In today's fast-paced world, where technology is often seen as a panacea for all our problems, it's crucial to take a step back and critically examine its potential pitfalls, especially in the realm of healthcare. The recent tragic case of baby Sidra Aliabase, who lost her life due to a medication error, raises a critical question: are electronic prescribing systems, designed to improve efficiency and reduce errors, inadvertently increasing the risk of 'look-alike sound-alike' (LASA) medication mistakes?
This article delves into the complexities of this issue, exploring the potential pitfalls of electronic systems and the human factors that contribute to medication errors. We will also examine the role of AI in both creating and mitigating these risks, and discuss the challenges and opportunities it presents for the future of medication safety.
The Human Factor in Medication Errors
One of the key insights from experts like Professor Bryony Dean Franklin is that medication errors are not solely a result of human fallibility. While mistakes are inevitable, a significant portion of errors can be attributed to system design and our understanding of cognitive processes.
In the case of LASA errors, the transition from paper-based to electronic prescribing systems has shifted the error stage. With paper, errors often occurred at the dispensing or administration stage due to illegible handwriting. However, with electronic systems, the error is more likely to occur at the prescribing stage, when selecting from a drop-down menu.
As Julia Scott, a pharmacist and chief information officer, puts it, "It's almost that we've taken away illegibility errors with one hand, and given drop-down menu errors with the other." This shift highlights the need to understand and address the cognitive mechanisms that lead to errors, rather than solely focusing on the fallibility of human nature.
Mitigating LASA Errors: A Multi-Pronged Approach
To reduce LASA errors, a combination of strategies is necessary. One such strategy is 'tall-man lettering', where certain letters in drug names are capitalized to distinguish them from similar-sounding drugs. While this method has shown some effectiveness, it does not eliminate the risk entirely.
Another approach suggested by Scott is to change how drugs are grouped and sorted in electronic systems. By forcing commonly confused drugs out of alphabetical order, the risk of selecting the wrong medication can be reduced. Additionally, Scott proposes implementing minimum character sets, requiring users to type more characters before a drop-down menu appears, thus narrowing the options and making it safer.
The Promise and Perils of AI Integration
The integration of AI into electronic prescribing systems holds great promise for medication safety. AI-powered clinical decision support systems can provide sophisticated prompts, ensuring that the selected medication aligns with the patient's diagnosis. For instance, if a doctor is prescribing for a chest infection but selects penicillamine, the system can prompt them, "You were writing about a chest infection, but you've selected penicillamine. I know those two things don't go together."
However, as Scott warns, there is a 'flip side' to AI integration. Ambient Voice Technology (AVT), also known as 'AI scribes', which transcribe conversations between patients and healthcare professionals, introduces a new category of sound-alike error risk. In this scenario, the AI scribe might mishear 'penicillin' as 'penicillamine', leading to potential errors.
This highlights the need for a balanced approach. While AI can enhance medication safety, it also introduces new error mechanisms. As Scott concludes, "Everything you introduce, there'll be a new error mechanism."
The Role of Reporting and Analysis
Currently, the scale of LASA errors, particularly in relation to electronic prescribing systems, is difficult to determine due to under-reporting. Professor Franklin notes that only about 1 in 100 prescribing errors and 1 in 1,000 administration errors are reported. This under-reporting is not due to a desire to hide mistakes but rather practical challenges, such as the time and energy required to report an incident, especially when healthcare professionals are in the midst of patient care.
AI, however, offers hope in analyzing LASA error reporting. With its ability to process large amounts of data quickly and accurately, AI can help identify patterns and insights that might otherwise be missed. The new LFPSE system, for instance, is designed to be more amenable to analysis, potentially improving our understanding of medication errors and their causes.
Conclusion: A Balanced Approach to Medication Safety
While LASA errors are unlikely to be fully eliminated, the future of medication safety lies in a balanced approach. This includes continuing to improve electronic prescribing systems, integrating AI safely and responsibly, and enhancing reporting and analysis processes.
As we navigate the complexities of medication safety in a rapidly evolving technological landscape, it's crucial to remember that while technology can enhance safety, it is not a silver bullet. A holistic understanding of the human factors, system design, and cognitive processes involved in medication errors is essential to developing effective strategies for reducing these errors and improving patient outcomes.
