This summary is based on peer-reviewed literature and current clinical guidelines, cited below.
Background
Clinical decision support systems (CDSS) — computerized physician order entry (CPOE) with embedded alerts, medication reconciliation tools, and increasingly, AI-based review assistants — are now standard infrastructure in hospital and health-system pharmacy. Their promise is straightforward: catch prescribing errors before they reach the patient. Their real-world performance is more nuanced, and understanding where these tools help, where they fall short, and where pharmacist judgment remains essential is core to practicing safely alongside them.
Clinical Considerations
CPOE/CDSS and Medication Reconciliation: What the Evidence Shows
A 2021 Cochrane systematic review of 65 studies (n=110,875) evaluating interventions to reduce medication errors in hospitalized adults found that CPOE/CDSS probably reduces medication errors compared with paper-based systems, and that improved CDSS configurations probably reduce errors further compared with standard CDSS (moderate-certainty evidence).2 Pharmacist-led medication reconciliation showed a similar pattern: low-certainty evidence suggests medication reconciliation performed by pharmacists, versus other professionals, may reduce medication errors, and moderate-certainty evidence shows reconciliation probably reduces adverse drug events compared with no reconciliation.2 Notably, the review also found low-certainty evidence that prioritized (tiered) alerts prevented more adverse drug events than non-prioritized alerts — a finding with direct implications for how alert systems should be configured.2
Alert Fatigue Remains the Central Limitation
A systematic review of 39 studies on medication safety alert design found that the most common CDSS format — interruptive pop-up modals — was accepted least often by prescribers.3 Of the alternative designs evaluated (risk tiering, override justification requirements, shortcuts for common corrections), only role-tailoring — routing certain alerts specifically to pharmacists rather than prescribers — reliably increased acceptance.3 This has a practical implication for pharmacy workflow design: alert systems that leverage the pharmacist’s role in the medication-use process, rather than generic interruptive alerts fired at anyone in the ordering chain, are better supported by the evidence.
Where AI-Based Tools Fit
Large language models (LLMs) are being studied as a layer on top of traditional rule-based CDSS. In a 2025 prospective, cross-over study across 16 medical and surgical specialties (91 error scenarios derived from 40 clinical vignettes), an LLM-based CDSS used as a “co-pilot” alongside a pharmacist outperformed either the pharmacist alone or the LLM alone, improving accuracy in detecting prescribing errors and increasing detection of serious-harm errors 1.5-fold over the pharmacist working unassisted.1 The pharmacist-plus-LLM combination, not the LLM in isolation, produced the best results — evidence for augmenting pharmacist review rather than replacing it.1
Bottom Line for Pharmacists
CPOE/CDSS and pharmacist-led medication reconciliation have real, evidence-supported effects on reducing medication errors and adverse drug events, but the magnitude of benefit depends heavily on implementation quality — tiered, role-appropriate alerting outperforms blanket interruptive pop-ups. Emerging AI/LLM-based tools show the most promise as a pharmacist co-pilot rather than an autonomous replacement. When evaluating or advocating for new decision-support tools in practice, ask how alerts are prioritized and routed, not just whether an alert exists.
References
- Ong JCL, Jin L, Elangovan K, et al. Large language model as clinical decision support system augments medication safety in 16 clinical specialties. Cell Rep Med. 2025;6(10):102323. https://doi.org/10.1016/j.xcrm.2025.102323
- Ciapponi A, Fernandez Nievas SE, Seijo M, et al. Reducing medication errors for adults in hospital settings. Cochrane Database Syst Rev. 2021;11(11):CD009985. https://doi.org/10.1002/14651858.CD009985.pub2
- Hussain MI, Reynolds TL, Zheng K. Medication safety alert fatigue may be reduced via interaction design and clinical role tailoring: a systematic review. J Am Med Inform Assoc. 2019;26(10):1141-1149. https://doi.org/10.1093/jamia/ocz095
great post!