Preventable and non-preventable adverse drug events in hospitalized patients: a prospective chart review in the Netherlands.

Background: Medication safety research and clinical pharmacy practice today is primarily focused on managing preventable adverse drug events (pADEs). Determinants of both pADEs and non-preventable adverse drug reactions (ADRs) have been identified. However, relatively little is known on the overlap between these determinants and the balance of preventable and non-preventable harm inpatients experience in modern computerized hospitals.

Objective: The aim of this study was to analyse the prevalence of pADEs and non-preventable ADRs as well as the determinants, including multimorbidity, of these ADEs, i.e. both pADEs and ADRs.

Methods: Adverse events experienced by patients admitted to two Dutch hospitals with functioning computerized physician order entry (CPOE) systems were prospectively identified through chart review. Adverse events were divided into pADEs (i.e. as a result of a medication error) and non-preventable ADRs. In both cases, a causal relationship between adverse events and patients' drugs was established using the simplified Yale algorithm. Study data were collected anytime between April 2006 and May 2008 over a 5-month period at each hospital ward included in the study, beginning from 8 weeks after CPOE was implemented at the ward.

Results: pADEs and non-preventable ADRs were experienced by 349 (58%) patients, of whom 307 (88%) had non-preventable ADRs. Multimorbidity (adjusted odds ratio [OR(adj)] 1.90; 95% CI 1.44, 2.50; OR(adj) 1.28; 95% CI 1.14, 1.45, respectively), length of stay (OR(adj) 1.13; 95% CI 1.06, 1.21; OR(adj) 1.11; 95% CI 1.07, 1.16, respectively), admission to the geriatric ward (OR(adj) 7.78; 95% CI 2.15, 28.13; OR(adj) 3.82; 95% CI 1.73, 8.45, respectively) and number of medication orders (OR(adj) 1.25; 95% CI 1.16, 1.35; OR(adj) 1.13; 95% CI 1.06, 1.21, respectively) were statistically significantly associated with pADEs and ADRs. Admission to the gastroenterology/rheumatology ward (OR(adj) 0.22; 95% CI 0.06, 0.77; OR(adj) 0.40; 95% CI 0.24, 0.65, respectively) was inversely related to both pADEs and ADRs. Other determinants for ADRs only were female sex (OR(adj) 1.77; 95% CI 1.12, 2.80) and use of drugs affecting the nervous system (OR(adj) 1.83; 95% CI 1.09, 3.07). Age was a significant determinant for pADEs only (OR(adj) 1.07; 95% CI 1.03, 1.11).

Conclusions: In this study more than half of the patients admitted to the hospitals are harmed by drugs, of which most are non-serious, non-preventable ADRs (after the introduction of CPOE). Determinants of both pADEs and ADRs overlap to a large extent. Our results imply the need for signalling early potential adverse events that occur during the normal use of drugs in multimorbid patients or those in geriatric wards. Subsequent therapeutic interventions may improve the well-being of hospitalized patients to a greater extent than focusing on errors in the medication process only.