Prevalence of drug-drug interactions and actual harms in acute care hospitals

Background: Drug-drug interactions (DDIs) have the potential to cause patient harm and this risk has increased with rising rates of polypharmacy and aging populations. The likelihood of harm associated with potential DDIs (pDDIs) depends on their clinical relevance. Clinically relevant DDIs (cDDIs) could lead to patient harm when a patient’s clinical profile, drug effects, and severity of potential outcome are considered. Limited evidence exists on how frequently pDDIs are also cDDIs. This study aimed to determine the prevalence of pDDIs, cDDIs, and subsequent actual harm among hospitalized patients.

Methods: This was a secondary analysis of control arm of a cluster randomised trial. Patients of all ages were randomly selected from three Australian hospitals. Retrospective medical chart review was conducted by clinical research pharmacists, independent from study hospitals. Clinical contextual factors (n=26) related to the drug, patient, and clinical setting were used to determine whether a pDDI was clinically relevant. Harm was assessed by an expert panel.

Results: Of 1186 patient admissions, 15,860 pDDIs were identified. Overall, 70.1%(n=831) of patients experienced a pDDI, 42.6% (n=505) a cDDI, and 0.9%(n=11) subsequent actual harm in hospital. Of all pDDIs identified, less than one third(27.0%, n=4,285) were classified as cDDIs. The median number of pDDIs and cDDIs per 10 drugs were 6 (interquartile range[IQR]:2-13) and 0 (IQR:0-2) respectively. The distributions of DDIs were similar for males and females. Patients aged 65 years and over experienced a relatively high number of pDDIs (median 7.9-9.2/10 drugs) and cDDIs(median 1.6-1.7/10 drugs) [insert the Figures attached with Filenames ‘Figure1.JPG and ‘Figure 2.JPG here]. Furthermore, age was a main patient contextual factor influencing the classification of 14% of pDDIs as cDDIs. Drug dose was the most common factor used to determine when a pDDI was unlikely to be a cDDI (29.7% of all non cDDIs).

Discussion: Large proportions of inpatients experienced potential DDIs but nearly three quarters were not clinically relevant to patients and very few were associated with actual harm during hospital stay. Contextual factors associated with clinically relevant DDIs identified in this study could be used to design more targeted interventions to improve medication safety in hospitals.