Interesting New Study on Classifying Non-adherence

By | September 30, 2013

News & Comment about medication adherence

New ways of describing adherence behavior types may allow better prediction and management of non-adherence.

Commitments to Reduce Non-adherence

CVS Caremark is probably best known as one of the largest U.S. pharmacy groups. What may be less well-known are its strong commitments to improve medication adherence. Commitments that include research partnerships with leading academic and hospital centers, like Harvard, Brigham and Women’s Hospital, and University of Pennsylvania.

A list of 29 adherence-related publications from these partnerships can be found here. Of particular interest for this post, a 30th paper from these research efforts was published earlier this month.

New Insights into Non-adherence Patient Types

The new publication in question can be found here.

To date, descriptions of adherence-related behavior types have been inexact. This is largely because the most common adherence measure – proportion of days covered (PDC) – is unable to distinguish between different patient situations. For instance, the same PDC value may apply to three patients who – over the same period of time – have very different behaviors, eg:

  • Patient #1 adheres early, but later shows non-adherence
  • Patient #2 is non-adherent early, but later shows adherence
  • Patient #3 shows ‘on/off’ adherence and non-adherence over the period

This latest study uses new methods to help identify different adherence behavior types. Defining these behaviors is important as different solutions to overcome non-adherence may apply to each behavior type.

In the study, nearly 265,000 patients receiving statin medications were followed for 15 months. Results from the study described six distinct behaviors. These are listed below, together with the percentage of patients that fell into each category.

  1. Patients that were nearly always adherent (23.4%)
  2. Patients that showed non-adherence early, then adhered later (11.4%)
  3. Patients whose adherence declined slowly over the study period (11.3%)
  4. Patients who only took the medication occasionally throughout (15.0%)
  5. Patients whose adherence declined rapidly right from the start (19.3%)
  6. Patients that had no refills after the initial prescription (23.4%)


Up to now a general figure of ‘50% non-adherence’ to chronic medications has commonly been used. These new data give deeper insights into the issue of non-adherence, suggesting multiple distinct patient behaviors types.

There are marked differences between the six identified behaviors. So it’s not unreasonable to think these behaviors may be driven by equally different underlying reasons for non-adherence. And if there are different underlying reasons, it’s also not unreasonable to imagine that separate interventions will be needed to address these different reasons.

Key steps to improve medication adherence.

Suggested key steps to improve medication adherence.

While the study provides interesting new insights, it sheds light on only a part of the overall problem. Identifying and classifying non-adherence is a critical first step. Beyond that, it will be just as important to investigate and understand the reasons why patients show these behaviors. And lastly, healthcare professionals need effective solutions that specifically align with each reason; plus the time to implement these solutions properly.

What are your thoughts on these new findings?