Financing COVID-19 vaccination in sub-Saharan Africa: lessons from a nation-wide willingness to pay (WTP) survey in Ghana

Table 3 Backwards stepwise regression on predictors of willingness to pay for COVID-19 vaccine (n = 559)

Willingness to pay	OR	Std.Err	P > z	[95%Conf	Interval]
Elderly male	0.199	0.225	0.153	0.022	1.820
Educated male	0.550	0.114	0.004	0.366	0.826
Adherent to COVID-19 protocols^a	0.887	0.107	0.320	0.700	1.123
Married Christian	0.549	0.173	0.057	0.296	1.018
Perception of vaccine^b	2.404	0.911	0.021	1.144	5.054
Educated elderly	3.178	3.738	0.326	0.317	31.871
Elderly married	5.205	4.892	0.079	0.825	32.844
Elderly Christian	0.373	0.290	0.204	0.082	1.709
Response to COVID-19^c	1.171	0.123	0.134	0.952	1.439
Educated married	2.188	0.791	0.030	1.077	4.445
Elderly Christian	1.422	0.325	0.124	0.908	2.225
Married health worker	0.428	0.149	0.015	0.217	0.845
_cons	0.573	0.298	0.285	0.207	1.589

Legend: ^aindexed score of overall adherence level to COVID-19 protocols on a five-point Likert scale where higher values depict better adherence and vice-versa; ^bindexed score of perception of COVID-19 where higher values depict positive perception and vice-versa; ^cindexed score on perceived government response strategy against COVID-19 on a five-point Likert scale where higher values depict better perceived response and vice-versa
Note: Sample size (n = 559) in the regression model is the valid responses of persons who will accept to take the COVID-19 vaccine and those who did not respond in the affirmative were dropped from the regression model
Step-wise backwards regression beginning with full model;
p = 0.9929 > = 0.3300 removing Male Christian
p = 0.9461 > = 0.3300 removing Educated health worker
p = 0.9266 > = 0.3300 removing Male health worker
p = 0.8447 > = 0.3300 removing Perceived impact of COVID-19
p = 0.6085 > = 0.3300 removing Perception of COVID-19
p = 0.5769 > = 0.3300 removing Christian health worker
p = 0.4735 > = 0.3300 removing Married male
Logistic regression (model specification)
Number of obs. = 559
LR chi2(12)= 31.99
Prob > chi2 = 0.0014
Log likelihood = -369.81995
Pseudo R2 = 0.0415

ISSN: 1471-2458