How our hospital reduced its paracetamol overdose prescribing errors to zero

Here at Queen Elizabeth Hospital Birmingham (QEHB), we see in excess of 400 paracetamol overdose cases every year. The regimen to protect patients from liver damage after paracetamol overdose is very difficult to calculate. 

The process requires three infusions of varying doses and dilutions of N-acetylcysteine, given over different durations. Each is sequentially prescribed as a separate prescription and, overall, it is an arduous process for prescribers.

In early 2019, Adam McCulloch, a consultant in the admissions unit at QEHB, saw that prescribers were struggling with N-acetylcysteine infusions. He had noticed that prescribing these infusions was tricky and was taking prescribers a long time to get right.

At the time, I was assisting a colleague with improving prescribing of infusion bundles for patients admitted with inherited metabolic conditions. It dawned on me that a similar solution could work for N-acetylcysteine infusions. Further discussions with pharmacy colleagues revealed that they were often having to provide guidance to prescribers on N-acetylcysteine infusions: not only on admissions and in the emergency department, but also on the wards.

Back in 2012, the Commission on Human Medicines reviewed and changed recommendations on the use of N-acetylcysteine for treating acute paracetamol overdose. And the Medicines and Healthcare products Regulatory Agency (MHRA) published a nomogram that aimed to weight-band and simplify the prescription of N-acetylcysteine. Despite this, uncertainty on prescribing these infusions continues.

This is where our project started. Using the nomogram as a template, we developed an order set (an electronic prescribing term used to describe a bundle of prescriptions that can be triggered in sequence). The only prescriber decision would be the diluent, and the rest would be automated.

Three prescriptions would be prescribed in sequence, and the prescriber would be asked to validate the prescription to populate the dose automatically, based on the patient’s weight. With very few variables, mathematically it was one of the simpler order sets to formulate, as it only had 16 well-defined branches. If the weight entry was older than a few days, the system would return an error and prompt for an updated weight.

However, weight was the first problem I encountered. Weight is a continuous variable, but owing to the way that we, as humans, write ranges, there are always lots of gaps. The MHRA nomogram had ranges for 40−49kg and 50−59kg, but our sharp-eyed programmers noticed this leaves a gap between 49kg and 50kg. So, we decided to opt for 40kg−<50kg and 50kg−<60kg ranges.

Once written, the tool was programmed by the drug dictionary analyst — an expert in the coding of the drug dictionary within our electronic patient record system, known as PICS (Prescribing, Information and Communication System). The tool was then uploaded into a test system and tested by a pharmacy colleague using weights from the bottom, middle and top of each weight bracket. It worked perfectly on the first test and after we showed it to consultants on admissions, it was signed off for release.

In order to find out if the tool actually reduced prescribing errors, McCulloch led a study looking at the accuracy of prescribing before and after implementation.

We discovered a 25% error rate before implementation of the tool. Following its implementation, the number of prescription errors was brought down to zero.

This was extremely satisfying, not only because it highlighted the importance of a multidisciplinary approach to solve healthcare problems, but also because it showed how quite a simple intervention could be used to completely eliminate the risk of error.

Eliminating these errors also eliminates the chance of a patient experiencing an adverse drug event or further liver damage and, in turn, reduces the amount of time patients need to be in hospital. It also makes the lives of our medical colleagues a little easier by reducing one of their sources of stress.

This work shows that pharmacy is not just about giving out drugs.