Objectives: This study aimed to identify factors associated with medication non-adherence at hospital discharge and calculate cut-off values for these factors. Methods: A logistic regression model was prepared using a retrospective dataset of 141 cases. Age, sex, the number of drugs prescribed after ICU discharge, the score on the motor and cognitive components of the functional independence measure (FIM), and the Charlson comorbidity index score were analyzed as potential independent risk factors for medication non-adherence after discharge. Cut-off values and diagnostic accuracy were evaluated by receiver operating characteristic curve, and the results were calibrated against another retrospective dataset of 72 cases, which was collected over a different period. Results: The age ≥75 years (OR 3.81, 95%CI 1.33–10.9) and scoring ≤4 on more than one item in the cognitive component of the FIM were found to be independent risk factors for medication non-adherence after ICU discharge. A FIM cognitive score cut-off of ≤26 had high diagnostic accuracy (sensitivity 62.5%, specificity 88.1%), and the accuracy ratio was highly consistent between the two datasets. Conclusions: The FIM cognitive score and age at ICU admission were independent risk factors for medication non-adherence after discharge, and the diagnostic accuracy was highest for a FIM cognitive score cut-off of 26 points or less.