IDENTIFYING DETERMINANTS OF MEDICATION PERSISTENCE IN RHEUMATOID ARTHRITIS USING MACHINE LEARNING AND REGRESSION ANALYSIS
Author(s)
Zulfiia Ditto, PhD, Andrew Wash, PharmD, PhD, Casey Fitzpatrick, PharmD, Ana I. Lopez Medina, PharmD, PhD;
CPS Solutions, LLC, Dublin, OH, USA
CPS Solutions, LLC, Dublin, OH, USA
OBJECTIVES: Persistence to rheumatoid arthritis (RA) therapy is critical for achieving disease control, yet real-world data consistently show persistence rates as low as 34%. Health-system specialty pharmacies (HSSPs) support patients with RA by facilitating medication access and providing clinical management and monitoring. In doing so, they collect a rich array of data that differs from traditional pharmacy records. This study aimed to identify factors influencing medication persistence among patients with RA in an HSSP setting.
METHODS: This retrospective, observational study utilized HSSP data for patients with RA who initiated specialty therapy from 2022-2024. Patients who dispensed their medication with the HSSP for a full year were included. There were 42 variables analyzed using feature importance analysis to identify potential predictors of persistence. The primary outcome was medication persistence, defined as whether a patient remained on therapy without a 60-day gap within one year of therapy initiation. Logistic regression was used to validate the significance and strength of association of selected variables. A p-value <0.05 was considered statistically significant.
RESULTS: A total of 930 patients were included in the analysis: 79% female and mean age of 59±13 years. One year after initiating therapy, 68% remained persistent. Feature importance analysis identified five variables with statistically significant associations with persistence. Logistic regression revealed that the following were associated with higher persistence: patients prescribed oral vs. injectable medications (OR: 2.04, 95%CI:1.14-3.66), improvement in RAPID3 scores (OR: 1.54, 95%CI:1.14-2.09), male sex (OR:1.47, 95%CI: 1.02-2.12), and receipt of financial assistance (OR: 1.96, 95%CI:1.34-2.87). Conversely, patients prescribed both oral and injectable therapies were less likely to persist on therapy (OR: 0.32, 95%CI: 0.23-0.47).
CONCLUSIONS: These findings highlight actionable areas for intervention, such as optimizing therapy type and ensuring financial support, to enhance persistence and improve outcomes. Further research is needed to explore causal relationships and develop predictive models for persistence.
METHODS: This retrospective, observational study utilized HSSP data for patients with RA who initiated specialty therapy from 2022-2024. Patients who dispensed their medication with the HSSP for a full year were included. There were 42 variables analyzed using feature importance analysis to identify potential predictors of persistence. The primary outcome was medication persistence, defined as whether a patient remained on therapy without a 60-day gap within one year of therapy initiation. Logistic regression was used to validate the significance and strength of association of selected variables. A p-value <0.05 was considered statistically significant.
RESULTS: A total of 930 patients were included in the analysis: 79% female and mean age of 59±13 years. One year after initiating therapy, 68% remained persistent. Feature importance analysis identified five variables with statistically significant associations with persistence. Logistic regression revealed that the following were associated with higher persistence: patients prescribed oral vs. injectable medications (OR: 2.04, 95%CI:1.14-3.66), improvement in RAPID3 scores (OR: 1.54, 95%CI:1.14-2.09), male sex (OR:1.47, 95%CI: 1.02-2.12), and receipt of financial assistance (OR: 1.96, 95%CI:1.34-2.87). Conversely, patients prescribed both oral and injectable therapies were less likely to persist on therapy (OR: 0.32, 95%CI: 0.23-0.47).
CONCLUSIONS: These findings highlight actionable areas for intervention, such as optimizing therapy type and ensuring financial support, to enhance persistence and improve outcomes. Further research is needed to explore causal relationships and develop predictive models for persistence.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
CO156
Topic
Clinical Outcomes
Disease
SDC: Systemic Disorders/Conditions (Anesthesia, Auto-Immune Disorders (n.e.c.), Hematological Disorders (non-oncologic), Pain)