Group-based trajectory models of integrated vaccine delivery and equity in low- and middle-income countries

Table 2 Group-based trajectory model diagnostics

	Geographic Equity (non-imputed model)	Geographic Equity (imputed model)	Socioeconomic Equity (non-imputed model)	Socioeconomic Equity (imputed model)
Model Diagnostics
Bayesian information criterion	-3786.16	-44042.75	109.11	1089.08
Entropy	0.951	0.896	0.701	0.735
Average posterior probability of group assignment	Group 1: 0.995 Group 2: 0.967 Group 3: 0.962 Group 4: 0.974 Group 5: 0.959	Group 1: 0.998 Group 2: 0.941 Group 3: 0.928 Group 4: 0.978 Group 5: 0.932	Group 1: 0.783 Group 2: 0.838 Group 3: 0.919 Group 4: 0.887	Group 1: 0.908 Group 2: 0.935
Odds of correct group classification	Group 1: 1419.48 Group 2: 135.58 Group 3: 70.24 Group 4: 91.29 Group 5: 159.73	Group 1: 2992.39 Group 2: 72.95 Group 3: 34.12 Group 4: 117.37 Group 5: 73.70	Group 1: 3.99 Group 2: 14.98 Group 3: 37.65 Group 4: 195.76	Group 1: 3.45 Group 2: 41.16

This table presents several diagnostics for each of our group-based trajectory models. Per Nagin, Bayesian information criterion values and entropy values should be as large as possible, average posterior probabilities should be at least 0.7, and the odds of correct classification should ideally be 5 or greater

ISSN: 1475-9276