Are electronic prescribing systems increasing the risk of ‘look-alike sound-alike’ medication errors?
The tragic death of a three-week-old baby, Sidra Aliabase, on May 10, 2024, due to a prescribing error, highlights the potential risks of 'look-alike sound-alike' (LASA) medication errors in electronic prescribing systems. This incident, where sodium acid phosphate was prescribed instead of sodium chloride, underscores the dangers of these errors, which can occur when similar-sounding medications are chosen from a drop-down menu. The coroner's report and the doctor's confirmation in court reveal a system that needs urgent improvement.
The issue of LASA errors is not isolated to this case. In November 2025, an alert was issued about the risk of healthcare staff incorrectly recording patients' penicillin allergies as penicillamine allergies, leading to potentially fatal anaphylactic reactions. This error, caused by the similarity in drug names and the alphabetical sorting of the drop-down menu, further emphasizes the need for better systems.
The implementation of electronic prescribing and medicines administration (ePMA) systems across the NHS, particularly since the push to go paperless, has been aimed at reducing medication errors by 30%. However, the available data suggests that while errors may not have increased, they have merely been replaced by new types of LASA errors. Bryony Dean Franklin, Professor of Medication Safety at University College London, and Julia Scott, a pharmacist and chief information officer, argue that the shift from illegibility errors to drop-down menu errors has not significantly impacted the overall error rate.
The transition from paper-based to electronic prescribing has led to a change in the nature of errors. Studies have shown that while errors involving incorrect doses and illegible orders are less common with ePMA, those involving duplication, omission, incorrect drug, and incorrect formulation are more prevalent. This shift highlights the need for new strategies to mitigate LASA errors.
One proposed solution is the use of 'tall-man lettering' in ePMA systems, which involves capitalizing certain letters in drug names to distinguish them from others. While this method has shown some effectiveness, it is not a complete solution. Scott suggests adding extra features to the sorting and filtering of drugs, such as forcing penicillamine and penicillin to appear separately in the list. Additionally, the integration of clinical decision support AI could help prevent LASA errors by applying logic and providing prompts.
However, the integration of AI also raises concerns about new types of errors, particularly with ambient voice technology (AVT) or 'AI scribes'. Scott warns that this technology could introduce a new category of significant sound-alike error risk, similar to the days of verbal orders. The challenge lies in mitigating these new errors, which may require layered AI-enabled clinical decision support systems.
Other methods to reduce LASA errors include the development of systems like 'Touchdose', which matches doses to indications and performs sensitive checks. A study comparing Touchdose with standard medication orders showed a significant reduction in prescribing errors. However, the under-reporting of errors remains a challenge, with only a small percentage of errors being reported as incidents.
The potential of AI to analyze LASA error reporting and capture near-misses is promising. Scott suggests that AI could enhance the current system by parsing incident reports and extracting relevant data. However, the integration of AI also requires addressing known issues, such as environmental impact and ethical concerns.
In conclusion, while LASA errors are unlikely to be fully eliminated, the LFPSE system and the potential of AI offer hope for reducing the rate of these errors. The NHS's commitment to using the latest technology, including machine learning, to improve patient safety is a step in the right direction. However, it will require significant investment in skills and knowledge to ensure the safe and effective implementation of these systems.
