Epidemiology and cost analysis of pharmacist interventions at Ghent University Hospital.

BACKGROUND

Performing bedside clinical recommendations is important for the prevention of adverse drug events. From an economic perspective, the economic value of adverse drug event avoidance needs to be weighed against the labour costs of pharmacists.

OBJECTIVE

To perform a cost analysis of pharmacist interventions with valproic acid, digoxin, methotrexate and penicillin.

SETTING

Ghent University Hospital in Belgium (1,062-beds).

METHOD: Interventions for valproic acid, digoxin, methotrexate and penicillin were selected from a clinical pharmacy database, CLINOR. The average number of registered interventions per year was 1,209 (period 2005-mid 2011).

MAIN OUTCOME MEASURE: Cost difference (cost value) between that of the avoided toxicity and that of the intervention (a positive cost value is cost saving).

RESULTS

Per annum, pharmacists performed interventions for valproic acid (n = 18) and digoxin (n = 21); the annual cost value of interventions for valproic acid was €18,853.7 with a standard deviation of €15,020.6; for digoxin it was €41,832.0 ± €15,348.5. With oral methotrexate, accidental toxicity occurs rarely but it can be life threatening. Two important pharmacist interventions were reported per year. The routine switching of penicillin therapy to alternative antibiotics, in patients with previous allergy, may invoke costs rather than benefits (two interventions per year). In half of cases, therapy was reinitiated without any further adverse drug event.

CONCLUSION

Clinically important pharmacy interventions are not automatically cost beneficial. Interventions that prevent digoxin and valproic acid toxicity were cost effective in this setting. The routine advice to switch the antibiotic class for every reported penicillin allergy is unlikely to avoid adverse drug events and challenges the cost value of this intervention. Interventions with methotrexate are relevant because they can be lifesaving. However, due to their low incidence, effective detection of these errors is crucial for reducing harm.