Over 116 million people worldwide have chronic pain and prescription dependence. In the US, opioids account for the majority of overdose deaths, and in 2014, almost 2 million Americans abused or were dependent on prescription opioids. Genetic factors may play a key role in opioid prescription addiction. Herein, we describe genetic variations between opioid addicted and non-addicted populations and derive a predictive model determining risk of opioid addiction. This case cohort study compares the frequency of 16 single nucleotide polymorphisms involved in the brain reward pathways in patients with and without opioid addiction. Data from 37 patients with prescription opioid or heroin addiction and 30 age and gender matched controls were used to design the predictive score. The predictive score was then tested on an additional 138 samples to determine generalizabilty. Results for Method Derivation of Observed data: ROC statistic=0.92, sensitivity=82% (95% CI: 66-90), specificity=75% (95% CI:56-87). TreeNet “learn” data: ROC statistic=0.92, sensitivity=92%, specificity=90%,precision=92%, and overall correct=91%. Results of Generalizability data: Sensitivity=97% (95% CI: 90 to 100), specificity=87% (95% CI:86 to 93), positive likelihood ratio=7.3 (95% CI: 4.0 to 13.5), and negative likelihood ratio=0.03 (95% CI:0.01 to 0.13). This negative likelihood ratio can be used as an evidence based measure to exclude patients with a high risk of opioid addicition or substance use disorder. By identifying patients with a lower risk for opioid addiction, our model may inform therapeutic decisions.