Prescription Timeliness Prediction Using Machine Learning

Abstract

Adherence is a very important aspect for each patient enrolled in a Prescription Drug Plan, a large gap between claim fills is an indicator that they might be disrupting their therapy. This could be due to unavailability, high drug costs or any other factor. The prediction of when each patient would go and get a refill or set a medical appointment to receive a new prescription would highly reduce such gaps and increase adherence. The purpose of the program is to predict when these occurrences might take place, using historical data, to better tailor adherence programs to a patient’s schedule. The methodology involves implementing Machine Learning utilizing R Services within SQL Server 2017. Using the information available an accurate prediction was not established using only demographic information. Additional historical information on an individual patient basis is necessary to be able to establish a more robust prediction Key Terms  Adherence, Machine Learning, Prescription Timeliness Prediction, Prediction using R in SQL Server.