Continued adherence to community-based health insurance scheme in two districts of northeast Ethiopia: application of accelerated failure time shared frailty models

Table 3 Multivariable analysis using the Weibull inverse-Gaussian shared frailty model on predictors of time to drop out

Variables	Categories		Coef	S. E		δ	p-value	95% CI for δ
Intercept			1.356	0.205		3.879	0.000	(2.598, 5.793)
Age in years	25–44		ref
	45–64		-0.014	0.076		0.986	0.850	(0.849, 1.145)
	65 +		0.172	0.119		1.188	0.148	(0.941, 1.501)
Gender	Male		ref
Gender	Female		-0.063	0.124		0.939	0.611	(0.735, 1.198)
Marital status	Divorced/widowed		ref
Marital status	Married		0.476	0.143		1.610	0.001	(1.216, 2.130)
Household size	Smaller (< 5)		ref
Household size	Larger (≥ 5)		0.155	0.072		1.168	0.032	(1.013, 1.346)
Self-rated health	Fair		ref
	Good		0.037	0.108		1.037	0.736	(0.839, 1.282)
	Very good		-0.081	0.113		0.922	0.471	(0.739, 1.150)
Chronic illness	No		ref
Chronic illness	Yes		0.353	0.102		1.424	0.001	(1.165, 1.740)
Hospitalization	No		ref
Hospitalization	Yes		0.267	0.080		1.306	0.001	(1.118, 1.527)
Perceived quality of health care	Low		ref
	Medium		0.127	0.085		1.135	0.135	(0.961, 1.340)
	Haigh		0.279	0.093		1.322	0.003	(1.100, 1.587)
Perceived risk protection	Low		ref
	Medium		0.030	0.094		1.031	0.748	(0.857, 1.240)
	Haigh		0.197	0.087		1.218	0.023	(1.027, 1.444)
Trust in scheme	Low		ref
	Medium		0.274	0.088		1.315	0.002	(1.107, 1.563)
	Haigh		0.549	0.098		1.731	0.000	(1.428, 2.098)
		ln (ρ) = 0.453 (p < 0.001)			γ = 0.636 (S.E = 0.028)
		ρ = 1.573 (S.E = 0.069)			τ = 0.059
		θ = 0.126 (S.E = 0.066)			AIC = 1779.13, BIC = 1871.23

CI Confidence Interval, coef regression coefficient, S.E Standard error; δ Acceleration Factor, ρ Shape parameter, γ Scale parameter, γ = 1/ρ; θ – Variance of the random effect, τ Kendall’s tau, τ = θ/θ + 2, where τ = ϵ (0, 1), ref reference category; AIC Akaike’s Information Criterion, BIC Bayesian information criterion

ISSN: 1475-9276