Table 4. Time-series regression results with the association of 76% three-dose PCV13 coverage on pneumonia outcomes highlighted.

	All clinical pneumonia a
Model and covariate a	IRR (95%CI);
	p-value
3dose (B1)	1.469 (0.865, 2.493);
	p = 0.154
L1.dres (B2)	1.015 (1.004, 1.025);
	p = 0.008
	Fast breathing pneumonia a	Chest indrawing pneumonia a	Danger sign pneumonia a	Hypoxemic pneumonia a	Mortality a	Proportion danger sign a
Model and covariate a	IRR (95%CI);	IRR (95%CI);	IRR (95%CI);	IRR (95%CI);	IRR (95%CI);	OR (95%CI);
	p-value	p-value	p-value	p-value	p-value	p-value
3dose (B1)	2.347 (1.389, 3.966);	1.393 (0.725, 2.678);	0.636 (0.313, 1.295);	0.532 (0.299, 0.945);	0.643 (0.416, 0.994);	0.351 (0.230, 0.536);
	p = 0.001	p = 0.320	p = 0.212	p = 0.031	p = 0.047	p<0.0001
L1.dres (B2)	1.009 (0.990, 1.028);	1.026 (1.008, 1.043);	1.042 (1.025, 1.059);	1.055 (1.018, 1.094);	1.029 (0.932, 1.135);	1.060 (1.041, 1.078);
	p = 0.379	p = 0.003	p<0.0001	p = 0.004	p = 0.572	p<0.0001

IRR = Incidence Rate Ratio; OR = Odds Ratio

^a 3dose is the population three-dose PCV13 coverage, scaled from 0 to 1 so that 1 = 76% to give an estimated effect (B₁) at 76% coverage, the level reached in the post-PCV13 period; L1.dres is the deviance residual for the previous month in an identical model without the dres term in it, which was added to adjust for residual autocorrelation in the time-series

Observations = 29; these are the 30 calendar months in time (January 2012 to June 2014) minus one due to the inclusion of the one-month lagged residual term L1.dres