Table 2.

Results from the linear regression. Compared to full-scope-of-practice states, states with restricted or reduced scope-of-practice had significantly higher NP-led HBPC visits per 1,000 FFS Medicare beneficiaries.

	Crude model		Adjusted model
	Beta coefficient (95% CI)	p-value	Beta coefficient (95% CI)	p-value
Full- scope-of-practice states	Reference		Reference
Reduced scope-of-practice states	0.74 (0.01–1.46)	0.046	0.92 (0.13–1.72)	0.023
Restricted- scope-of-practice states	1.15 (0.36–1.94)	0.005	0.91 (0.03–1.79)	0.043

Note: The adjusted model controlled for the following state-level covariates: median household income, proportion of population who were Black, Hispanic, and the proportion of those living below the poverty line who were 65 years or older. The outcome of the model was the natural-log transformed HBPC utilization rate (visits per 1000 FFS Medicare beneficiaries). In the log scale, the beta coefficient is the difference in the expected geometric means of the log of HBPC rates between the full-scope- of-practice and reduced- or restricted-scope-of-practice states. An example of interpreting the beta coefficient in the original scale is to exponentiate the coefficient and subtract one. For example, exp(0.91) – 1 =1.46, which can be interpreted as switching from full-scope- of- practice states to restricted- scope- of- practice states, we expect to see approximately 146% increase in the geometric mean of HBPC utilization rate.