THE IMPACT OF AN ELECTRONIC ALERT IN PREVENTING DUPLICATE ANTICOAGULANT PRESCRIBING (submitted in 2019)
European Statement
Patient Safety and Quality Assurance
Author(s)
Alison Brown, Gillian Cavell, Nikita Dogra, Cate Whittlesea
Why was it done?
Anticoagulants are high-risk drugs. An NHS England Patient Safety Alert was published in 2015 highlighting harm from inappropriate co-prescription of anticoagulants1.
What was done?
A ‘duplicate anticoagulant alert’ (Anticoagulant MLM) was implemented within our electronic prescribing system (EPMA) to alert prescribers if co-prescription of two or more anticoagulants was attempted, with the intention of preventing the completion of a potentially harmful prescription. We conducted a retrospective review of the impact of the Anticoagulant MLM on preventing co-prescription of low-molecular weight heparin (LMWH) and direct oral anticoagulants (DOACS)
How was it done?
The study took place in a 950 bed UK acute teaching hospital. A report of all Anticoagulant MLM alerts generated for adult inpatients between 26th June 2017 and 8th October 2018 was extracted from EPMA. Data on drugs prescribed, alert acceptance or override and duplicate anticoagulant administration were collected. Where alerts were overridden, appropriateness of the override was assessed by an anticoagulation specialist pharmacist. Ethics approval was not needed.
What has been achieved?
The Anticoagulant MLM triggered on 894 occasions; 113 in response to attempted prescription of a LMWH for a patient already prescribed a DOAC. 65 of 113 alerts were overridden (duplicate prescription completed). 48 alerts were accepted (duplicate prescription avoided). Of the 65 overridden alerts, consecutive doses of both anticoagulants were scheduled appropriately. No duplicate doses were administered in 44 cases (44/65, 67.7%). 15 duplicate prescriptions were either cancelled before administration or not administered concurrently (15/65, 23.1%). Duplicate doses were administered against 6 prescriptions (6/65, 9.2%), on 3 occasions. No patient harm was identified. The alert prevented inappropriate co-prescription of anticoagulants to 48 patients. Overrides were justified in 44 cases. Anticoagulants were correctly prescribed for 92/113 (81.4%) patients. It was outside the scope of this project to investigate why alerts were overridden. ‘Alert fatigue’2 and alert frequency3 are recognised factors limiting the effectiveness of electronic alerts in changing a planned course of action.
What next?
The alert remains in place as a barrier to error. Further work is needed to identify reasons for anticoagulant alert overrides.