Investigating patients’ preferences for new anti-diabetic drugs to inform public health insurance coverage decisions: a discrete choice experiment in China

Table 4 Estimation of preference heterogeneity from interaction-effects models

Interaction terms	Age (Model 1)		Income (Model 2)		Diabetes Complications (Model 3)
Interaction terms	Mean	SE	Mean	SE	Mean	SE
Improvement in HbA_1c
As expected	0.005	0.121	0.542^***	0.136	0.152	0.141
Serious side effects
Occasionally	0.358^**	0.125	-0.001	0.124	0.184	0.129
Never or rarely	-0.333^**	0.116	0.025	0.117	-0.193	0.124
Extended life
Extended life (per year)	0.030	0.122	0.122	0.132	0.404^**	0.141
Change in HRQoL
No improvement	0.178	0.131	-0.103	0.139	-0.208	0.142
Improvement	0.249	0.184	0.145	0.172	0.402^*	0.200
Dosing frequency
Once every other day	0.076	0.121	0.126	0.123	0.106	0.126
Once a week	0.027	0.128	0.160	0.126	-0.195	0.133
Out-of-pocket costs per month
Cost (per CNY50)	-0.107^**	0.040	-0.175^***	0.042	-0.156^***	0.043
Log likelihood	-2586.191		-2580.122		-2586.387
Observations	10,224		10,224		10,224

Age, income, and diabetes complications were treated as categorical variables in interaction-effects models. Age: Young and middle-aged = 0, Elderly = 1; Income: More than CNY 6000 = 0, CNY 6000 and below = 1; Diabetes complications: Without = 0, With = 1; SE, standard error; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001

ISSN: 1471-2458