Modifications to the drug burden index calculation may limit interpretation of its association with clinical outcomes in older adults.

Abstract

We read with great interest the article by Floroff and colleagues [1], which reported that potentially inappropriate medication use and increasing Drug Burden Index (DBI) exposure are associated with clinical outcomes in critically ill older inpatients with neurological injury. This research demonstrates important new findings and the utility of the DBI in a clinical setting not previously investigated. However, two adaptations to the DBI calculation used in this study may hinder interpretation of the findings. Firstly, Floroff and colleagues state ‘‘the DBI for each medication was calculated by dividing the daily dose by the recommended minimum daily dose (Dose/Minimum Daily Dose)’’. The original equation by Hilmer and colleagues [2] calculates the DBI using a dose response model, as the sum of Daily Dose/(Daily Dose + Minimum Daily Dose) for each individual medication. The original DBI has been validated across a number of countries and has been associated with poorer physical function [2, 3], falls [4], frailty [5], hospitalization [6], and mortality [7] in older adults across many clinical settings. Secondly, the calculation articulated by Floroff and colleagues details that ‘‘medications with anticholinergic and sedative properties were defined using the updated Beers and STOPP criteria independent of diagnosis, as well as previously published studies’’. By definition, the DBI uses country-specific registered medication information to identify medications with clinically significant anticholinergic and sedative effects [2–5, 8–10]. The main advantage of this approach is that the DBI can be applied in various settings internationally [11]. The approach used by Floroff and colleagues included medication classes that do not satisfy these criteria (e.g., H2 receptor antagonists) and others that may not have been considered or clearly defined (e.g., sedating antihistamines or dopamine agonists). The formula used by Floroff and colleagues considerably changes the DBI exposure for each patient, and may impact on the association of DBI with important clinical outcomes in older adults. Changes to the calculation of DBI may under or overestimate the actual exposure and the association of DBI with outcomes in this population. Furthermore, use of a different DBI calculation method with the same name, which has not previously been validated, can alter the interpretation and validity of the DBI for further research and translation to clinical practice. If the data presented by Floroff and colleagues were further analyzed using the original DBI equation, this would provide better understanding of the anticholinergic and sedative burden in this subset of older adults, and its implications for neurological injury and care.