WOONSOCKET, R.I. — A new approach to classifying patients by their long-term medication adherence behavior may increase the accuracy of compliance studies.

In a study published in the September issue of Medical Care, researchers at CVS Caremark Corp. and Brigham and Women’s Hospital followed more than 264,000 statin users over a 15-month period and created measures to account for different adherence ­behaviors.

The researchers identified key groups into which various patients would fall in an effort to predict and compare their long-term adherence patterns.

The study utilized a new method of categorizing adherence called group-based trajectory modeling, which is based on observed patterns of medication filling over time. This method allows researchers to more accurately capture and describe adherence compared to techniques that simply classify patients as adherent or not, using average levels of ­adherence.

In addition, the trajectory patterns that are created should allow for more targeted interventions to address nonadherence. Trajectory modeling is especially useful for accounting for variable patterns of intermittent and long-term adherence behaviors.

“Our findings could help facilitate research on medication adherence and medication effectiveness in a variety of ways,” said Dr. Niteesh Choudhry, associate physician in the Division of Pharmacoepidemiology and Pharmacoeconomics at Brigham and Women’s, and an associate professor at Harvard Medical School.

“For example, by using group-based trajectory modeling, we could help identify patients with distinct patterns of adherence so that health care professionals could appropriately target interventions. In addition, data related to the quantity and timing of medication availability could help improve our understanding of the effects of nonadherence on clinical outcomes.”

Based on the study group, researchers categorized patients into six distinct groups:

• Nearly always adherent (23.4%).
• Brief gap in medication use or filled irregularly during the first nine months, but improved during the last six months (11.4%).
• Slowly declining adherence throughout the 15-month period (11.3%).
• Used statins only occasionally across the period (15.0%)
• Rapid decline in statin use after initiation (19.3%).
• Virtually no fills after the initial fill (23.4%).

Group trajectory modeling also identified certain characteristics that impact medication adherence. For example, among the study participants, on average, those patients with the best adherence were older, more likely to have a higher income, more likely to be a high school graduate and less likely to be black.

Also, those with the best adherence were more likely to be a Medicare Part D beneficiary or live in New England. Those with the lowest adherence rates tended to be generally younger, male and less likely to have an initial prescription that provided them with more than a 30-day supply of medication.