Associations of the Korean patient placement criteria matching among individuals with alcohol-related problems with treatment completion and abstinence: an observational study

Table 4 Association of KPPC-matched treatment with the number of months of treatment completion with and without adjustment for confounders (n = 225)

Predictor variable and category vs reference category	Odds ratio [95% CI]	p	Fit statistics
Predictor variable and category vs reference category	Odds ratio [95% CI]	p	AIC	SC	−2LL
No. of KPPC-matched treatments
Unadjusted model^a			541.98	559.06	531.98
Once vs none	3.54 [1.61–7.76]	< 0.01
Twice vs none	3.76 [1.87–7.53]	< 0.001
Adjusted model^b
Once vs none	3.28 [1.48–7.26]	< 0.01	541.62	568.95	525.62
Twice vs none	3.19 [1.56–6.52]	< 0.01
Sex
Male vs female	0.67 [0.38–1.17]	0.16
Employment status
Employed vs unemployed	0.65 [0.32–1.30]	0.22
Previous hospitalization due to AUD
Yes vs no	1.42 [0.84–2.40]	0.19

Two ordinal logistic regression models satisfied the proportional odds assumption at the significance level of 0.05. None of the confounders were significant. Although the model-fit statistic AIC was slightly smaller in the adjusted model, the likelihood ratio test based on −2LLs from the two models was non-significant, and the SC was smaller in the unadjusted model
^aThe unadjusted model includes only the independent variable, No. of KPPC-matched treatments
^bThe adjusted model adds Sex, Employment status, and Previous hospitalization due to alcohol use disorder

ISSN: 1940-0640