Primary medication nonadherence calculation method specifications impact resulting rates

Autumn D Zuckerman; Josh DeClercq; Nisha B Shah; Victoria W Reynolds; Megan E Peter; Aaron M Pavlik; Leena Choi

doi:10.1016/j.sapharm.2021.03.016

Primary medication nonadherence calculation method specifications impact resulting rates

Res Social Adm Pharm. 2022 Mar;18(3):2478-2483. doi: 10.1016/j.sapharm.2021.03.016. Epub 2021 Apr 20.

Authors

Autumn D Zuckerman¹, Josh DeClercq², Nisha B Shah³, Victoria W Reynolds³, Megan E Peter³, Aaron M Pavlik⁴, Leena Choi²

Affiliations

¹ Department of Pharmaceutical Services, Vanderbilt University Medical Center, Nashville, TN, USA. Electronic address: Autumn.zuckerman@vumc.org.
² Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN, USA.
³ Department of Pharmaceutical Services, Vanderbilt University Medical Center, Nashville, TN, USA.
⁴ Department of Health Informatics, Vanderbilt University Medical Center, Nashville, TN, USA.

PMID: 33926825
DOI: 10.1016/j.sapharm.2021.03.016

Abstract

Background: Previous literature has illustrated a wide range of primary medication nonadherence (PMN) rates due to inconsistent calculation methods and parameters, but the impact of parameter specifications on PMN rates has not been assessed.

Objectives: The objective of this study was to evaluate the impact of lookback window (LBW), duplicate window (DW), and fill window (FW) specifications on PMN rates in patients prescribed specialty self-administered oncology medications.

Methods: This was a single-center, retrospective cohort analysis. Patients receiving a new electronic specialty oncology prescription January-December 2018 were included; excluded if re-routed to an external pharmacy within 2 days, fell within a DW, or cancelled within a FW. Twenty-four methods were used to calculate PMN based on combinations of the following parameters: (i) absence of prior specialty self-administered oncology medication fill within LBW (90, 180 days); (ii) absence of a duplicate prescription within DW (2, 7, 30 days); and (iii) sold status within FW (14, 30, 60, 90 days). For each method, PMN was calculated as the number of unsold prescriptions within the FW divided by all eligible prescriptions.

Results: We evaluated 4,482 prescriptions, resulting in PMN ranging from 16% to 23%. Patients were commonly male (53%) and white (83%), with a median age of 64 years (interquartile range, IQR, 54, 72). Increasing the LBW from 90 to 180 days resulted in exclusion of 72 (2%) prescriptions and minimally impacted PMN rates. Most duplicate prescriptions (87%) occurred within two days of original prescription and PMN rates were minimally affected by DW. Most fulfilled prescriptions were filled within FW 30 days, 98% with a method of LBW 180, DW 2, and FW 30. Adjusting the FW consistently impacted PMN rates.

Conclusions: Because various PMN definitions can significantly impact results, a thorough explanation of all parameter specifications should be reported in research using PMN.

Keywords: Medication adherence; Oncology; Primary medication nonadherence; Specialty pharmacy.

MeSH terms

Cohort Studies
Electronic Prescribing*
Humans
Male
Medication Adherence
Middle Aged
Pharmacies*
Retrospective Studies