AI/ML Methodologies Predict Roadblocks to Patient Access to Therapy
IntegriChain, a leading life sciences commercial data and analytics company that helps patients start therapy faster and stay on therapy longer, announced it has now implemented more than 10 artificial intelligence/machine learning (AI/ML) models and predictive analytics to its ICyte Platform. IntegriChain’s advancements in AI/ML and predictive analytics were recently highlighted in a Gartner report, “Life Science CIOs Can Accelerate Commercial Effectiveness With New Applications of Artificial Intelligence,” which is available on Gartner.com.
“We are first and foremost a data science company, employing leading-edge technologies and algorithms to deliver insights that help patients get access to critical medicines and to save patient days of therapy,” said IntegriChain CEO Kevin Leininger. “This milestone exemplifies our commitment to remaining positioned at the forefront of data science and implementing enabling technologies that help us better serve our customers and their patients.”
“By employing AI/ML methodologies in our ICyte platform, we deliver higher quality, actionable data and predictive analytics that identify patients at risk for not getting on therapy and not continuing therapy as prescribed by their doctors,” said Lucas Dan, IntegriChain Manager of Data Science. “This allows healthcare stakeholders to proactively target patient services to help at-risk patients overcome these barriers to therapy.”
IntegriChain employs a number of AI/ML models including Random Forest Classification, XGBoost Classification, Kalman Filter, K-Nearest Neighbor, K-Means Clustering, and ARIMA Modeling to deliver insights including:
- Patient Initiation Risk Scores. We use a variety of data sets to predict the likelihood that patients who are referred for Specialty medication will not be able to initiate on therapy due to patient journey roadblocks such as their insurance company denying coverage.
- Patient Adherence Risk Scores. We use a variety of data sets to predict the likelihood that patients who are already on therapy for Specialty medication will discontinue therapy before their next refill.
- Inventory Demand Projection. We employ disparate data sets to project future inventory demand for each channel and National Drug Code (NDC) combination.
- Automated Pharmacy Selection for Pharmacy Call Programs. We produce statistically significant lists of pharmacies for the Pharmacy Call Programs that allow for intelligent, focused, and efficient outreach to pharmacies, maximizing the investments in these programs.
- NDC/Brand Similarity Algorithms. We employee NDC and brand-level features–such as channel level sales distribution, therapeutic category, and WAC price–to determine similar NDCs/brands for various analytic projects.