Context This quality improvement project was carried out at a large general paediatric hospital. The project was performed on our medical and surgical inpatient wards, involving all prescribing healthcare professionals.

Problem There had been numerous minor and moderate prescribing errors identified via DATIX and CPSQ meetings, the majority not affecting patient care and were identified prior to administration of medication, thus avoiding serious incidents. However, they caused delays in treatment and sub-optimal, timely patient care. This inefficient practice wastes time and resources of nurses, doctors and pharmacists in identification of errors and re-prescribing of medication.

Assessment of problem and analysis of its causes This problem had already been identified at CPSQ meetings and the medication errors were ongoing despite improvement in prescribing training at staffing induction.

A multidisciplinary meeting involving doctors, senior nursing clinicians and the pharmacy lead was held to identify common errors. The prescribing errors were primarily illegible handwriting, inaccurate history taking, confusion with drug names, inappropriate use of decimal points, incorrect completion of drug charts and use of verbal orders and abbreviations. Clinicians were often unaware of errors, with colleagues amending drug charts, with no direct feedback to the person making the error.

Intervention Our quality improvement intervention was to implement a “zero tolerance policy” of prescribing errors, such that no medication would be given until correctly prescribed. No verbal orders were accepted. All errors once identified were photographed by paediatric pharmacists. All photographic evidence was then communicated to the consultant who directly discussed it with the individual responsible, either face-to-face, or via email. Individuals’ results were published monthly. The intervention was timed so that it would there would be no staff change over.

The information was relayed back to the multi-disciplinary team in a teaching session, highlighting the numbers and types of errors using graphical evidence.

Strategy for change The implementation was simple once agreement had been reached with the pharmacy team who wished to prioritise a reduction in prescribing errors. The main stakeholders involved were the pharmacy team, senior nurses and consultants. Results were disseminated as previously described. It was expected that change would occur slowly initially, with the project needing to be ongoing as medical staff rotate every four months.

Measurement of improvement A total of 204 errors were identified in the four months prior to the zero tolerance policy being implemented, and assigned according to staffing posts. These included 26 errors at consultant level, 53 at registrar, 74 SHO errors, and 51 for other prescribing posts.

In the first month of our project being initiated, there were a recorded 74 errors, with the biggest reduction being between this and the second month (34), with a further 34 errors in the third month, reducing to only 10 by the fourth month. This demonstrated a clear trend in reduction of prescribing errors.

Effects of changes There was a dramatic decrease by 81% in total errors.

Additionally there has been a generalised perceived but unmeasurable improvement in efficiency of distribution of medication.

Lessons learnt This was a surprisingly simple intervention with excellent results. However it has been hard to maintain momentum at continuing this analysis.

Message for others This was a simple project to implement and to demonstrate improvement. It improves individual accountability and self-awareness.

There is potential to develop a phone application capturing errors and sending feedback to prescribers, producing a league table.

Since incorrect medical prescribing has been identified as being the single most preventable cause of patient harm (previous studies identifying errors in one in eight charts), this could form part of a formal assessment in the e-portfolio, with doctors providing annual error reports.