How are the costs of drug-related morbidity measured?: a systematic literature review

Abstract

Background: Drug-related morbidity has been associated with increased healthcare costs and has been suggested as one of the leading causes of death. Previous reviews have identified heterogeneity in research methods in studies measuring the cost of drug-related morbidity. To date, no attempt has been made to analyse different methods and cost sources used when estimating the costs of drug-related morbidity.

Objective: The aim of this review was to evaluate and compare methods and data sources in cost estimates of drug-related morbidity.

Methods: A literature search was conducted in three electronic databases (CINAHL, EMBASE and MEDLINE) to identify peer-reviewed articles written in English and published between January 1990 and November 2011. Articles were included if estimating the direct or indirect costs of drug-related morbidity based on clinical data from general patient groups. The general patient groups were defined as patients visiting, being admitted to, treated at or discharged from a general hospital, excluding studies from nursing homes or specialized hospitals. Study information was collected using a standardized data collection sheet. Studies were categorized according to the type of costs included in the cost analysis. Thereafter, the cost analyses of included studies were reviewed regarding viewpoint, costing methods and adjustments for timing of costs.

Results: In total, 9569 articles were identified, of which 25 publications were included in this review, and four additional articles were identified from reference or citation lists of publications already included. Eighteen studies measured either the total or attributable costs of drug-related morbidity, while seven studies estimated the increased costs using matched controls or regression analyses. Six studies measured costs from a payer perspective, while the other 23 measured costs to the hospital. One study included costs resulting after discharge, and discounted future costs, while the remaining 28 studies measured costs during the initial admission only and involved no adjustment for timing of costs.

Conclusions: The data sources and costs measured in the included studies varied considerably in terms of perspectives and use of data sources. Even though there is a trend towards more studies estimating costs from the payer perspective, the identified studies still focused on costs resulting from patients attending hospital, therefore underestimating the cost of drug-related morbidity. There is thus a need for more research on the costs of drug-related morbidity to providers other than hospitals, and costs occurring outside of hospitals and after the initial care episode. Such studies require clear descriptions of how the costs of drug-related morbidity are measured, and should adhere to published guidelines for observational studies and economic evaluation studies.