OBJECTIVES : This study attempts to identify and correlate the impact of tailored support services led clinicians on patient adherence and identify key variances impacting adherence risk.

METHODS : A retrospective cohort study of patients enrolled in a patient support program between January 2017 and May 2019 was conducted to evaluate adherence to medication across two patient cohorts: unenrolled and remained enrolled in clinician led support services provided by Lash Group. Patient adherence was measured by medication possession ratio (MPR) using a patient-specific, variable follow-up period from start of therapy to last available service date. Enrolled patients were segmented using the patient activation measure (PAM) and assessed for barriers found within the 5 dimensions of adherence. To describe and quantify inter-cohort differences in outcomes and patient characteristics, chi-square tests and Student’s t-tests were conducted. Multivariate logistic regression models predicted the likelihood of adherence 80% or higher.

RESULTS : A total of 24,842 patients were included in the analysis and the population had an average of 12 prescription fills during the follow up period. It was determined that those who remained enrolled had a 0.79 MPR vs. a 0.64 MPR for patients who unenrolled (p-value <.0001). Primarily, patients suffered from therapy related barriers (79.9% for unenrolled and 76.2% for stayed, p-value<.001) – notably mastery of certain techniques. In the multivariate analysis, those who continued to receive nursing support services were 1.9 times more likely to have an MPR of 0.80 compared to those who unenrolled (OR=1.91, 95% CI:1.78, 2.05). Older age, female gender, region in the west or NE, PAM score, and number of patient barriers identified proved to be significant predictors of MPR above 0.80.

CONCLUSIONS : Patient continued utilization of clinician led support services proves to be an approach to ensure patients remain on treatment as prescribed. Further research will show the importance of predictive variables.