Evaluation of medication administration timing variance using information from a large health system’s clinical data warehouse

Abstract

Disclaimer In an effort to expedite the publication of articles related to the COVID-19 pandemic, AJHP is posting these manuscripts online as soon as possible after acceptance. Accepted manuscripts have been peer-reviewed and copyedited, but are posted online before technical formatting and author proofing. These manuscripts are not the final version of record and will be replaced with the final article (formatted per AJHP style and proofed by the authors) at a later time. Purpose An analysis to determine the frequency of medication administration timing variances for specific therapeutic classes of high-risk medications using data extracted from a health-system clinical data warehouse (CDW) is presented. Methods This multicenter retrospective, observational analysis of 1 year of medication administration data from 14 hospitals was conducted using a large enterprise health-system CDW. The primary objective was to assess medication administration timing variance for focused therapeutic classes using medication orders and electronic medication administration records data extracted from the electronic health record (EHR). Administration timing variance patterns between standard hospital staffing shifts, within therapeutic drug classes, and for as-needed (PRN) medications were also studied. Calculated variables for delayed medication administration (ie, administration time variance) were created for documented administration time intervals of 30-59, 60-120, and more than 120 minutes before or after medication orders. Results A total of 5,690,770 medication administrations (3,418,275 scheduled and 2,272,495 PRN) were included in the normalized data set. Scheduled medications were frequently subject to delays of ≥60 minutes (15% of administrations, n = 275,257) when scheduled for administration between 9-10 AM and between 9-10 PM. By therapeutic drug class, scheduled administrations of insulins, heparin products, and platelet aggregation inhibitors (most commonly heparin flushes and line-management preparations) were the most commonly delayed. For PRN medications, medications in the anticoagulant and antiplatelet agent class were most likely to be administered early (<60 minutes from the scheduled time of first administration). Conclusion The findings of this study assist in understanding patterns of delayed medication administration. Medication class, time of day of scheduled administration, and frequency were factors that influenced medication administration timing variance.

Using routinely collected clinical data for circadian medicine: A review of opportunities and challenges

“…Although clinical care takes place around the clock in hospitals, specific procedures or treatments are provided at specific times of day and not at other times. The timing of medication orders and administration are influenced by hospital rounding and may not be driven by the time of greatest clinical need [ 65 , 66 ]. Similarly, sampling for laboratory tests typically occurs in the early morning hours [ 67 , 68 ] and other care activities are—not surprisingly—unevenly distributed around the clock [ 69 ].…”

Cohort profile for development of machine learning models to predict healthcare-related adverse events (Demeter): clinical objectives, data requirements for modelling and overview of data set for 2016–2018

“…In many countries Clinical Data Warehouses (CDW) collect and reuse healthcare data from EMR for various applications covering all domains of medicine. [12][13][14][15][16][17][18][19][20] To integrate and analyse EMR data from patients treated at our university hospital for the purposes of research, education and institutional management, the hospital has established a CDW. 21 This data can be supplemented with publicly available data from the French National Institute of Statistics and Economic Studies (INSEE) (eg, data on social determinants of health (SDOH)).…”

“…Moreover, the availability of EMR data facilitates the use of new approaches such as state-of-the-art machine-learning methods for risk prediction. In many countries Clinical Data Warehouses (CDW) collect and reuse healthcare data from EMR for various applications covering all domains of medicine 12–20. To integrate and analyse EMR data from patients treated at our university hospital for the purposes of research, education and institutional management, the hospital has established a CDW 21.…”