Computerised decision support software (CDSS) is increasingly used in the management of patients taking oral anticoagulant therapy.

CDSS use dosing algorithms to recommend both anticoagulant dosage and time until next test (recall times). A number of studies have shown CDSS to increase the ‘time in range’, possibly improving the quality of anticoagulation monitoring. They also facilitate audit of anticoagulant control. By increasing the ‘time in range’ (and allowing it to be easily calculated and audited) it is believed that the adoption of CDSS will reduce the complications of anticoagulant therapy (bleeding or recurrent thrombosis). Recent guidance from the British Committee for Standards in Haematology and National Patient Safety Agency recommend that CDSS are used to assist in the monitoring of patients taking oral anticoagulant drugs.

There are many different CDSS for oral anticoagulant monitoring available, using different algorithms for dosing. Most systems have the option for centres to locally adjust the CDSS dosing algorithms. In addition, all systems allow for manual ‘override’ of the suggested anticoagulant dose and time until next test- allowing for manual practitioner dosing. This has the potential to result in considerable variation in the recommended anticoagulant dose and recall times.

In 2006, NEQAS circulated anticoagulant dosing exercises to participants in the UK NEQAS Near Patient Testing (NPT) and the UK NEQAS for Blood Coagulation (hospital laboratory) programmes. Participants were asked to suggest a warfarin dose and recall time for each of 3 clinical scenarios (in each scenario three historical INR results and warfarin dosages were provided).

32% of the 955 UK NEQAS for Blood Coagulation participants, and 49% of the 793 UK NEQAS Near patient Testing (NPT) participants returned the exercises. 64% of the NPT participants used CDSS to recommend warfarin dose and recall times for each of the clinical scenarios, whereas only 25% of the Blood Coagulation participants used CDSS for the purpose of the exercise. A number of different CDSS systems were used; the majority of the Blood coagulation participants used DAWN AC wheras INRstar and DAWN AC were the most frequently used CDSS systems amongst the NPT participants.

In the first scenario, a patient stably anticoagulated with 4 INR results between 2.5 and 2.7 on 5mg/day warfarin; all but 3 centres recommended continuing on the same dose of warfarin. These three centres recommended warfarin doses of between 1.8 and 4.9 mg/day. The recommended recall times varied between 7 and 42 days with most recommending a recall time of 21 days. The variability in recall times occurred whether dosed manually or by CDSS.

The second and third scenarios were of patients with less stable anticoagulation. In these scenarios there was marked variability in both recommended warfarin dosages and recall times, both in patients dosed manually and those dosed by software systems.

The exercise highlighted the considerable variability that exists in the dosing and recall of patients taking oral anticoagulant therapy, whether dosed with the assistance of CDSS or manually.

Despite the use of CDSS systems, there appears to be little standardisation of recommended warfarin dosage and recall times.

Thanks to Ian Jennings from UK NEQAS who circulated the questionnaires and analysed the data. Also thanks to all those who participated in the exercise.