Journal of Environmental and Public Health / 2011 / Article / Tab 3

Research Article

Distance Traveled and Cross-State Commuting to Opioid Treatment Programs in the United States

Table 3

Multilevel logistic regression model predicting patient travel across state to OTP.

Predictor95% Confidence interval of the unadjusted odds ratio95% Confidence interval of the adjusted odds ratio
LowerOdds ratioUpperLowerOdds ratioUpper

Block one
 U.S. Region
  Southeast versus Northeast0.3142.51620.1400.1261.0518.737
  Midwest versus Northeast1.04411.826133.9700.8218.47387.475
  West versus Northeast0.0030.0631.4800.0020.0380.839
  250  versus Metro 1.64910.95972.8481.87812.08177.697
  <250 K versus Metro >1 M0.1051.38418.1600.2303.14842.997

Block two
 Patient ZIP code area0.9210.9831.0490.9180.9801.046
 Program ZIP code area1.0242.0073.9331.3172.9286.510

Block three
  30–43 versus 18–290.9431.1151.3200.9501.1251.332
  43–81 versus 18–290.8911.1261.4230.9001.1421.448
  Hispanic versus white0.2620.5591.1920.2660.5701.220
  Black versus white0.1980.4160.8750.1910.4020.847
  Other versus white0.1960.4030.8300.1950.4030.829

Block four
 Pain a reason for treatment0.8821.0321.2090.8801.0321.211
 First methadone treatment0.7220.8490.9970.7340.8681.026
 Strong urge to use0.6670.8611.1120.6760.8951.185
 Severe withdrawal0.7780.9341.1230.8010.9781.194
 Prescription opioid use in past 30 days0.6540.9881.4930.6020.9351.450
 Heroin use in past 30 days0.8541.0641.3270.8581.0841.369

Patient ZIP code area and program ZIP code area are continuous measures; all other covariates are coded 0, 1.