OBJECTIVES: This study aims to examine the potential benefits and implications of transitioning from traditional paclitaxel delivery to nano-drug delivery for breast cancer using AI/ML models.

METHODS: We used the Optum® de-identified Market Clarity Dataset, a real-world patient database covering 80 million lives. Unstructured data (physician notes) were also leveraged to determine staging information in breast cancer patients. Incident breast cancer patients from 2019-20 were identified. The first documented claim of Paclitaxel was the index event. We ensured 12-month pre- and 24-month post-index medical and pharmacy eligibility. TNM Staging data were collected via physician notes, identified using NLP. Propensity score matching was used to remove any bias. The cohort was divided into patients on nano-drug delivery versus traditional drug delivery for Paclitaxel. The data were divided into an 80:20 ratio for model training and testing and model performance was evaluated based on precision, recall and F1 score.

RESULTS: For the entire cohort, 192,743 patients were identified as receiving traditional drug delivery for Paclitaxel, in contrast to 51,981 patients who were administered Paclitaxel via nano drug delivery. Patient’s making switch from traditional to nano drug or vice-versa were excluded from the analysis.

CONCLUSIONS: Our study provides critical insights for healthcare stakeholders by predicting the drug switch from traditional to nano drug delivery in cancer patients. This model can also be replicated for other therapeutic areas, leading to potential cost savings, improved patient outcomes, and efficient resource utilization.