Blecker Saul, Zhao Yunan, Li Xiyue, Kronish Ian M, Mukhopadhyay Amrita, Stokes Tyrel, Adhikari Samrachana
Department of Population Health, NYU Grossman School of Medicine, New York, NY, USA.
Division of Cardiology, Department of Medicine, NYU Grossman School of Medicine, New York, NY, USA.
J Gen Intern Med. 2025 Mar;40(4):811-817. doi: 10.1007/s11606-024-09216-5. Epub 2024 Nov 25.
Medication non-adherence, which is common in chronic diseases such as heart failure, is often estimated using proportion of days covered (PDC). PDC is typically calculated using medication fill information from pharmacy or insurance claims data, which lack information on when medications are prescribed. Many electronic health records (EHRs) have prescription and pharmacy fill data available, enabling enhanced PDC assessment that can be utilized in routine clinical care.
To describe our approach to calculating PDC using linked EHR-pharmacy data and to compare to PDC calculated using pharmacy-only data for patients with heart failure.
We performed a retrospective cohort study of adult patients with heart failure who were prescribed guideline-directed medical therapy (GDMT) and seen in a large health system. Using linked EHR-pharmacy data, we estimated medication adherence by PDC as the percent of days in which a patient possessed GDMT based on medication pharmacy fills over the number of days the prescription order was active. We also calculated PDC using pharmacy-only data, calculated as medications possessed over days with continued medication fills. We compared these two approaches for days observed and PDC using a paired t-test.
Among 33,212 patients with heart failure who were prescribed GDMT, 2226 (6.7%) never filled their medications, making them unavailable in the assessment of PDC using pharmacy-only data (n = 30,995). Linked EHR-pharmacy data had slightly longer days observed for PDC assessment (164.7 vs. 163.4 days; p < 0.001) and lower PDC (78.5 vs. 90.6, p < 0.001) as compared to assessment using pharmacy-only data.
Linked EHR-pharmacy data can be used to identify patients who never fill their prescriptions. Estimating adherence using linked EHR-pharmacy data resulted in a lower mean PDC as compared to estimates using pharmacy-only data.
药物治疗不依从在心力衰竭等慢性病中很常见，通常使用覆盖天数比例（PDC）来评估。PDC通常使用药房的药物配药信息或保险理赔数据来计算，这些数据缺乏药物处方时间的信息。许多电子健康记录（EHR）都有处方和药房配药数据，从而能够进行更完善的PDC评估，并可用于常规临床护理。
描述我们使用关联的电子健康记录 - 药房数据计算PDC的方法，并与使用仅药房数据计算的心力衰竭患者的PDC进行比较。
我们对在大型医疗系统中接受指南指导药物治疗（GDMT）的成年心力衰竭患者进行了一项回顾性队列研究。使用关联的电子健康记录 - 药房数据，我们根据药物药房配药情况，以患者拥有GDMT的天数占处方订单有效天数的百分比来通过PDC估计药物依从性。我们还使用仅药房数据计算PDC，计算方法是拥有药物的天数除以持续配药的天数。我们使用配对t检验比较这两种方法在观察天数和PDC方面的差异。
在33212例接受GDMT处方的心力衰竭患者中，2226例（6.7%）从未取药，这使得在使用仅药房数据（n = 30995）评估PDC时无法纳入这些患者。与使用仅药房数据的评估相比，关联的电子健康记录 - 药房数据在PDC评估中观察到的天数略长（164.7天对163.4天；p < 0.001），且PDC较低（78.5对90.6，p < 0.001）。
关联的电子健康记录 - 药房数据可用于识别从未取药的患者。与使用仅药房数据的估计相比，使用关联的电子健康记录 - 药房数据估计依从性导致平均PDC较低。
