Cost of adverse drug events related to potentially inappropriate medication use: A systematic review

Abstract

Background: The use of potentially inappropriate medications (PIMs) for older people is associated with worse health outcomes owing to the occurrence of adverse drug events (ADEs) and drug interactions, leading to increased health care costs.

Objectives: Identify the costs of ADEs related to PIMs use, in addition to the costs predictors.

Methods: A systematic review was conducted in the PubMed and Scopus databases (until February, 2022), and the report of this study was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Interventional and observational studies that reported costs of ADEs regardless of perspective (i.e., payer) were considered. Reporting and methodological quality were assessed using the tool proposed by Larg and Moss for evaluating cost-of-illness studies.

Results: A total of 20 (21 publications), published between 2001 and 2020, were included (236,888,744 older people). The ADEs costs related to PIMs use were mostly related to the use of health services (hospitalization [n = 7], health care expenses [n = 7], and emergency department visits [n = 3]). Among the 8 studies that reported P value, 7 identified higher costs for PIM users than non-PIM users. Three studies reported cost predictors, being highest number of PIMs in use, age older than 75 years, male gender, general health status in older people in use of benzodiazepines, and drug interactions in older people diagnosed as having dementia. Regarding to assessment of reporting and methodological quality, all studies had at least one limitation (answer "no").

Conclusions: Our findings suggest that PIMs use is associated with higher costs of hospitalization, health care expenses, and visits to emergency department owing to ADEs, regardless of PIMs in use, health service, perspective, and screening tools used for PIMs identification. However, these findings should be interpreted with caution because all studies had at least one methodological limitation.