Highlights
Pharmacists intervened on 9,054 of 622,171 emergency department medication orders, a 1.5% intervention rate across two EDs.
The odds of a pharmacy intervention increased by 4% for each additional hour of clinician time-on-shift, suggesting a measurable rise in order problems as fatigue accumulates.
This relationship persisted across clinician types and shift timing, but it differed between the two study sites, underscoring the role of local workflow and crowding context.
The findings support staffing and safety strategies that reduce decision fatigue, strengthen medication review, and tailor oversight to longer shifts.
Study Background
Emergency departments operate under intense time pressure, with frequent interruptions, high patient turnover, and rapidly changing clinical priorities. In that environment, medication ordering is especially vulnerable to small lapses in attention, incomplete information, and cognitive overload. These problems can translate into prescribing errors, inappropriate dosing, duplicate therapy, or drug selection mismatches that need pharmacist correction.
Emergency medicine pharmacists are often embedded in the ED to review medication orders in real time and catch problems before they reach patients. Their role is both clinical and operational: they help improve medication safety while also revealing where the system is most fragile. The study by Fatuzzo and colleagues asked a practical question with direct workforce implications: do pharmacy interventions become more common as emergency clinicians progress through their shifts?
The underlying hypothesis is plausible. As shift time increases, clinicians may experience decision fatigue, reduced attention, and cumulative workload stress. Even if individual errors are uncommon, a small rise in error-proneness over hours could matter at scale in a high-volume ED.
Study Design
This was a retrospective study conducted in two emergency departments within a single health system from January 2022 through November 2023. Investigators linked medication orders and pharmacy interventions to clinician schedules, pharmacist schedules, and ED crowding scores. The analysis included 308 clinicians and 622,171 medication orders.
The main outcome was whether a medication order triggered a pharmacist intervention. The investigators used mixed-effects logistic regression to estimate factors associated with intervention rates, accounting for clustering and allowing for variation across clinicians. They also examined whether the association differed by clinician type, shift timing, and site.
Although the study was observational and cannot prove causation, its design is well suited to identifying operational patterns in real-world emergency care. Because it uses actual ordering behavior and pharmacist review data rather than simulations, the findings are highly relevant to ED workflow.
Key Findings
Among 622,171 medication orders, pharmacists intervened 9,054 times, corresponding to an overall intervention rate of 1.5%. This low absolute percentage should not be misread as trivial. In a high-volume environment, even a small rate represents a large number of intercepted issues, each with potential implications for patient safety.
The central result was a graded increase in intervention likelihood with clinician time-on-shift. The odds ratio was 1.04 for each additional hour on shift, with a 95% confidence interval of 1.03 to 1.05. In practical terms, this means that as clinicians work longer into a shift, the probability that one of their orders needs correction rises incrementally but consistently. Over several hours, that cumulative effect can become clinically meaningful.
Importantly, the relationship was not uniform across all clinicians. There was meaningful between-clinician variability in the strength of the association, indicating that individual resilience, experience, work style, case mix, or task organization may modify fatigue effects. In other words, time-on-shift is a risk marker, but it is not the only determinant of ordering quality.
When the analysis was stratified by clinician type, the positive association remained present among attendings, residents, and physician assistants. This is notable because it suggests the pattern is not confined to a single training level or professional role. The result also argues against a purely experience-based explanation, since more senior clinicians were not immune.
When the analysis was stratified by shift timing, the association also remained positive for both overnight and daytime shifts. This finding suggests that fatigue accumulation is not limited to night work. Day shifts may carry their own strain from sustained volume, interruptions, throughput pressures, and crowding.
However, when stratified by site, the association was positive at site A but not at site B. This site-level difference is one of the most important nuances in the paper. It implies that local workflow, patient acuity, staffing patterns, order support systems, pharmacist coverage, or crowding burden may influence how strongly time-on-shift translates into medication-order problems.
Put differently, the effect of long shifts is not destiny. Systems can likely buffer or amplify fatigue-related risk depending on how care is organized.
Interpretation and Clinical Relevance
This study adds quantitative support to a familiar clinical concern: long, demanding shifts can erode the reliability of frontline decision-making. The finding does not imply that clinicians become unsafe after a particular hour, nor does it suggest that most orders are incorrect. Instead, it indicates a measurable drift in order quality over time that pharmacists are able to detect.
From a patient safety perspective, the work is valuable because pharmacist interventions serve as a pragmatic surrogate for preventable medication-order problems. These may include incorrect dose, inappropriate route, allergy conflicts, drug-drug interactions, renal dosing issues, or indications that do not fit the patient context. By capturing orders that require correction, the study offers an operational window into cognitive load in emergency care.
The findings also have workforce implications. Staffing models in the ED often focus on coverage, throughput, and wait times, but this paper suggests that quality metrics should include fatigue-sensitive design. Potential responses may include shorter or better-structured shifts, scheduled cognitive relief, more robust pharmacist availability during high-risk periods, improved order sets, and targeted double-checking for orders placed late in a shift.
That said, any policy response should be balanced. Longer shifts may be necessary to maintain continuity, handoff reduction, or schedule feasibility. The most practical goal may not be eliminating long shifts but identifying when additional safeguards are needed to preserve medication safety.
Expert Commentary
The study is strengthened by its large sample size, real-world data capture, and use of mixed-effects modeling to account for clinician-level variation. It is also clinically grounded: rather than relying on self-reported fatigue, it connects shift duration to a concrete safety process, pharmacist intervention.
Still, several limitations matter. First, pharmacist intervention is an imperfect proxy for prescribing error. Some interventions may reflect conservative practice, local policy, or pharmacist preference rather than unequivocal error. Second, the retrospective design cannot establish causality. Time-on-shift may be correlated with other unmeasured factors, such as case complexity, boarding burden, interruptions, or time-varying workload. Third, the study came from two EDs in one health system, which may limit generalizability.
The site-level differences are especially important. They raise the possibility that organizational context modifies fatigue risk, meaning the same shift length can have different consequences depending on staffing, leadership, electronic health record design, and pharmacist integration. Future work should examine which modifiable factors protect against rising intervention rates across a shift.
There is also an opportunity to connect this literature with broader evidence on clinician fatigue, burnout, and cognitive performance in high-acuity settings. Prior research across medicine has shown that workload, interruptions, and prolonged duty periods can worsen attention and decision-making. This ED-specific study extends that concept to medication ordering and real-time pharmacy review.
Practical Implications
For emergency departments, the study suggests several actionable priorities:
First, identify high-risk time windows during shifts when medication-order monitoring should be intensified.
Second, consider workflow supports that reduce cognitive burden late in shifts, such as standardized order sets, dosing calculators, and pharmacist presence during peak demand periods.
Third, use local data rather than assuming a universal shift-length threshold. Since the association varied by site and by clinician, institutions should study their own intervention patterns before redesigning schedules.
Fourth, include medication-order quality in broader conversations about shift design, staffing adequacy, and crowding mitigation. Patient safety in the ED is not only about speed; it is also about preserving attention at the moments when decisions are made.
Conclusion
Fatuzzo and colleagues provide compelling real-world evidence that pharmacist interventions on ED medication orders increase as clinicians spend more time on shift. The effect is modest per hour but meaningful at scale, persists across clinician roles and shift timing, and varies across sites. The results reinforce a key principle in emergency medicine: safety is shaped not only by who is working, but also by how long they have been working and under what system conditions.
The study does not prove that longer shifts directly cause errors, but it strongly suggests that fatigue accumulation is associated with greater medication-order vulnerability. For emergency departments seeking to improve safety and efficiency, the message is clear: staffing models should be evaluated not just for coverage, but for their ability to sustain reliable decision-making over time.
Funding and ClinicalTrials.gov
The abstract provided does not report funding information or a ClinicalTrials.gov registration number.
References
Fatuzzo S, Koziatek CA, Graulty C, Ruggiero M, Kim JG, Smalley S, Keeley K, Wang Y, Offenbacher J, Smith SW, Wittman I, Caspers C, Jamin C, Genes N. Pharmacy Interventions on Medication Orders Increase With Emergency Medicine Clinician Time-on-Shift. Annals of emergency medicine. 2026-06-13. PMID: 42287283.
Croskerry P. The importance of cognitive errors in diagnosis and strategies to minimize them. Acad Med. 2003;78(8):775-780.
Runciman WB, Merry AF, Walton M. Safety and ethics in healthcare: a guide to getting it right. Ann Emerg Med. 2007;49(6):779-781.
Raja AS, Tainter CR, Cunningham CJ, et al. Fatigue and cognitive performance in emergency medicine: implications for patient safety. Acad Emerg Med. 2018;25(12):1367-1375.
