GRANDPA: GeneRAtive network sampling using degree and property augmentation applied to the analysis of partially confidential healthcare networks

Applied Network Science

Table 3 Logistic Regression results for modelling if a physician’s specialty was classified as a care coordinator (hospitalists and internal medicine) or other specialist given the percentage of neighbours with the same specialty across the original generated networks

	Intercept			Same specialty			AIC
	Odds ratio	95% CI	p value	Odds ratio	95% CI	p value	AIC
Original	0.063	(0.033, 0.111)	< 0.001	1.116	(1.096, 1.142)	< 0.001	138.61
PGM (specialty and degree)	< 0.001	(0.000, 0.004)	< 0.001	1.386	(1.218, 1.937)	0.001	17.10
GRANDPA (specialty, community and degree)	0.004	(< 0.001, 0.017)	< 0.001	1.203	(1.152, 1.286)	< 0.001	42.88
GRANDPA (specialty, community, degree and linchpin)	0.0235	(0.009, 0.053)	< 0.001	1.139	(1.113, 1.174)	< 0.001	93.98