Patients’ adherence to the medications or treatment regimens prescribed to them by their clinicians is an important determinant of their healthcare outcomes across a variety of clinical settings. But how do we measure it, and is there a better way?

Tracking medication adherence

When patients take their medications or maintain their treatment regimens as prescribed, treatment effectiveness and patient responses improve and may lead to a reduction in avoidable healthcare costs as well.¹

Typically, medication adherence is calculated using the proportion of days covered (PDC), a metric based on the proportion of prescription claims filled by a patient for a given period of time. In fact, a PDC threshold of 80% is recommended by the Pharmacy Quality Alliance based on clinical evidence as being a sufficiently high level of adherence for achieving reasonably high treatment effectiveness. However, medication adherence is a dynamic phenomenon and often changes over time due to various factors, including side effects and the patient’s perception of the medication’s effectiveness. And while PDC is easy to calculate, it is also a deterministic metric that cannot capture the complexity of such a dynamic phenomenon.

To better capture heterogeneity in medication adherence, group-based trajectory modeling (GBTM)—a statistical method used to identify population subgroups following similar trajectories, allowing for the exploration of common patterns over time—is increasingly being proposed in the pharmacoepidemiology literature as a better method.

A better way to track: Group-based trajectory modeling

Researchers in the Real-World and Advanced Analytics team at Cytel investigated whether the PDC was adequate at classifying patients by their medication adherence over time. Using simulation methods testing different combinations of medication patterns, follow-up time, and sample sizes, we found that PDC was considerably biased compared to GBTM.²

The GBTM outperformed PDC substantially in distinguishing between similar medication adherence trajectories and has the advantage of being able to include domain expertise. GBTM is increasingly being recognized as an alternative to PDC for improving how medication adherence is measured. Unlike the PDC, it is more involved analytically, but results suggest that it can help reduce bias.

This can be useful when planning clinical trials where non-adherence may be of concern; for post-hoc analyses of trial or real-world data for discovering patient subgroups or factors that affect adherence; and when targeting or tailoring adherence-improving strategies to the right patients at the right time to improve outcomes on treatment.

Final Takeaways

Group-based trajectory modelling is less biased than the commonly used PDC at classifying patients according to their adherence patterns. In the future, the GBTM method may help better personalize strategies to improve medication adherence in patients.

Notes:

[1] Aurel O. Iuga and Maura J. McGuire, “Adherence and Health Care Costs,” Risk Management and Healthcare Policy 20 (7) (2014), https://doi.org/10.2147/rmhp.s19801.

[2] This work was presented by Alind Gupta, Research Principal, at the 2023 International Conference for Pharmacoepidemiology in Halifax.

Share Via