Brief Review of Literature

https://doi.org/10.1177/00185787221145110

Hospital Pharmacy 2023, Vol. 58(3) 309 –314 © The Author(s) 2022 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/00185787221145110 journals.sagepub.com/home/hpx

Original Research Article

Introduction

A medication error is any preventable event that may cause or lead to inappropriate medication use or patient harm while the medication is in control of the health care professional or patient.1 Medication errors, although likely occurring since the dawn of medicine, does not appear in the literature until the 1960s. It was found that preparing a single-dose specifi- cally for a patient, also known as unit-dose packaging, decreased medication errors within hospital wards.2 Even less was known about medication errors within the operating room. It was accepted that errors occur in procedural areas however the etiology of these errors received little attention. Cooper et al conducted staff interviews and 359 preventable incidents were identified. Eighty-two percent of the prevent- able incidents were due to human error with drug syringe errors being one of the most common errors.3 This study

shows that medication errors are especially problematic within the operating room, however, this was based on staff recall after the medication error had occurred.

One major reason medication errors occur within the operating room is that anesthesia providers are typically the only practitioner involved in the medication-use process: prescribing, formulating, dispensing, administering, and monitoring.4 This process removes the risk reduction strat- egy of redundancies that exist in other hospital areas. To compound this dilemma, it also has the highest number of administered medications.5 Due to this, The Anesthesia

1145110 HPXXXX10.1177/00185787221145110Hospital PharmacyWolf et al research-article2022

1UK Healthcare, Lexington, KY, USA

Corresponding Author: Mark Wolf, UK Healthcare, 800 Rose Street, Lexington, KY 40536, USA. Email: [email protected]

Evaluation of Detected Medication Errors Within the Operating Room at an Academic Medical Center

Mark Wolf Jr.1 , Justin Rolf1, Dylan Nelson1, Devlin Smith1, and Elizabeth Hess1

Abstract

Background: Medication errors are preventable events that lead to inappropriate medication use and potential patient harm. This is especially prevalent within the operating room (OR) where one practitioner is involved in the entire medication- use process. Despite recent implementation of BD Pyxis™ Anesthesia ES, Codonics Safe Label System, and Epic One Step at the University of Kentucky Healthcare (UKHC) to prevent medication errors, errors continue to be reported. Curatolo et al found human error was the most frequent cause of medication error within the OR. Clumsy automation may be an explanation for this, which imposes burdens and promotes work arounds. This study endeavors to assess potential medication errors via chart review to identify risk reduction strategies. Methods: This a single-center retrospective cohort review of patients admitted to a UK HealthCare Main Operating Room, defined OR1A-OR5A and OR7A-OR16A, who were administered medications from 8/1/2021 to 9/30/2021. Results: Over a 2-month period, 145 cases were conducted at UK HealthCare. Of the 145 cases, 98.6% (n = 143) involved a medication error and 93.7% (n = 136) of the errors involved a high- alert medication. The top 5 classes of drugs involved in errors were all high-alert medications. Lastly, 46.6% (n = 67) of cases had documentation that Codonics was utilized. In addition to analyzing medication errors, the financial analysis found that $3154.04 in drug cost was lost in the study period. When extrapolating these results to all BD™ Pyxis Anesthesia Machines at UK HealthCare, $107 237.36 of drug cost is potentially lost per year. Conclusions: These findings add to previous data that have described the increased rate of medication errors when conducting chart review rather than rely on self-reported data. In this study, 98.6% of all cases involved a medication error. In addition, these findings provide additional insight in the increased use of technology within the operating room despite medication errors still occurring. These results can be applied to like institutions to critically evaluate anesthesia workflow to determine risk reduction strategies.

Keywords CQI, drug information, medication errors, medication process, medication safety, operating room pharmacies

310 Hospital Pharmacy 58(3)

Patient Safety Foundation (APSF) identified changes and developed methods for improving the practice of anesthesia. The APSF developed a new paradigm based on 3 principles: standardization, technology, and prefilled or premixed syringes to help facilitate workflow and reduce medication errors. More specifically, the recommendations were stan- dardization of the anesthesia workplace, adoption of tech- nology such as barcode medication administration (BCMA), and addition of a clinical pharmacist to the operating room team to facilitate preparation and standardization of medications.6

Various technologies have been implemented to reduce medication errors. BD Pyxis™ Anesthesia Station ES was developed to provide safe clinical workflows and security enhancements that promote medication safety and compli- ance efforts.7 These anesthesia stations allow for increased medication safety and availability of medications for anes- thesia providers, control, and security for information tech- nology (IT), and ability to further build a medication management system through partnerships with other tech- nologies. One of the partnerships is Codonics Safe Label System (SLS). The Safe Label System utilizes barcode tech- nology to scan information from a drug container, electroni- cally verify it against the hospital-approved drug database, audibly and visually confirm the medication, and prints a TJC compliant label.8 Additionally, the electronic medical record also has inherent safety features compared to paper- based charting. One example is Epic One Step, which is a real-time order form that prompts providers to scan the patient bracelet, select which medications are being adminis- tered, and directly document into the chart. Theoretically, these interventions should create a safer environment. However, studies have shown that computer systems can cre- ate new burdens and complexities for those operating high- consequence systems.9

UKHC is a 900-bed tertiary academic medical center that completes approximately 150 procedures requiring anesthe- sia daily. In August 2020, BD Pyxis™ Anesthesia ES and Codonics Safe Label System were implemented in the peri- operative areas. In addition, UKHC underwent an electronic medical record change in June 2021 that moved anesthesia documentation away from paper-based charting to Epic One Step. Prior to implementation of these new technologies, it was not possible to effectively audit medication errors in the perioperative space. Even with these interventions in place, medication errors are still being reported to the internal med- ication event reporting software. Due to the propensity for self-reported data to underestimate true rates of medication errors, this study endeavors to objectively assess medication errors via chart review as defined by the NCC MERP Taxonomy.

Methods and Materials

This is a retrospective cohort review of patients admitted to a UKHC Operating Room between 8/1/2021 and 9/30/2021.

Included patients were ≥18 years of age and had medica- tions administered while in a main operating room defined as locations OR1A-OR5A and OR7A-OR16A. Patients were excluded from the study if they underwent trauma or emer- gent cases as defined as procedural location OR6A or under- went procedures in any locations outside the specified Main Operating Room spaces. Data was extracted via chart review of Epic Hyperspace Medical Record, BD Pyxis™ Enterprise Server, and Codonics Safe label System Administration Tools. The wholesale acquisition cost (WAC) was then uti- lized to determine the total cost lost from medications that were dispensed from Pyxis, not administered or charged to the patient, and no record of return back into Pyxis.

The primary endpoint was to determine the rate of any detected potential medication error. The secondary endpoints were to determine the potential event type based on NCC MERP Taxonomy, the potential harm score for each medi- cation error based on the NCC MERP Index for Categorizing Medication Errors, the rate of potential medication errors with high-alert medications as defined by the Institute for Safe Medication Practices, and the rate of labeling requirements based on the Joint Commission Medication Management Standards. A post-hoc analysis was conducted to determine the financial impact of the unaccounted-for medications that had no associated patient charge utilizing Wholesale Acquisition Cost (WAC).

BD Pyxis™ Enterprise Server dispense data was recon- ciled with Epic Hyperspace Medical Record One Step administration report and Codonics Safe Label System. BD Pyxis™ Enterprise Server and Epic One Step Administration report were reconciled via patient ID and Codonics Safe Label System was reconciled via date of scanning. Any dis- crepancies detected between reports is what the authors would then categorize as a potential medication error. The term “potential medication error” was utilized as the method of chart review relied on documentation of what anesthesia providers did during the case rather than direct observation of what occurred. Discrepancies between reports were not analyzed if the medication is not stocked within the Main Operating Room Pyxis machine. Events were categorized as “dose omission” when there was documentation of medica- tion administration but no associated Pyxis dispense. “Wrong drug” events occurred when medications were dispensed from Pyxis, not documented as administered in Epic One Step, and had no documentation of returning the drug back to Pyxis or wasting. “Wrong dose—overdose” occurred when the documented administered dose is less than what was dis- pensed from Pyxis. Lastly, “Wrong dose—under Dose” occurred when the documented administered dose is greater than what was dispensed from Pyxis. The event type of “other” was utilized when an event occurred that did not fit these definitions (Table 1).

Descriptive Statistics was used to evaluate the rate of potential medication errors. The primary and secondary end- points were modeled via descriptive statistics and reported as a percentage rate. The potential number of errors per drug

Wolf et al 311

class, potential NCC MERP Event Type, and potential NCC MERP Harm Score were represented via a bar graph. The percentage financial loss of charge omissions was deter- mined via adding the total cost of one Main Operating Room Pyxis Station and dividing the cost of medications catego- rized via the Event Type “Wrong Drug.” The charge omis- sion data was then extrapolated to all Pyxis Anesthesia Stations at UKHC to report a total amount of drug cost lost over 1 year. Microsoft Excel was utilized for descriptive statistics.

Results

Over a 2-month period, 8/1/2021 through 9/30/2021, 145 cases and 785 potential medication errors were included for the analysis of the primary and secondary endpoints. 98.6% (n = 143) of cases had an event that could have resulted in a medication error. The average number of potential medica- tion errors per case was 7 (0-18) and the median was 6. 93.7% (n = 136) of cases had a potential medication error that involved a high-alert medication. 46.2% (n = 67) of cases had record that Codonics was utilized to obtain a label (Table 2).

The most common medication classes involved in a medi- cation error were vasopressors (n = 118), opioids (n = 118),

neuromuscular blockers (n = 96), sedatives (n = 66), and anesthetics (n = 59) (Figure 1). The most common medica- tions involved in an error were fentanyl (n = 88), rocuronium (n = 73), propofol (n = 66), phenylephrine (n = 48), and midazolam (n = 44). The most common potential Harm Score and Event Type were Category F (n = 257) and Dose Omission (n = 354) (Figures 2 and 3).

A post-hoc analysis was conducted to determine the financial impact of the unaccounted-for medications that did not have a patient charge. Thirty-seven medications were included in the financial analysis that were classified as a “wrong drug” event type. The most expensive medication kept within the Pyxis ES™ Anesthesia Machine was sug- gamadex ($107.96 WAC) and the least expensive medica- tion was ketorolac ($0.83 WAC). The estimated total drug cost of one Pyxis Anesthesia station is $7730.77. The total amount lost over the 2-month period in the specified Pyxis ES™ Anesthesia Machine in the Main Operating Rooms was $3154.04 total or $1577.02 per month. This result can be extrapolated to all 85 BD™ Pyxis Anesthesia Station at UK HealthCare for a yearly drug cost loss of $107 237.36 total or $1261.62 per station. This accounts for roughly 16% of the total cost of one Pyxis station that is lost in drug cost alone.

Table 1. Results of Medication Error Surveillance in 145 Operating Room Cases.

Endpoint Number (%)

Potential medication errors 785 Potential medication errors with high-alert medications 526 (67%) Potential medication error per case 143 (98.6%) Potential medication errors per case that involved a high-alert medication 136 (93.7%) Average number of potential medication errors per case 7 (range 0-18) Median number of potential medication errors per case 6 Documentation of codonics utilization per case 67 (46.2%)

Table 2. NCC MERP Taxonomy* Definitions and Examples of Medication Error Types.

NCC MERP event type Definition Example

Dose omission Documentation of medication administration but no associated Pyxis dispense

Provider documented administration of rocuronium but did not dispense from Pyxis

Wrong dose— overdose

The documented administered dose is less than what was dispensed from Pyxis without an associated return or waste

Provider dispensed 5 mL of fentanyl, documented administration of 3 mL, and no documentation of wasting

Wrong dose—under dose

Documented administered dose is greater than what was dispensed from Pyxis

Provider documented administration of 7 mL of hydromorphone but dispensed 1 mL

Wrong technique When there is no record of utilizing Codonics obtain a label

Provider documented administration of fentanyl but no associated Codonics data of scanning is found.

Wrong drug Medication is dispensed from Pyxis but not documented as administered in One Step with no documentation of returning back to Pyxis

Provider dispensed a vial of sugammadex, did not administer to the patient, and did not return back into Pyxis at end of case

Other Event occurred that did not meet any of the other definitions

Provider pulled and wasted remifentanil but no documentation the drug was administered

Note. *The National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) taxonomy is a guidance tool to aid in the categorization and coding of medication errors to provide standard language and structure in developing medication error databases.

312 Hospital Pharmacy 58(3)

Discussion

Within UKHC, various pharmacy technologies have been implemented to reduce medication errors including BD Pyxis™ Anesthesia Station ES, Codonics Safe Label System, and Epic One Step. Even with all these system interventions designed to create a safer environment, we continue to see medication errors in the operating room. A retrospective review at a large tertiary care center from 2007 to 2015 sought to distinguish preventable adverse events from non- preventable events based on self-reported data. After case review, a total of 747 cases were included in final analysis

and about half (42.8%) were deemed to be preventable. The study identified that system causes were more frequent for overall errors, however, human causes was most frequent when it came to medication errors.10 The cause of these errors, that being human error, is defined as a failure of a planned action to achieve a desired outcome.11 This trend of human causes with medication errors can potentially be due to clumsy automation. Clumsy automation is a form of poor coordination between human and machine, whereas costs and burdens imposed by technology occur at periods of peak workload.12 As a healthcare institution, it is important to remember that human must not be blamed. Rather, we must

Figure 3. NCC MERP potential event type in 145 operating room cases (N = 785). Note. NCC MERP = national coordinating council for medication error reporting and prevention.

Figure 1. Number of errors by drug class in 145 operating room cases (n = 785). Note. *Other class includes drug classes who’s total equaled less than 10. Includes anticholinergics (n = 7), insulins (n = 7), vasodilators (n = 5), diuretics (n = 5), Anticoagulants (n = 4), Acetylcholinesterase inhibitors (n = 3), Antihistamines (n = 2).

Figure 2. NCC MERP potential harm score in 145 operating room cases (N = 785). Note. NCC MERP = national coordinating council for medication error reporting and prevention.

Wolf et al 313

recognize that errors made by individuals may be due to a flawed system.13 Unfortunately, it must also be recognized that the true rate of medications errors is likely much higher.

One major limiting factor when it comes to medication error identification within the operating room is that it is typically based on self-reported data. Passive reporting sys- tems, which rely upon voluntary reports from staff, are known to result in far fewer medication error reports than active surveillance systems.1 Jha et al and colleagues sought to quantify this difference. A trained reviewer would exam- ine patients’ hospital records when a computer-based moni- toring program identified alerts to determine if an adverse drug event (ADE) occurred. The study found that chart review identified 398 ADE’s compared from 23 from volun- tary reporting.14 Based on recommendations from NCC MERP, there is no national benchmark for accepted rate of medication errors. However, the ISMP White Paper states that there is at 5% error rate of adverse events and/or medica- tion errors in surgeries.15 Within this study, the authors found that 98.6% of cases had a potential medication error, much higher than the potential rate ISMP suggests. While error rates should not be benchmarked, it is imperative hospital leaders utilize this information to critically evaluate the sys- tems put in place to decrease medication errors and lessen the burden of clumsy automation.

To the authors knowledge, this is one of the first studies that sought out to objectively assess potential medication errors within the operating room via chart review. BD™ Pyxis Enterprise Server dispense data, which served as a surrogate of what anesthesia providers could physically administer, was primarily reconciled with Epic One Step administration report, which served as what the anesthesia provider meant to administer. An important point to note within the methods of this study is that only potential medi- cation errors were assessed. An example of a discrepancy between both reports would include documented administra- tion of 15.43 mL of hydromorphone but only 1 mL was docu- mented as dispensed from Pyxis. In this example, the potential NCC MERP event type would be “Wrong Dose— underdose” and assigned a potential NCC MERP Harm Score of “E.” “E” was assigned due to the potential risk of the patient being in pain that was not being adequately con- trolled due to the anesthesia providers only physically hav- ing 1 mL of hydromorphone to administer.

While the methods utilized in the study prevented the investigators from identifying true Harm events, the utiliza- tion of potential Harm Score can provide valuable insight of the risk that can occur within the operating room. The use of potential harm showed immense value in an initiative con- ducted by Vizient. Vizient implemented a standardized pro- cess for interception and documentation of prescribing error by pharmacists at order verification and would classify potential errors based on a modified NCC-MERP Index for potential capacity for harm. The initiative found that 46.6% interventions prevented potentially serious consequences

and a potential in $874 000 avoided.16 In addition to the high rate of potential medication errors per case, this study found that 93.7% of all cases had a medication error that involved a high-alert medication. High-alert medications are drugs that bear heightened risk of causing significant patient harm when they are used in error.17 In this study, the top 5 most common classes of medication involved in medication errors are high-alert. Identifying the potential harm that could result from medication errors is imperative, especially when best practices include placing special safe guards to reduce the risk of error for high-alert medications.18

While the impact of medication errors and resulting patient harm is vital, it is well documented that errors have a lasting financial impact within the health system. The Betsey Lehman Center for Patient Safety sought to quantify this impact. The center utilized Medicare data from 2013 and found $518 million in excess insurance claims were associ- ated with patient harm.19 Unfortunately, this excess cost does not specify the cause of the harm event, rather, lumps every potential cause of harm such as medication errors, falls, and hospital associated infection. The authors found approxi- mately a 16% loss of the total inventory cost of one BD™ Pyxis Anesthesia Station per year. This equates to roughly $107 000 loss in drug cost alone per year. It is important to realize this loss only accounts for the WAC pricing of medi- cation and not the loss in revenue for the health system, which is likely much more.

This study provides additional insight into the challenges faced within the operating room regarding medication safety. Particularly when it comes to the use of technology and the reliance on the user to appropriately document their actions. Despite these merits, the study and its findings are not with- out limitations. The primary limitation of this study includes its retrospective design and its reliance on Pyxis, Epic, and Codonics data. The use of this method prevents the authors from identifying true harm events, rather, assume the worst- case scenario with every discrepancy. The data utilized by the authors relied on anesthesia staff to appropriately docu- ment dispensed and administered medications. There is the possibility that anesthesia staff dispensed and administered the appropriate medication and dose, however, did not appro- priately document their actions. Even with this limitation, it is important to realize that improper documentation has last- ing legal implications.20 The authors of this study were unable to reconcile Codonics data via patient ID due to limi- tations with Pyxis, Epic, and Codonics interface capabilities. Rather, the use of date of scanning was used to reconcile the data. While this is a significant limitation in identifying appropriate labeling in the context of this study, it also proves to be a significant limitation in the use of Codonics in general that should be addressed by the manufacturer. Lastly, the authors did not evaluate the loss of revenue experienced by the health system due to unaccounted for drug. The loss of revenue would most likely have provided a higher financial impact; however, the drug cost alone is not insignificant.

314 Hospital Pharmacy 58(3)

Conclusion

At UKHC, 145 cases occurred within the main Operating Rooms from 8/1/2021 to 9/30/2022. Of those 145 cases, 98.6% (n = 143) involved a medication error and 93.7% (n = 136) of the errors involved a high-alert medication. The top 5 classes of drugs involved in errors were all high-alert medications. Lastly, 46.6% (n = 67) of cases had documenta- tion that Codonics was utilized. These findings add to previ- ous data that have described the increased rate of medication errors when conducting a chart review rather than reliance on self-reported events.10,14 These findings provide additional insight in the increasing use of technology within the operat- ing room, yet, medication errors are still occurring.10,12 These results can be applied to like institutions to determine risk reduction strategies to help improve anesthesia workflow, particularly when it comes to the interface between human and machine.

Acknowledgments

Aric Schadler, PhD, MS.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, author- ship, and/or publication of this article.

ORCID iD

Mark Wolf https://orcid.org/0000-0002-9371-7056

References

1. NCC MERP. 2021. National Coordinating Council for Medi- cation Error Reporting and Prevention. Accessed September 9, 2021. https://www.nccmerp.org/

2. Barker KN. Unit-dose injectables and the hospital medication error problem drug distribution in hospitals. Unit-dose inject- ables and the hospital medication error problem. Bull Parenter Drug Assoc. 1966;20:157-165.

3. Cooper JB, Newbower RS, Long CD, McPeek B. Preventable anesthesia mishaps: a study of human factors. Anesthesiology. 1978;49:399-406.

4. Wahr JA, Abernathy JH, Lazarra EH, et al. Medication safety in the operating room: literature and expert-based recommen- dations. Br J Anaesth. 2017;118(1):32-43.

5. Brown LB. Medication administration in the operating room: new standards and recommendations. AANA J. 2014;82(6): 465-469.

6. Eichborn JH. APSF hosts Medication Safety Conference: con- sensus group defines challenges and opportunities for improved practice. APSF Newslett. 2010;25(1):1-8.

7. BD. BD Pyxis Anesthesia Station ES: A Platform Approach to OR Medication Management. BD; 2018.

8. Codonics. Label Me Safe. Codonics; 2014. 9. Cook RI, Woods DD. Adapting to new technology in the oper-

ating room. Hum Factors. 1996;38:593-613. 10. Curatolo CJ, McCormick PJ, Hyman JB, Beilin Y. Preventable

anesthesia-related adverse events at a large tertiary care center: a nine-year retrospective analysis. JT Comm J Qual Saf. 2018;44:708-718.

11. NOPSEMA. 2020. Human error. Accessed March 8, 2021. https://www.nopsema.gov.au/offshore-industry/safety/human- factors

12. Sarter NB, Woods DD. How in the world did we ever get into that mode? Mode error and awareness in supervisory control. Hum Factors. 1995;37:5-19.

13. PSNet. 2019. Systems approach. Accessed April 8, 2021. https:// psnet.ahrq.gov/primer/systems-approach#:~:text=The%20 systems%20approach%20holds%20that,distinguish%20indi- vidual%20from%20system%20errors.

14. Jha AK, Kuperman GJ, Teich JM, et al. Identifying adverse drug events: development of a computer-based monitor and comparison with chart review and stimulated voluntary report. J Am Med Inform Assoc. 1998;5(3):305-314.

15. ISMP. A call to action: The case for Medication Safety Officers (MSO)—A white paper. ISMP; 2018:1-15.

16. Cabri A, Barsegyan N, Postelnick M, et al. Pharmacist inter- vention on prescribing errors: use of a standardized approach in the inpatient setting. Am J Health Syst Pharm. 2021;78: 2151-2158.

17. ISMP. ISMP list of high-alert medications in acute care set- tings. ISMP; 2018:1-1.

18. Billstein-Leber M, Carrillo CJD, Cassano AT, Moline K, Robertson JJ. ASHP guidelines on preventing medication errors in hospitals. Am J Health Syst Pharm. 2018;75(19): 1493-1517.

19. Fain B, Healey M, Sudders M, Palleschi E, Campbell R. The Financial and Human Cost of Medical Error and How Massachusetts Can Lead the Way on Patient Safety. Betsy Lehman Center for Patient Safety; 2019.

20. Yankowsky KW. Avoiding unnecessary litigation: communi- cation and documentation. Adv Skin Wound Care. 2017;30(2): 66-70. doi:10.1097/01.ASW.0000511697.33786.fa

Copyright of Hospital Pharmacy is the property of Sage Publications Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.