Table 2.

Relationship of “need” (operationalized as the prevalence of newly diagnosed cases of injection-related AIDS per 10,000 residents), racial/ethnic composition, and poverty rates to temporal trajectories in spatial access to pharmacies enrolled in the Expanded Syringe Access Demonstration Program (ESAP) in New York City health districts over time (2001–2006)

Predictor	Model A	Model B	Model C	Model D
Fixed effects
Intercept	12.71***	12.74***	16.48***	16.42***
Year	1.32***	1.31***	1.68***	1.68***
Need		1.56	10.23***	10.43***
Need × year		0.43*	1.25***	1.25***
Percent of residents who are Non-Hispanic white			38.56***	37.56***
Percent of residents who are non-Hispanic white × Year			3.61***	3.61***
Percent of residents who are non-Hispanic white × Need			20.05**	19.72*
Percent of residents who are non-Hispanic white × need × Year			1.94**	1.94**
Poverty rate				−5.39
Random effects
Variance among the intercepts	168.92***	165.76***	96.98***	96.52***
Correlation of intercepts and slopes	0.96***	0.99***	0.99***	0.99***
Variance among the slopes	1.77***	1.48***	0.87***	0.87***
Ar(1)a	0.60*	0.66*	0.66*	0.66*
Residual variance	1.50*	1.78	1.76	1.79

*p < 0.05; **p < 0.001; ***p < 0.0001

^aWithin district covariance, assuming a first-order autoregressive error structure