Table 1.

Parameter estimates from the multiple indicator HMM, the univariate HMM for the total encounters, and the direct Poisson lagged regression model (also based on total encounters).

	Multiple Indicator HMM		Total Encounters HMM		Total Encounters Poisson regression
Parameter	mean	sd	mean	sd	mean	sd
Intercept	-3.48	0.040	-3.43	0.038	-0.23	0.029
lag 1 montha	4.59	0.026	4.59	0.026	0.13	0.009
Alcoholism Treatmentb
After Treatment	-0.07	0.018	-0.07	0.018	-0.06	0.013
During Treatment	0.36	0.030	0.33	0.029	0.25	0.026
gender(male)	-0.12	0.017	-0.12	0.017	-0.13	0.013
Age categoryc
30-39 years	0.18	0.031	0.18	0.029	0.25	0.022
40-49 years	0.24	0.031	0.24	0.029	0.31	0.022
50-59 years	0.41	0.033	0.41	0.032	0.51	0.025
≥ 60 years	0.58	0.043	0.57	0.042	0.72	0.043
Insurance Regiond
West	-0.18	0.028	-0.18	0.027	-0.26	0.022
South	-0.12	0.029	-0.10	0.027	-0.20	0.024
Northeast	-0.13	0.035	-0.13	0.033	-0.08	0.027

^alag 1 month is the previous month’s hidden state for HMM models and is the previous month’s observed count for the Poisson model.

^bbefore treatment is the reference criteria.

^c18-29 years old is the reference category

^dmidwest is the reference category.