Author summary Polypharmacy continues to grow in importance because of multimorbid, aging populations. In healthcare, polypharmacy is well-known for its detrimental consequences related to adverse events. Polypharmacy is also a key challenge included in the precision medicine agenda. So-called precision dosing is important for addressing concomitant diseases and medications, and their impact on treatment responses. In this work, we present a study aiming at investigating this problem using real-world data from ~185 million treatment episodes leading to significant statistical power across drug-cocktails. We introduce a comprehensive analysis of dosage changes aiming to identify those significant pairs that may influence each other. More than half of the identified drug pairs were associated with readmission, mortality or longer stays and we also observed major differences in relation to disease and laboratory tests. Under the premise that drug-drug interactions are manageable through patient monitoring and dosage adjustments, we collected and cross-referenced information from a wide range of clinical and bioinformatics drug-drug interaction databases that could be related to pairs associated with dosage changes. Overall, this work shows how distinct, changing dosage patterns can facilitate the identification of drug-drug interactions in the context of polypharmacy complementing longitudinal analyses of disease progression.