Proportion of days covered (PDC) is the most widely endorsed measure of medication adherence. For instance, the Centers for Medicare & Medicaid Services (CMS) uses PDC as its preferred adherence measure for its Quality Rating System for many chronic conditions (e.g., hypertension, diabetes, and hyperlipidemia). However, your PDC measure may be flawed depending on what dates you are using. A paper by Zheng et al. (2026) explains the issue:
Many approaches rely on the adjudication (fill) date, the point at which a claim is billed, but this may not accurately reflect when the patient actually receives or begins using the medication. This disconnect is pronounced with automatic refills, mail-order pharmacies, or delays between filling and sold.0 Additionally, at the prescribing stage, substantial delays occur between prescription writing and filling: among 32,586 new chronic disease prescriptions, only 75% were filled within 7 days and 81% within 30 days of being written
A key question does the date used–i.e., when the prescription was written, when it was filled (adjudicated), and when it was sold (picked up)–make a difference? The authors also use two different PDC definitions: exposure PDC, which is the standard approach, and adherence PDC which adjust for large treatment gaps. In the adjusted approach, the authors define a maximum gap (MaxGap) variable that they would consider to be non-adherent (vs. viable decision to stop medication). Specifically, “…if the original gap exceeded this MaxGap, we treated the excess as an excluded gap, and only the MaxGap was counted as an adjusted gap in the observation window.” So the patient had a 120 gap between medication fills but MaxGap was set to 90 days, only the 90 days would count as being non-adherent; the remaining 30 days would be ignored.
To answer this question, the authors used Surescripts data between 2009 and 2019 for patients prescribed a cardiovascular medication (i.e., angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), calcium channel blockers (CCBs), or diuretics).
The authors found that:
Exposure PDC varied substantially across sources (median = 0.747-0.794; P < 0.001), whereas adherence PDC was uniformly high (0.959-0.977; P < 0.001). Written dates produced the lowest exposure PDC because of inflated observation windows and gaps, and sold dates yielded the highest…At the 0.80 threshold, adherence classification was stable for adherence PDC (>94% concordance across sources) but highly sensitive for exposure PDC, with more than 15% of patients switching classification when written dates were used instead of fills.
The authors spend a fair amount of time in the discussion section of the article aruging for the merits of their ‘adherence PDC’ metric over the ‘exposure PDC’ metric traditionally used. You can view the full paper here.
