Introduction The advantages of electronic prescribing systems are well recognised offering superiority over paper-based systems in terms of legibility, dose calculation, assignment of dose limits, and report generation. Such systems do not come with a populated drug file so each institution must design and build their own. This is a substantial project that is complex, labour-intensive and risk-prone. A road-map was not available at the time for how this could be achieved.

Aims To describe the process of building an electronic drug file for a paediatric intensive care unit (PICU).

Method (1) Commencement: Initial work consisted of setting out certain standards, such as acceptable abbreviations, units of measure and reference sources. As the unit caters for a wide range of patients from premature neonates to teenagers, it was a risk to use one upper dose limit for each drug. Patient categories were created, consisting of (a) premature (b) neonate (c) 1 month to 1 year (d) over 1 year up to 40 kg (e) over 40 kg. For each of these patient categories, a minimum and maximum body weight was assigned which would be used to calculate dose limits. (2) Configuration: Four limits for each drug dose were inputted: minimum, low normal, high normal and maximum. It was decided that all minima would be zero and low normal would reflect the lower end of the normal dose range (hence, dose reduction for renal impairment would be highlighted as abnormal). Drug configuration and corrections were done by one PICU pharmacist, who logged the reference for any unusual decisions. Each draft was checked by a second PICU pharmacist. A selected list of drugs and their various assigned limits was reviewed by a consultant intensivist. An ‘Issues Log’ was maintained for liaison with the vendor where issues were logged as they arose and their progress to resolution noted. (3) Testing: Teams of testers drawn from consultants and senior nursing staff volunteered for testing. They were assigned specific drugs and formulations to test, which they did by prescribing each drug for each patient category. A form was used to log details of the drug tested and any difficulties that arose. Progress of the testing phase was logged on a spreadsheet. (4) Miscellaneous: 556 drug formulations were configured, consisting of 3756 limits in total. Aciclovir alone has 180 limits because of the complexity of dosing for different age groups and indications, as well as different formulations (injection, cream, etc). (5) Time Frame: Work on this project took two pharmacists 9 months to complete, working 3–7 h/day. Testing was conducted over a 14 week period.

Conclusions A formal structured approach is key when undertaking a large complex project. Tools, such as issues and progress logs are important to track the multitude of questions that arise. Close working relationships with all stakeholders need to be maintained throughout the process.

In times of reduced financial and human resources, sharing of information is critical to reducing waste. This roadmap will help other institutions undertaking a similar project.