In the quest for a more efficient and streamlined healthcare system, the adoption of electronic prescribing systems has been a significant step forward. However, as we delve into the potential risks associated with these systems, particularly in relation to 'look-alike sound-alike' (LASA) medication errors, we uncover a complex and critical issue that demands our attention.
The Human Factor in a Digital Age
The tragic case of Sidra Aliabase, a three-week-old baby who lost her life due to a prescribing error, serves as a stark reminder of the consequences of LASA errors. In an era where technology is meant to enhance accuracy, the question arises: are we inadvertently creating new avenues for mistakes?
A Transition's Tale
The rollout of electronic prescribing and medicines administration (ePMA) systems across the NHS began in the 1990s, with a push towards a paperless system gaining momentum in recent years. While the UK government claims a 30% reduction in medication errors with electronic prescribing, the reality is more nuanced.
Data: A Murky Picture
Obtaining precise data on LASA incidents is a challenge. The transition between incident reporting systems and the lack of a specific LASA category in reports create a murky picture. As Bryony Dean Franklin, a professor of medication safety, points out, the data may not reflect an increase in errors, but rather a shift in their nature.
Errors in Transition
Franklin and Julia Scott, a pharmacist and chief information officer, suggest that errors have simply moved from one stage to another. With paper prescribing, errors might occur due to illegible handwriting, while electronic systems introduce new challenges, such as selecting the wrong drug from a drop-down menu. It's a trade-off, they argue, and one that needs careful consideration.
Mitigating Strategies
One tactic to prevent LASA errors in paper-based systems is 'tall-man lettering', which capitalizes certain letters in drug names to distinguish them. Scott suggests implementing this in e-prescribing systems, along with changing drug grouping and improving drop-down menus. She also highlights the potential of AI-integrated clinical decision support to provide sophisticated prompts and reduce errors.
The AI Conundrum
While AI offers promising solutions, it also introduces new risks. Scott warns of 'ambient voice technology' (AVT) or 'AI scribes', which can lead to sound-alike errors. She questions how these errors can be mitigated, suggesting the need for AI-enabled clinical decision support to retrospectively identify and highlight errors in transcripts.
Other Approaches
Franklin highlights 'Touchdose', a system that allows prescribing by indication, reducing the likelihood of LASA errors. This tool, developed by Dosium, has shown promising results in reducing prescribing errors.
Under-Reporting: A Hidden Challenge
The true scale of LASA errors remains elusive due to under-reporting. Franklin notes that only a tiny fraction of prescribing and administration errors are reported, making it difficult to gain a clear understanding of the issue. She expresses hope that AI can improve the analysis of error reports.
The Role of AI: Promise and Pitfalls
Scott agrees that AI has the potential to revolutionize error reporting, capturing near-misses and providing valuable data. However, she acknowledges the need to address environmental and ethical concerns associated with AI.
A Complex Web
The issue of LASA errors in electronic prescribing systems is a complex web of challenges and potential solutions. While technology offers advancements, it also presents new risks and errors. As NHS England moves forward with its LFPSE system, the hope is that, coupled with AI, it can reduce the rate of LASA errors. The journey towards safer prescribing practices is ongoing, and it requires a careful balance of human insight and technological innovation.
