Background: Medication adherence may be measured either by (a) patient-reported measurement or (b) objective measurement using medication dispense data. Proportion of days covered (PDC) is an objective measure of adherence becoming more readily available in the electronic health record (EHR) and a threshold of < 80% is commonly used to identify a potential adherence problem. Using PDCs may be more efficient and clinically practical than using longer patient-reported adherence questionnaires. However, it is unknown how well the PDC and patient reported measures correlate.

Methods: As part of a pragmatic randomized trial to evaluate an intervention designed to improve medication adherence in a Midwest health care system, 1,423 adults with type 2 diabetes (T2D) were identified at an index visit in primary care based on suboptimal diabetes control (AIC >8%) and at least one oral diabetes medication with PDC < 80%. They were surveyed shortly after the index visit to assess adherence using a validated 12 question ASK-12 instrument, with a response rate of 29% (80/280) . We evaluated the relationship between PDC and ASK-12 total/subscale scores using linear regression.

Results: There were no significant associations found between PDC scores and the total ASK-12 score (β=-0.4, p=0.47) , or ASK-12 subscales of inconvenience/forgetfulness (β=-1.1, p=0.42) , treatment beliefs (β=-0.4, p=0.77) , behavior (β=-0.7, p=0.55) , or barriers (β=-0.7, p=0.44) .

Conclusions: For patients above recommended A1C goal with a medication PDC < 80%, there was no correlation between PDC scores and patient-reported measures of adherence using ASK-12 total or component scores. This analysis raises questions about the validity and usefulness of EHR-derived PDC scores to identify poor medication adherence for patients with T2D.


P.Pawloski: None. M.Pankonin: None. D.Rehrauer: None. C.Frail: None. J.Anderson: None. L.Chumba: None. J.Ziegenfuss: None. P.J.O'connor: None. S.P.Dehmer: None. H.Ekstrom: None. J.Haapala: None. J.M.Sperl-hillen: None. M.Becker: None. A.Truitt: None.


NHLBI #R01HL136937

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at