A study billed as the largest ever to focus on the administration of blood-thinning drugs has concluded that calculating dosages by computer is at least as safe and reliable as having them measured by expert medical professionals.

The study, which was published in the Journal of Thrombosis and Haemostasis, was carried out in 32 medical centres across the European Union and associated countries. It analysed nearly 400,000 tests for the international normalised ratio (INR), or normal blood clotting, on 13,219 patients. Of these tests, 193,890 were based on manual dosing of anticoagulants and 193,424 on computer-assisted dosing.

As Professor Leon Poller, who co-ordinated the international research team from a central facility at the University of Manchester’s Faculty of Life Sciences in the UK, explained, the need for computer assistance in determining the correct dosage to prolong the INR in patients taking blood-thinning drugs such as warfarin “arises from the massive demand for oral anticoagulants, following their success at treating an increasing number of thrombotic and embolic conditions”.

But prescribing the right oral dose can be problematic, even for experienced medical staff, Manchester University pointed out. Individual patients have widely varying responses to any given dose, while the response of a single patient can also change over the course of an illness. If too high a dose is given, the blood can thin out too much, potentially triggering internal bleeding. Too low a dose can make the blood clot too readily.

Previous studies supporting the use of computer-assisted dosage have relied solely on laboratory results and have not be large enough to assess whether prolongation of the INR resulted in clinical benefit and improved safety, the university said.

In the large-scale clinical trial, the proportion of manual tests found to give the correct INR was 64.7% compared with 65.9% for computer-assisted dosage. Moreover, the number of INR tests resulting in clinical complications was 7.6% lower in all the clinical groups with computer-assisted dosage.

While this overall reduction was not statistically significant, in the 3,209 patients with deep vein thrombosis or pulmonary embolism, the number of clinical events following treatment was significantly lower with computer-assisted dosages than with manual dosages – 6.1 per 100 patient years compared with 9.1 per 100 patient years when the dosages were calculated by medical staff.

“The results are even more impressive when you consider that the comparisons were made against medical professionals based at centres that specialised in prescribing oral anticoagulants,” Poller commented.

“At the very least, our study confirms the clinical safety and effectiveness of computer-assisted dosage using the two systems we tested [the PARMA 5 and DAWN AC programmes] and should help to bring relief to overstretched medical professionals while providing reassurance to patients,” he added.