Evolution of Intravenous Medication Errors and Preventive Systemic Defenses in Hospital Settings—A Narrative Review of Recent Evidence

Objectives Intravenous drug administration has been associated with severe medication errors in hospitals. The present narrative review is based on a systematic literature search, and aimed to describe the recent evolution in research on systemic causes and defenses in intravenous medication errors in hospitals. Methods This narrative review was based on Reason’s theory of systems-based risk management. A systematic literature search covering the period from June 2016 to October 2021 was conducted on Medline (Ovid). We used the search strategy and selection criteria developed for our previous systematic reviews. The included articles were analyzed and compared to our previous reviews. Results The updated search found 435 articles. Of the 63 included articles, 16 focused on systemic causes of intravenous medication errors, and 47 on systemic defenses. A high proportion (n = 24, 38%) of the studies were conducted in the United States or Canada. Most of the studies focused on drug administration (n = 21/63, 33%) and preparation (n = 19/63, 30%). Compared to our previous review of error causes, more studies (n = 5/16, 31%) utilized research designs with a prospective risk management approach. Within articles related to systemic defenses, smart infusion pumps remained most widely studied (n = 10/47, 21%), while those related to preparation technologies (n = 7/47, 15%) had increased. Conclusions This narrative review demonstrates a growing interest in systems-based risk management for intravenous drug therapy and in introducing new technology, particularly smart infusion pumps and preparation systems, as systemic defenses. When introducing new technologies, prospective assessment and continuous monitoring of emerging safety risks should be conducted.

2][3][4] Observational multisite studies have reported a high prevalence of MEs and procedural failures, 5,6 while some have estimated that up to 10% of IV medication administrations include an error. 4When compared to other administration routes, it has been found that errors are as much as five times more likely to occur when an IV route is used. 3Moreover, many intravenously administered drugs are high-alert medications, bearing a heightened risk of causing significant patient harm if used in error. 7[10][11][12][13][14][15] Studies have found preparation and administration errors to be the most common IV ME types, probably due to these phases of the MMU process being the most widely studied. 4,16,17However, the number of IV prescribing errors is likely to be higher as, when also considering other administration routes, prescribing and monitoring are associated with the highest prevalence rates of preventable MEs. 18Wrong administration rate errors have been identified as the most common error type in IV drug administration, followed by the wrong time of administration, and the remaining error types include administration of wrong dose, inputting wrong pump settings, and dose omission. 4hen considering error defenses, a systematic review exploring IV preparation errors compared incidence by preparation site and/or method, finding the error incidence to be lower for doses prepared within a central pharmacy versus the nursing ward and lower for automated preparation versus manual preparation. 17In the same study, the error types and reported rates were found to vary substantially.
In 2016, we conducted systematic reviews exploring systemic causes of, 19 and systemic defenses to prevent, 20 IV MEs.Insufficient actions to secure safe use of IV high-alert medications, lack of knowledge of the drug, calculation tasks, failure in double-checking procedures, and confusion between look-alike sound-alike medications were identified as the leading causes of IV MEs. 19Most studies exploring systemic defenses to prevent IV MEs focused on the administration stage, with smart infusion pumps being the most widely studied systemic defense. 20A limited number of studies explored other stages of the MMU process, which represented a crucial area for future research.Moreover, most of the systemic defenses involved new technological solutions and features related to closed-loop medication management systems.The aim of the present narrative review was to describe the recent evolution of research on systemic causes of, and defenses to prevent, IV MEs in the MMU process in hospitals.A specific emphasis was to investigate whether the research on systemic causes has shifted more toward prospective risk management, and what the phases are of the IV MMU process to which the studied systemic defenses have been integrated.

Theoretical Framework
][23] In health care, risk management is defined as "clinical and administrative activities undertaken to identify, evaluate, and reduce the risk of injury to patients, staff, and visitors and the risk of loss to the organization itself." 22These activities or measures aim to prevent, remedy, or mitigate the occurrence or reoccurrence of an actual or potential patient safety event. 24The starting point of the theory is that errors are inevitable when human action is involved, which is why systems relying on perfect performance by individuals to prevent errors are doomed to fail.Hospitals are considered as high-reliability organizations which should introduce safety culture to learn from errors and to implement systemic defenses for ensuring safer care. 12,21,23In recent decades, medication risk management has evolved from retrospective error detection and a person-centered approach toward prospective risk management and a just culture environment, where the accountability of errors is divided between the systems and the individuals. 12,21,22,24,25

Search Strategy and Study Selection
This narrative review was carried out according to the Scale for the Assessment of Narrative Review Articles (SANRA), which is a brief critical appraisal tool for the assessment of nonsystematic review articles (Supplementary File 1, http://links.lww.com/JPS/A611). 26A systematic literature search was performed on Medline (Ovid) in October 2021.We used the same search strategy as in our previous publications 19,20 to review the most recent evidence.Medline (Ovid) was selected as a source of information as it represents a key database for medication safety research.Studies published within the period from 2016 to October 2021 were included.One reviewer (SK) selected the studies based on titles, abstracts, and full texts (Fig. 1).The PICO tool (participants, interventions, comparison, and outcomes) developed for our earlier publications 19,20 was used to select studies for inclusion.The search found 435 articles, of which 63 articles were included in the narrative review.

Data Extraction and Analysis
One author (SK) carried out the data extraction and analysis, and the results were carefully reviewed by the other authors (A-RH, MA).At first, reference, country, study design, setting, and the main results of the included studies were extracted to 2 tables (Supplementary Files 2, http://links.lww.com/JPS/A612, and 3, http://links.lww.com/JPS/A613).After that, a classification and comparison of frequencies were performed to summarize key findings of the articles included in this narrative review (n = 63) (Fig. 2).Finally, the key findings were compared to our previous systematic reviews 19,20 to illustrate the evolution of the IV medication MMU process in hospitals.
The articles on systemic causes of IV MEs (n = 16) were classified according to the stages of the MMU process they focused on (Fig. 2).The study designs employed in the articles were analyzed to recognize whether the studies aimed to identify ME risks prospectively or retrospectively (Fig. 2). 12,24Study designs focusing on learning from previous incidents to minimize similar errors in the future were defined as retrospective risk management (e.g., retrospective analysis of ME reports, observations, and medical records, or interviewing healthcare professionals about MEs they have encountered). 12,24,27][29] The number of studies utilizing a design considered as prospective risk management was compared between our earlier systematic review 19 and the present narrative review (Fig. 2).
The articles on systemic defenses to prevent IV MEs (n = 47) were classified according to the MMU process stages they focused on and compared to our earlier systematic review (Fig. 2). 20he evolution of systemic defenses was described using the "Swiss cheese model" comprising a central component of the systems-based risk management theory by James Reason (2000). 12inally, the main findings related to the most studied medication management technologies within all the included articles were summarized in 2 tables.
To investigate the types and frequency of intravenous medication errors associated with smart pumps in the USA.
60% of the observed infusions in 10 hospitals were associated with one or more MEs.The MEs were predominantly associated with violations of hospital policy (e.g., labeling errors, bypassing the smart pump and the drug library).Smart infusion pumps with DERS as a systemic defense (n = 8) Giuliano et al, 2018 55 A retrospective one-way analysis of variance and a descriptive overview of smart pump alerts.
To improve the overall understanding of IV smart pump drug library compliance by using the REMEDI data set to describe end-user compliance.
There are differences in IV smart pump compliance both within and between hospital systems.IV smart pump type and the number of drug library profiles may be influencing factors for smart pump compliance.Ibarra-Perez et al, 2021 80 A descriptive retrospective analysis of the smart pump alert reports.
To investigate the efficacy of IV smart pumps with drug libraries and DERS to intercept programming errors entailing high risk for patients in an adult ICU.
Drug library compliance was 70%.MEs were intercepted in 30% of infusions when using a drug library.Upper hard limit alerts accounted for 26% of pump reprogramming events.Jani et al, 2020 63 A retrospective review of ME reports and observed MEs.
To explore the role of smart infusion devices in preventing or contributing to medication administration errors using retrospective review of 2 complementary data sets.
Smart pumps both prevent and contribute to MEs.Using any infusion device rather than gravitational administration may have prevented 8-13% of MEs.EHR-integrated pumps could have prevented 52-73% of MEs.Manrique-Rodriguez et al, 2016 77 A prospective, observational interventional study with analytical components.
To estimate the impact of smart pump implementation in a PICU in terms of number and type of administration errors intercepted.
Drug library compliance was 84% and 283 MEs were intercepted for a study period of 62 months.A high-risk drug was involved in 58% of prevented MEs, which is why smart pump implementation was proven effective in intercepting programming errors of high-risk drugs.Moreira et al, 2020 75 A systematic review and meta-analysis.
To identify the scientific evidence for the frequency of handling errors of conventional and smart pump infusions in IV insulin therapy in ICUs.
An error rate of 10%-40% was associated with conventional pumps and 0-14% with smart pumps.Meta-analysis of 2 studies favored smart pumps to reduce the relative risk of programming errors by 51%.Palacios Rosas et al, 2019 79 A retrospective observational study with a pre-post design.
To evaluate the economic impact of the implementation of smart infusion pumps in the consumption of IV solutions in an ICU.
Implementation of smart infusion pumps allows savings by reducing the annual consumption of IV solutions measured in both units (18%) and liters (22.3%).Silva et al, 2019 74 A mathematical modeling for economic analysis to analyze cost-effectiveness.
To analyze cost-effectiveness and to calculate incremental cost-effectiveness ratio of the use of infusion pumps with drug library to reduce errors in IV drug administration in pediatric and neonatal patients in ICUs.
Infusion pump with a drug library may be the best strategy to avoid IV administration errors.
Although it has the lowest cost, the conventional pump also has lower effectiveness.

DISCUSSION
Our narrative review demonstrated a recent expansion of, and growing interest in, research related to medication safety of IV drug therapy.When compared to our previous systematic reviews, 19,20 the present updated literature search resulted in similar numbers of publications, although the search timeframe was shorter than in the original search, and the current search was conducted only in one database.While our previous review of systemic causes of IV MEs highlighted the need for more research into prescribing, preparation, and administration phases, 19 the present narrative review indicated that research interest in drug preparation and administration has remained, but research into causes of prescribing errors is still lacking.In addition, the transition from using retrospective research methods and study designs toward prospective risk management was identified in studies investigating systemic causes of IV MEs, because methods such as FMEA 32 and inductive PHA 41 were increasingly used in recent studies.This may represent a currently increasing trend of study focus in preventive IV medication risk management and has not been demonstrated in such an extent in previous literature.
In relation to studying systemic defenses to IV MEs, the stages of prescribing, preparation, and administration remained dominant. 20However, the identified widely expanding research interest in systemic defenses to secure safe IV drug preparation represents a key finding.This is central, as manual IV drug preparation has been identified as an especially high-risk task. 4,16,17If an IV workflow management system is not available, it is recommended that a second individual performs an independent verification of A retrospective review of drug library reports.
To establish baseline data to show how metrics in the set-up and programming phase of IV medication administration can be produced from medication library near-miss error reports from infusion pumps.
Drug library compliance was 74%.Wrong drug selections (LASA errors) represented 3% of drug library alerts and 22% of canceled infusions.Wrong dose selection was responsible for 3% of alerts and 19% of canceled infusions.Smart infusion pump interoperability with EHR as a systemic defense (n = 2) Furniss et al, 2020 69 A sociotechnical investigation utilizing in-depth qualitative observations and interviews.
To gain an in-depth understanding of the patterns of working that had evolved in an ICU where a closed-loop IV medication administration documentation system had been implemented and of the consequent effects on patient safety.
Several benefits were identified (e.g., hard and soft limits for different drugs, bolus limits for certain drugs, automatic recording of drug administration).Challenges included inadequate and outdated drug libraries, pump or medications not mapped with the EHR system, and inconsistency in dosing units between the drug library, EHR and usual pump-programming practices.Wei et al, 2021 62 Retrospective analysis of infusion pump log data.
To describe the patient safety and financial impact of pump-EHR interoperability at a community hospital.
Pump-EHR interoperability leads to safer administration of IV medications based on improved drug library compliance (74% versus 83%) and more accurate smart pump programming (infusions generating alerts 4% versus 3%).
DERS, dose error reduction software; LASA, look alike sound alike; ME, medication error; PICU, pediatric intensive care unit; REMEDI, Regenstrief National Center for Medical Device Informatics.

TABLE 2. A Summary of the Key Findings of Studies Investigating IV Preparation Workflow Technologies or IV Robotic Systems
(n = 8) 32,47,51,57,59,60,89,93 Reference Study Design Objectives Key Findings Systemic causes of MEs related to IV preparation workflow software (n = 1 study) Feemster et al, 2021 32 FMEA.
To identify potential failure points in a new chemotherapy preparation technology and to implement changes that prevent or minimize the consequences of those failures before they occur.
FMEA is useful for risk mitigation and workflow optimization prior to the implementation of an IV compounding technology.The chemotherapy workflow was defined as a 41-step process with 16 failure modes.IV preparation workflow software as a systemic defense (n = 5) Bucci et al, 2019 47 A retrospective analysis of TAWF error reports and manual error reports.
To evaluate the impact of a gravimetric-based TAWF system on the nonhazardous compounded sterile product error capture rate, production times, and pharmacy staff perceptions of compounding methods.
IV preparation with a gravimetric TAWF system is slower than manual volumetric preparation but can improve the error capture rate (41% versus <1%).Staff perceived the TAWF method to be the safest and most accurate.Eckel et al, 2019 51 A retrospective analysis of TAWF error reports and manual error reports.
To evaluate the benefits of TAWF compared with manual workflow (non-TAWF) on IV room efficiency, costs, and safety at hospitals with more than 200 beds.
The use of TAWF in the IV room was associated with detecting 14 times more MEs than the use of non-TAWF, demonstrating the different frequency of error in the results.TAWF also led to a faster preparation time with a lower cost.Higgins et al, 2019 57 A retrospective analysis of TAWF error reports and manual error reports.
To evaluate the benefits of TAWF compared to manual workflow (non-TAWF) on IV room efficiency, costs, and safety at community hospitals with less than 200 beds.
The TAWF sites detected errors at a significantly higher rate (4%) compared to the non-TAWF sites (<1%) ( P < 0.05).The top error-reporting category for the TAWF sites was incorrect medication (72%).The use of TAWF may be associated with a decrease in turnaround time and a decrease in overall cost.Lin et al, 2018 59 A retrospective analysis of the error report sample hospital information system database.
To determine the financial costs associated with wasted and missing doses before and after the implementation of an IVWMS and to quantify the number and the rate of detected IV preparation errors.
The adoption of the IVWMS significantly reduced the amount of wasted and missing IV doses from 21% to 7%.The overall cost savings of using the system was $144,000 over three months.The total number of MEs detected was 1,160 (1%) after using the IVWMS.Terkola et al, 2017 93 A large-scale, multicenter, multinational, retrospective study of preparation error reports.
To detect medication errors with possible critical therapeutic impact as determined by the rate of prevented medication errors in chemotherapy compounding after implementation of gravimetric measurement.
A gravimetric IV workflow system detected and prevented MEs that would not have been recognized with traditional methods.Eight percent of the antineoplastic drug doses had ME levels outside the accepted tolerance range.The proportion of doses with deviations >10% ranged from 0%-5% (mean 2.25%), and >20% ranged from 0%-1% (mean 0.71%).IV preparation robotic system as a systemic defense (n = 2) Amodeo et al, 2019 89 A simulated experimental controlled observational study.
To analyze the advantages of using the IV robotic system, compared to the manual preparation of injectable drugs in terms of accuracy, cost, and time in NICU.
The median error observed during reconstitution, dilution, and final therapy of the drugs prepared by the IV robotics ranged within ±5% accuracy, with narrower ranges of ME compared to manual preparation.The IV robotics consumed fewer materials, reduced costs, decreased preparation time, and optimized the medication process.Pang et al, 2021 60 A retrospective analysis of preparation reports from the IV preparation system's internal database.
To compare an IV gravimetric TAWF platform to an IV robotic system.
Implementing either an IV gravimetric TAWF system or an IV robotics system will result in similar compounding accuracy and precision.Preparation time was less with the use of the IV gravimetric TAWF versus the IV robotic system, but the IV robotic system required less human intervention.
FMEA = failure mode and effects analysis, IV = intravenous, IVWMS = intravenous workflow management system, ME = medication error, TAWF =technology-assisted workflow.manual additions to ensure that the proper medications (and diluents) are added, including confirmation of the proper volume of each medication (and diluent) before its addition to the final container. 15However, without technological assistance, manual double-check procedures have been recognized as complex, laborious, and rarely executed to full extent. 94,95Our present review identified a growing interest in IV preparation workflow software and robotic systems that could secure and, at least to some extent, replace risky manual work steps. 47,51,57,59,60,89,93Implementation of these systems has also been highlighted elsewhere, accompanied by technology utilized in telepharmacy operations. 15,27n addition to streamlining the preparation workflow and equipment, standardizing IV infusions is one of the key components of error prevention both in adult and pediatric populations. 10,65,68,72,76,96,97owever, the most powerful safeguard against preparation errors is to dispense IV medications in ready-to-administer packaging without further manipulation by the person administering the medication (e.g., withdrawing doses from containers or reconstituting powdered drug products). 27,96,97This trend was also noticeable in our review, as safety and cost-effectiveness of ready-to-use IV injections and infusions represented a new individual research area. 48,56,64,87,90Overall, our work shows that several alternative systemic defenses to secure safe preparation of IV medications are available, with the aim to either reduce or secure risky manual work steps.When it comes to IV medication guidelines, user-testing seems to be worthwhile to ensure practical and easy-to-understand instructions. 38mart infusion pumps remained one of the most widely studied technological defense in IV drug therapy. 55,63,74,75,77,79,80,82Examples of new perspectives include studies employing larger datasets collected from multiple hospitals 31,55 and smart infusion pump cost-effectiveness. 74In the literature, smart infusion pumps have been associated with alert fatigue resulting in insufficient compliance with drug library use and high override rates of soft limits. 10,20,98,99Other barriers include limitations in pump capabilities, availability of pumps in clinical care areas, unfunctional programming workflow, risks associated with secondary infusions, pump data analysis not utilized in quality improvement purposes, and deficiencies in drug library use and updates (e.g., omitting specific drugs or IV fluids). 10,100,101Despite these challenges, the use of smart pumps with dose-error reduction software is recommended for every drug infusion in all care settings as they may serve as an effective defense if their use is optimized. 10,29owever, implementation of new technologies may also create new, unexpected risks. 10,15,27,100This was observed in studies focusing on smart infusion pumps, 6,31,34 as well as IV workflow systems, 32 and patient-controlled analgesia. 43Especially smart pumps were found to introduce their own class of MEs that have not been considered in taxonomies of current ME reporting systems, which may hinder their detection. 34Consequently, it is important to evaluate possible new medication safety risks caused by defense implementation, as well as to improve and monitor their usability, which helps to avoid creation of risky shortcuts. 10,15,27,1000]29 Nonetheless, during the analysis process we observed that the current search strategy was not sensitive enough to detect all studies related to the EHR interoperability with smart infusion pump [102][103][104][105][106] and patient-controlled analgesia pumps. 107herefore, more targeted research objectives and search strategies are needed in future systematic reviews of new technologies to secure safe IV MMU process.
The findings of the present review are preliminary and indicative in nature, as the literature search was performed in only one database.Because of the increased use of technological solutions and closed-loop medication management systems, there is a continuous need to explore the emerging new risks throughout the IV MMU process.Moreover, it will be important to investigate effectiveness of systemic defenses in stronger outcomes research designs able to provide high quality evidence, both in terms of medication safety and cost effectiveness.As more data is available from technology integrated into MMU systems (e.g., compounding error reports within IV workflow management systems, smart pump alert log data, reports of EHR system-generated alerts), it is essential to use this information in academic research, practice development, and in knowledge-based leadership within hospital organizations.There is also a need for further studies to explore safety of IV drug administration in other settings apart from inpatient care.This would be crucial as IV medications are commonly administered in ambulatory settings, such as home infusion chemotherapy, pain management, and antimicrobial therapy.

CONCLUSIONS
This narrative review demonstrates that interest is still growing in systems-based risk management for IV drug therapy.The use of study designs and methods based on prospective risk management has increased in studies investigating systemic causes of IV MEs.This is a reasonable direction for development as safeguarding the IV MMU process has shifted toward application of technologies to replace or support manual error-prone work steps, which, in turn, may create new unexpected risks.In relation to studying systemic defenses to IV MEs, the stages of prescribing, preparation, and administration remained dominant.However, monitoring the usability of new defenses is recommended to ensure their feasible use and to avoid the development of shortcuts posing unnecessary safety risks.