Introduction Prescription errors in neonatal intensive care units (NICU) are not uncommon and can lead to significant mortality and morbidity. Electronic prescribing has been shown to reduce prescription errors. We speculate that the use of a computer-based spreadsheet program would reduce calculation errors by physicians.

Methods Our neonatal drug dose calculator (NDDC) is a spreadsheet based program (Excel, Microsoft) incorporating our drug formulary. Weight limits were set between 450 and 4500 g to minimise input errors. The NDDC also calculates the amount of drug to be added to a fixed volume for drug infusions. Two tests (A and B) based on commonly used drugs were designed. Each doctor was asked to undertake one test by using the NDDC and another test with a handheld calculator. Each test was timed and errors were analysed. Results were compared using the Mann–Whitney test (Sigmastat version 3).

Results A total of 10 doctors participated in 20 tests. Of the 10 tests (5As and 5Bs) undertaken using the NDDC there were three errors in 271 prescriptions (1.1%) compared with 18 errors in 269 prescriptions (6.7%) when using a handheld calculator (p = 0.005). The median time taken to complete the tests with the NDDC was significantly lower than the other group (44 vs 63 minutes, p = 0.007). There were four >10-fold errors using the handheld calculator but none with the NDDC.

Conclusions The use of our NDDC significantly reduces calculation errors and prescription time by physicians. Drug errors of 10-fold or more could be prevented using the NDDC.