Table 4.

MSMs of the number of hospitalizations

	M1 Poisson			M2 Negative Binomial			M3 - Zero-Inflated Poisson
	RR	[95% CI]		RR	[95% CI]		RR	[95% CI]
ᾱ = (0, 0, 1)	1.660	0.725	3.802	1.657	0.725	3.788	2.184	0.846	5.636
ᾱ = (0, 1, 0)	2.165*	1.201	3.901	2.177**	1.209	3.921	0.645	0.357	1.166
ᾱ = (0, 1, 1)	1.030	0.784	1.352	1.029	0.784	1.351	1.178	0.780	1.777
ᾱ = (1, 0, 0)	1.217**	1.056	1.404	1.218**	1.056	1.405	0.980	0.756	1.272
ᾱ = (1, 0, 1)	1.492**	1.268	1.755	1.493**	1.269	1.757	1.215	0.944	1.565
ᾱ = (1, 1, 0)	1.360**	1.211	1.528	1.361**	1.212	1.529	1.181	0.939	1.485
ᾱ = (1, 1, 1)	1.024	0.988	1.062	1.024	0.988	1.062	1.016	0.952	1.084
ᾱ = (0, 0, 1)							1.455**	1.132	1.870
ᾱ = (0, 1, 0)							7.50e–8**	5.69e-9	9.89e-7
ᾱ = (0, 1, 1)							1.206	0.763	1.906
ᾱ = (1, 0, 0)							0.722	0.491	1.061
ᾱ = (1, 0, 1)							0.735	0.498	1.087
ᾱ = (1, 1, 0)							0.811	0.613	1.073
ᾱ = (1, 1, 1)							0.988	0.907	1.076
No. Obs.	199,682			199,682			199,682
QIC	1.478			1.389			1.399
BIC	−2.224e+06			−2.16E+06			−2.16E+06
log-likelihood	−147552			−138717			−139683
p	<0.001			<0.001			0.108
Expected no. of zeros+	171,823			178,243			178,195

All regressions adjust for propensity score and marginal structural weights. For the ZIP, the top panel indicates the Poisson model for the counts, while the bottom panel indicates the parameters for the any/no outcome component of the model using a logistic function

⁺Actual number of zeros is 177,762.

^*p<.01,

^**p<.001 based on Wald tests