Adjusting for geographic variation in observational comparative effectiveness studies: a case study of antipsychotics using state Medicaid data

Table 1 Variance Components and Intra-class Correlation (ICC) for Random Effects Models

	Var	SE	p-value	ICC
Mental-health Related Hospitalizations
Unadjusted, MHSA Random Effect	0.034	0.017	<0.0001	1.0%
Adjusted, MHSA Random Effect	0.017	0.014	0.033	0.5%
Unadjusted, County Random Effect	0.044	0.020	<0.0001	1.3%
Adjusted, County Random Effect	0.023	0.018	0.029	0.7%
Mental-health Related ED Visits
Unadjusted, MHSA Random Effect	0.071	0.027	<0.0001	2.1%
Adjusted, MHSA Random Effect	0.027	0.016	0.001	0.8%
Unadjusted, County Random Effect	0.070	0.023	<0.0001	2.1%
Adjusted, County Random Effect	0.037	0.019	0.001	1.1%

ISSN: 1472-6963