Are electronic prescribing systems as safe as we think?

Electronic prescribing (e-prescribing) systems are embedded across NHS hospitals and are viewed as a major advance in patient safety. 

By automating checks and providing clinical decision support, they are used to reduce prescribing errors and improve medicines management. 

Yet prescribing errors remain a significant source of preventable harm in healthcare. In England, medication-related errors are estimated to cause 712 deaths yearly and contribute to 1,708 more^​1​. Digital systems are increasingly relied upon as an important mitigation strategy. How well do they perform once embedded into routine practice?  

While e-prescribing systems undergo extensive testing at implementation, their ongoing safety performance is not routinely evaluated once they are embedded into everyday clinical practice. Nationally, there are clinical safety standards for digital systems, alongside information and interoperability requirements for e-prescribing and medication administration (ePMA)^​2​. However, they largely focus on risk management processes and data exchange, rather than defining minimum performance expectations for how effectively live systems should mitigate prescribing risks.  

As an ePMA pharmacist with a clinical background as an independent prescriber in critical care, my practice revolves around ensuring a safe and optimal digital system. A national evaluation using the electronic prescribing risk and safety evaluation (ePRaSe) tool, published in April 2026, has offered rare quantitative insight into how NHS e-prescribing systems perform in a real-world setting, raising important questions about safety beyond implementation.  

Real-world e-prescribing safety 

The ePRaSE tool was developed to assess how well live hospital e-prescribing systems mitigate known prescribing risk^​3​. It uses a bank of prescribing scenarios, each with predefined expected system responses, including both high-risk situations and no-harm controls.

Participating trusts manually enter these test scenarios into their local e-prescribing systems and record how the system responded. Responses are then categorised as “good mitigation” (where prescribing is prevented), “some mitigation” (for example, an alert is generated) or “no mitigation” (where no intervention took place). The tool also includes control scenarios, where no intervention is expected; if the system intervenes unnecessarily, this is scored as “over-mitigation”.

Between October 2022 and January 2023, 45 NHS trusts using 13 different prescribing systems (e.g EPIC, Oracle Health, System C) used ePRaSE to test e-prescribing systems^​3​. Although trusts completed different scenario sets, all were required to assess five mandatory extreme-risk scenarios based on “never events”, categorised by NHS England (see Table^​3​).  

Table: Mandatory high-risk prescribing scenarios and expected e-prescribing system responses assessed using the ePRaSE tool

The results showed that across all systems, the mean “good mitigation” score was about 50%, with wide variation between platforms. One e-prescribing system achieved an average “good mitigation” rate of about 60%, while another achieved closer to 20%, indicating that known high-risk prescribing scenarios were not consistently being prevented^​3​. Importantly, no individual trust mitigated all five mandatory extreme-risk scenarios, and fewer than half of trusts prevented inappropriate daily oral methotrexate prescribing, an error associated with severe toxicity, bone marrow suppression and potentially fatal outcomes^​4,5​.

Performance varied by clinical category. Drug allergy alerts showed the strongest mitigation, with more than 80% classified as “good mitigation”, whereas safeguards relating to omissions and laboratory linked prescribing were weaker^​3​. Even within the same e-prescribing systems, mitigation scores differed substantially between trusts, highlighting that local implementation plays a significant role in safety performance.  

Taken together, these findings suggest that prescribing safety is shaped not only by the choice of e-prescribing platform, but how systems are configured, governed and optimised in practice. The ePRaSE study provides quantitative evidence that implementation alone does not guarantee safety. The findings raise questions about the extent to which prescribing safety can be optimised locally, rather than requiring safer default functionality and stronger baseline safeguards from system suppliers.   

Beyond implementation: how do we define a safe e-prescribing system  

The findings of the ePRaSE study should be interpreted in context. Scores may have been influenced by how organisations interpret mitigation criteria, while organisations were not always assessed against the same prescribing risks, making direct comparisons more challenging^​3​. In addition, a lower score does not necessarily mean that safeguards are absent. A trust may have multiple layers of protection in place — such as structured order sets, prescribing restrictions and clinical decision support rules — yet still receive a lower score if a prescribing error remains technically possible^​3​.  

However, these limitations do not diminish the importance of the study. Instead, they expose a wider challenge in how digital medicines safety is assessed across the NHS.

National standards require organisations and suppliers to identify and manage clinical risk, but there is no national benchmark defining how effectively e-prescribing systems should mitigate common high-risk medication errors. This is a notable gap. It is difficult to think of another area of medicines safety where the NHS would invest so heavily in technology without defining what good safety performance looks like or how it should be measured. Consequently, organisations using the same system may implement different safeguards, while hospitals investing in different platforms have limited objective information on how their systems compare^​3,6​. 

The study highlights the challenge of balancing safety interventions against alert fatigue. ePRaSE found that systems demonstrating higher levels of mitigation were also more likely to exhibit over-mitigation, suggesting that stronger safeguards may come at the cost of increased interruptions to clinical workflows³. Alert fatigue is recognised as a patient safety issue, with evidence showing that excessive or low-value alerts are frequently overridden^​7,8​. Simply responding to poor performance by adding more alerts may therefore improve ePRaSE scores without necessarily improving overall safety. The more difficult challenge is designing systems that effectively mitigate risk while remaining usable in busy clinical environments^​7,8​. 

The response to digital safety remains largely reactive, with new safeguards frequently introduced following incidents or national alerts rather than through systematic prospective evaluation. Although ePRaSE is not a perfect measure of safety, prospective testing tools, such as ePRaSE, may represent an important step towards a more proactive approach to digital medicines safety.

Human factors in using technology 

While the ePRaSE findings may appear striking, they reflect challenges that have been described in the literature for over a decade. A UK-focused review found that e-prescribing systems can reduce some types of medication errors, but evidence of overall safety benefit is mixed, and unintended consequences are common^​9​. More recently, a multi-method research programme highlighted how workflow disruption, usability issues and local adaptation strongly influence outcomes of e-prescribing implementation^​10​.  

These findings help explain the wide variation observed in ePRaSE. Prescribing safety is shaped not only by the presence of an electronic system, but by how clinical decision support is configured, how order sets are designed, how alerts are prioritised and how the system integrates into busy clinical environments. Encouragingly, targeted optimisation can make a difference: reductions in high-risk prescribing can result from the implementation of well-configured clinical decision support tools^​11​. In particular, the system reduced errors relating to incomplete prescriptions and prescribing documentation by standardising order entry and requiring crucial prescribing information to be completed before a medication could be prescribed^​11​.  

Human factors play a critical role: e-prescribing systems do not operate in isolation but instead interact continuously with clinicians, workflows and organisational culture. In 2020, Aufegger et al. showed that user interface design directly affects prescribing behaviours, with poorly aligned systems encouraging workarounds that may introduce new risks^​7​. For example, if a prescriber has to navigate multiple screens to locate the correct medicine or formulation, the risk of selecting the wrong product may increase.  

From my experience of working in ePMA optimisation, many safety gaps sit quietly within configuration choices and often only become visible following incidents or targeted reviews. The ePRaSE findings quantify what many pharmacists recognise anecdotally.  

Why this matters for UK pharmacy and digital strategy  

These findings come at a critical time for the NHS and the pharmacy profession. National policies, such as the NHS ten-year health plan for England, increasingly position digital transformation as central to the future of healthcare delivery^​12​. Alongside this, the Royal College of Pharmacy has placed growing emphasis on digital capability within the workforce^​13​.  

In practice, pharmacy teams already sit at the centre of e-prescribing safety. Pharmacists configure clinical decision support, manage formularies, design order sets, support training and respond to incidents arising from system behaviour.

The ePRaSE findings reinforce the need for pharmacists to take a more proactive leadership role in digital medicines governance. Without routine assessment of how systems perform in real-world settings, safety gaps may persist unnoticed, particularly as configurations evolve over time. For pharmacy professionals, this represents both a responsibility and an opportunity to shape safer digital systems, influence organisational priorities and ensure that technological progress translates into meaningful improvements in patient care.  

While pharmacy teams play a crucial role in optimisation and governance, baseline system design and vendor-provided functionality shapes what is achievable. The variability observed within the same e-prescribing platforms indicates that prescribing safety is co-produced by healthcare organisations and system suppliers. Clearer national expectations of minimum safety functionality, greater transparency around system performance and procurement decisions informed by demonstrated safety capability could help drive improvement^​6​. 

Digital transformation offers enormous opportunity, but without systematic evaluation, unsafe configurations may persist unnoticed, allowing known high-risk prescribing errors to remain inadequately mitigated despite significant investment in digital systems. E-prescribing has transformed care, but its safety must be actively managed through ongoing optimisation, evaluation and pharmacist leadership. For hospital pharmacists, this means moving beyond simply implementing systems to continually challenging whether they are performing as safely as intended.