Introduction: Opiates are traditionally used in STEMI patients to relieve pain, decrease pulmonary congestion, and anxiety. Yet, according to in vitro measurements, they delay and diminish the effect of all currently used oral platelet P2Y12 receptor antagonists. To be able to draw causal conclusions on the potential effects of opiates on clinical outcomes using real-world registries, the bias in treatment allocation may be minimized by propensity score methods. Therefore, we constructed a simple model of prehospital opiate administration using the database of the National Ambulance Service. Methods: We analyzed data of 10,139 subjects who were diagnosed with STEMI within 24 hours from symptom onset between November 2018 and June 2021. Patients with unsuccessful out-of-hospital CPR were excluded. Opiates were applied in 3,992 cases (39.4%). Besides onset-to-door time and need for CPR due to OHCA, basic demographic data like age and sex were studied as candidate predictors using logistic regression. Non-linearity of the continuous predictors was assessed by restricted cubic splines. Relative importance of the predictors was characterized by the Akaike Information Criterion (AIC). Results: Overall model fit was good: likelihood ratio chi-square=610.40, d.f.=10, p<0.0001. Opiate use was strongly associated with onset-to-door-time (p<0.0001, AIC=333.83) and Wald testing for linearity suggested a non-linear relationship (p<0.0001) with the highest odds at about 120 minutes. Similarly, it was also highly (p<0.0001, AIC=193.78) and non-linearly (p<0.0001) related to age with an inverted U-shaped curve. Moreover, it was negatively dependent on the need for CPR (p=0.0019, AIC=7.64) and unrelated to the sex of the patient (p=0.15, AIC=0.09). Conclusion: Our propensity score model may help to evaluate the role of opiate administration on clinical outcomes of STEMI patients and underscores the importance of non-linear modeling of continuous predictors.