Result of discrete choice experiment regarding criteria for prioritizing the HIV response, main effects model, disaggregated by agency
From: Criteria for prioritization of HIV programs in Viet Nam: a discrete choice experiment
Dependent variable (0 = if the basic profile is chosen; 1 = if an alternative profile is chosen) | ||||||
|---|---|---|---|---|---|---|
Dvlp. Partners | Government | Civil Society | ||||
Criteria | β | std.err. | β | std.err. | β | std.err. |
Independent variables (coding under variables) | ||||||
Δ Effectiveness | 2.021* | 0.776 | 0.848 | 0.515 | −0.356 | 0.632 |
0 = no change, remains 4 million years of life saved | ||||||
1 = increases, 5 instead of 4 million years of life saved | ||||||
Δ Cost-effectiveness | 0.081 | 0.729 | 0.413 | 0.467 | −0.061 | 0.545 |
0 = no change, remains $6 return for every $1 invested | ||||||
-1 = decreases, $5 instead of $6 return for every $1 invested | ||||||
Δ Sustainability | 0.028 | 0.736 | 1.561* | 0.533 | 2.125* | 0.685 |
0 = no change, spending increases and then decreases | ||||||
-1 = changes to spending constantly increases | ||||||
Δ Treatment/Prevention | −0.277 | 0.569 | 1.190* | 0.503 | 1.140* | 0.640 |
0 = no change, greater investment in prevention | ||||||
-1 = changes to greater investment in treatment | ||||||
Δ Feasibility | 3.113* | 0.828 | 1.514* | 0.492 | 2.202* | 0.641 |
0 = no change, remains low feasibility | ||||||
1 = increases, high feasibility instead of low feasibility | ||||||
Constant | −1.073 | 1.465 | 1.414 | 1.173 | 1.355 | 1.345 |
Number of observations | 147 | 182 | 140 | |||
Number of groups | 21 | 26 | 20 | |||
Log likelihood function | −60.46 | −90.12 | −66.92 | |||
Wald chi2 | 21.63 | 27.93 | 22.18 | |||
prob > chi2 | 0.001 | 0.000 | 0.001 | |||