Table 3.

Performance of the Authors’ Novel Method When Applied to Acute Flaccid Paralysis Data Simulated Under a Homogeneous-Temporal or Heterogeneous-Temporal Model of Oral Poliovirus Vaccine Coverage^a

Vaccination Model and Average Coverage Used in Simulations, %		Best-Fit Model Based on rDIC (% of Simulations)b				Estimated Average Vaccination Coverage, %
Time Period 1 (12 Months)	Time Period 2 (12 Months)	Homogeneous	Heterogeneous	Homogeneous- Temporal	Heterogeneous- Temporal	Time Period 1		Time Period 2
						Median	95%CrI	Median	95%CrI
Homogeneous-Temporal Model
80	40	0	0	98	2	80.1	77.3, 82.9	40.4	37.3, 43.5
80	50	0	0	96	4	80.2	77.3, 83.1	50.3	47.1, 53.5
80	60	0	0	95	5	80.0	77.1, 82.9	60.1	56.8, 63.3
80	70	5	3	89	3	80.4	77.4, 83.2	70.1	66.8, 73.2
80	90	1	0	98	1	79.9	77.0, 82.6	89.9	87.1, 92.5
Heterogeneous-Temporal Modelc
52	28	0	0	0	100	45.1	41.9, 48.3	24.9	22.0, 27.8
52	34	0	0	0	100	44.9	42.1, 47.8	29.8	26.9, 32.8
52	40	0	0	0	100	44.5	41.8, 47.1	34.4	31.6, 37.2
52	46	0	37	31	32	44.5	42.0, 47.0	39.8	37.1, 42.5
52	58	0	13	35	52	44.3	41.8, 46.7	49.3	46.8, 51.7

Abbreviations: CrI, credible interval; rDIC, rescaled deviance information criterion.

^a These models assume that coverage varies between 2 time periods. In addition, in the heterogeneous-temporal model, a proportion of the children with acute flaccid paralysis are assumed to come from an undervaccinated group, as for Table 2 (see Methods for details).

^b Underlined values correspond to the correct model choice.

^c Assuming that 40% of children are in an undervaccinated group with 10% coverage that does not change over time.