Skip to main content
. 2019 Oct 21;20:140. doi: 10.1186/s12875-019-1031-7

Table 3.

Bivariate and multivariate multinomial regression modeling of treatment engagement predicted by 30-day follow up scores (n = 305)

Model X2 Nagelkerke r2 Counseling + Behavioral Interventiona Medication or Meds+Counselinga
β Wald t OR 95%CI lower 95%CI upper β Wald t OR 95%CI lower 95%CI upper
Bivariatea
 PHQ-9 34.98*** .128 −.025 .525 .975 .912 1.043 .102 8.122** 1.108 1.032 1.188
 DKM 11.23** .043 .253 7.589** 1.288 1.076 1.542 .343 11.235** 1.409 1.153 1.723
 SCMHC 2.684 .010 −.386 1.375 .680 .356 1.296 −.678 2.717+ .508 .227 1.137
 LSAS .677 .003 .015 .037 1.015 .874 1.178 −.025 .098 .975 .835 1.139
 SD .367 .001 .020 .058 1.020 .869 1.197 .047 .290 1.048 .883 1.243
Multivariateab 49.305*** .177
 PHQ-9 .252 7.210** 1.286 1.070 1.546 .403 14.066*** 1.497 1.212 1.848
 DKM −.011 .093 .989 .924 1.059 .123 10.964** 1.131 1.051 1.216

+ p < .10

** p < .01

*** p < .001

aReference category is no treatment

bBoth PHQ-9 and DKM scores entered into same model in a multivariate model