PRACTICE REPORTS Barcode technology
PRACTICE REPORTS
Effect of barcode technology with electronic medication administration record on medication accuracy rates Heather H. Seibert, Ray R. Maddox, Elizabeth A. Flynn, and Carolyn K. Williams
M
edication errors and consequent deaths continue to escalate, costing health care systems billions of dollars each year.1 It is estimated that at least 1 medication error occurs per day per hospitalized patient. An estimated 450,000 adverse drug events—medication errors that result in patient harm— occur annually, approximately 25% of which are preventable.2 Various technologies have been introduced to help improve the accuracy of medication administration, including automated dispensing cabinets, computerized prescriber order entry (CPOE), “smart” (computerized) i.v. infusion pumps, barcodeassisted medication administration (BCMA) systems, electronic medication administration records (eMARs), and wireless connectivity and integration with hospital information technology. A recent report on the adoption rates of medication-safety technologies revealed that health care organizations have implemented smart pumps more than other technologies.3 BCMA is the second most
Purpose. The effect of barcode-assisted medication administration (BCMA) with electronic medication administration record (eMAR) technology on the occurrence of medication administration errors was evaluated. Methods. A pretest–posttest nonequivalent comparison group was used to investigate the effect of BCMA-eMAR on the medication administration accuracy rates at two community-based hospitals. Patient care units included three matched pairs in the two hospitals—two medical– surgical, two telemetry, and two rehabilitation units—plus a medical–surgical intensive care unit, an emergency department, and both an inpatient oncology unit and an outpatient oncology service at one of the hospitals. Medication administration accuracy rates were observed and recorded before (phase 1) and approximately 6 and 12 months after (phases 2 and 3, respectively) the implementation of BCMA-eMAR. Results. The overall accuracy rate at hospital 1 increased significantly from phase 1 (89%) to phase 3 (90%) (p = 0.0015); if
commonly implemented technology, followed by CPOE with decision support.
Heather H. Seibert, Pharm.D., M.B.A., is Manager and Clinical Pharmacy Specialist, Centers for Medication Management; and Ray R. Maddox, Pharm.D., FASHP, is Director, Clinical Pharmacy, Research and Pulmonary Medicine, St. Joseph’s/Candler Health System, Savannah, GA. Elizabeth A. Flynn, Ph.D., is Independent Research Consultant, Artesia, NM. Carolyn Williams, B.S.Pharm., is Medication Safety Specialist, Clinical Pharmacy, St. Joseph’s/ Candler Health System.
wrong-time errors are excluded, the accuracy rate improved from 92% in phase 1 to 96% in phase 3 (p = 0.000008). The overall accuracy rate did not change significantly from phase 1 to phase 3 at hospital 2; when wrong-time errors were excluded from consideration, the accuracy rate improved from 93% in phase 1 to 96% in phase 3 (p = 0.015). Conclusion. Implementation of BCMAeMAR in two hospitals was associated with significant increases in total medication accuracy rates in most study units and did not introduce new types of error into the medication administration process. Accuracy rates further improved when wrongtime errors were excluded from analysis. The frequency of errors preventable by BCMA-eMAR decreased significantly in both hospitals after implementation of that technology. BCMA-eMAR and direct observation were more effective than voluntary reporting programs at intercepting and recording errors and preventing them from reaching patients. Am J Health-Syst Pharm. 2014; 71:209-18
Medication errors can be identified and quantified through at least four different processes. The most
Address correspondence to Dr. Seibert ([email protected]). The authors have declared no potential conflicts of interest. Copyright © 2014, American Society of Health-System Pharmacists, Inc. All rights reserved. 1079-2082/14/0201-0209$06.00. DOI 10.2146/ajhp130332
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
209
PRACTICE REPORTS Barcode technology
common way of identifying a medication error is through the voluntary reporting of events recognized during the medication-use process by clinicians caring for the patient. However, this method vastly underreports the number, frequency, and outcomes of medication errors. 4 Other commonly used methods involve computerized monitoring and manual chart review, though computerized monitoring may identify fewer errors than chart surveillance.5 A more-thorough method of identifying medication errors is direct observation of medication administration by caregivers in patient care areas. Direct observation is resource intensive and seldom used but is recognized as the gold standard of identifying and documenting medication errors.4 Lastly, data from various medication-safety technologies (smart pumps, BCMA, CPOE) that intervene and prevent errors from occurring present a picture of the “potential” medication error rate in health systems where these technologies are deployed.6-8 Studies have shown that the percentage of medication misadventures attributable to errors in drug administration ranges from 2.4% to 11.1% but may be as high as 34–49%, according to some international evaluations.9 Few of these errors are intercepted before reaching the patient. For this reason, the use of BCMA to help improve the accuracy of medication administration at the point of care has seemed particularly promising. However, the implementation of BCMA has proved challenging, with fewer than half of nonfederal hospitals having adopted this technology.3 Published research substantiating the efficacy of BCMA in decreasing the frequency of medication errors is limited.8,10-15 An Institute of Medicine (IOM) report cited a lack of solid evidence demonstrating the effect of technology on medication errors. 1 The IOM Committee on Identifying and Preventing Medication 210
Errors recognized the potential value of bedside BCMA verification but noted that data from observational studies of medication administration are needed.1 There is also a need for evidence comparing the accuracy of such electronic systems to the recognized gold-standard error-detection method (direct observation).4 This study evaluated the effects of a BCMA-eMAR system on the rate of medication administration errors in both regular and special care units in two community, nonteaching hospitals and analyzed the differences in event rates from voluntary reporting, interventions caught by BCMA technology, and interventions identified through direct observation. Methods Setting. St. Joseph’s/Candler Health System comprises two tertiary care, community hospitals totaling 644 beds, with an annual patient volume of 22,807. The hospital staff includes 455 community-based, private practice physicians, 1,245 nurses, and 53 pharmacists. Although most patients in these hospitals are adults, one hospital is a high-volume provider of obstetric services. Technology used for BCMAeMAR. At the time of BCMA-eMAR implementation, the health system used health information technology that was integrated across both hospitals (Meditech Magic 5.61, Westwood, MA). During the course of the BCMA-eMAR implementation, an upgrade to a higher version of the software (Magic 5.64) was completed, but no significant changes to the BCMA-eMAR component were made. Mobile medication carts with thin client computers and tethered scanners were initially provided for each staff member who administered medications. Within a few months, the tethered scanners were replaced with wireless scanners. Shortly after the system went live, the decision was made to install inroom terminals
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
with wireless scanners in the critical care areas and negative-pressure isolation rooms. The change in hardware resulted in significant improvements in staff acceptance, the scanning rate, and system use. BCMA-eMAR implementation. In 2006, the medication safety team evaluated the state of BCMA technology at both hospitals, completed the American Hospital Association (AHA) and Institute for Safe Medication Practices (ISMP) Readiness Plan for BCMA,16 and identified infrastructure improvements that would be needed before implementation could proceed. Improvements identified as a result of the AHA/ISMP assessment included the need for all medications to have a machinereadable barcode, new patient wristbands that contain machine-readable barcodes, significant changes in the computer database, and equipment and training for pharmacy, nursing, and respiratory staff. Completion of the project was divided over three years, allowing the health system to spread the cost of equipment over several budget cycles. Implementation of BCMA-eMAR began in fall 2007. Before the BCMA-eMAR system went live, the decision was made to conduct a research study on implementation to assess whether medication errors were prevented or new errors were introduced through the implementation of the BCMA-eMAR system. Data collection. A pretest–posttest nonequivalent comparison group design was used to investigate the effect of BCMA used in conjunction with an eMAR on the rate of medication administration errors. Observations of medication administration errors were made before (phase 1) and approximately 6 and 12 months after (phases 2 and 3, respectively) implementation of the BCMA-eMAR system. Postimplementation data were collected via direct observation after the study unit staff were fully trained, the system was operational
PRACTICE REPORTS Barcode technology
for at least 6 months, and study unit nurses achieved an electronic scanning rate of at least 80%. Postimplementation data were not collected at the same time for all study units. Study units were reevaluated approximately 12 months after BCMA-eMAR implementation. The study population comprised randomly observed nurses who administered at least one medication to adult patients on the selected units during the observation times. Nurse subjects were those working in the selected patient care units when observation occurred. In order to maintain nurse confidentiality and due to the complexity of trying to collect sufficient doses, no attempt was made to match the preimplementation and postimplementation observations for individual nurses. Instead, observations were made by patient care unit. To facilitate extrapolation of the results to other institutions, data were collected on patient care units that had a diversified adult patient population to whom a wide variety of medications was administered. Patient care units were selected for study inclusion if their use of a medication distribution system was deemed either typical (i.e., similar to like units in other hospitals, such as internal medicine and critical care) or included special interest units (i.e., emergency department and outpatient chemotherapy infusion center). Patient care units included three matched pairs in the two hospitals— two medical–surgical, two telemetry, and two rehabilitation units—plus a medical–surgical intensive care unit (ICU), an emergency department, and both an inpatient oncology unit and outpatient oncology service at one of the hospitals. Investigators estimated the sample size (number of doses to be observed) needed to achieve 90% confidence that the true medication administration error rate was being measured. Based on similar studies,17 a 10%
medication administration error rate was used to estimate the sample size needed for 90% confidence. The direct observation method of Barker et al.18 was used to collect data. Direct observation is a scientifically validated technique for measuring medication errors and provides an accurate description of how many errors occur and insight into how errors may be prevented. This method of collecting errors is nonjudgmental and nonpunitive. Errors were documented and tabulated in the AU MEDS system (MedAccuracy LLC, Lenexa, KS). This system is a nationally standardized method for monitoring medication error rates based on the direct observation of nurses preparing and administering medications.19 Data were collected by 15 licensed health care professionals (7 pharmacists, 8 nurses) and the research pharmacist, all of whom were employed by the health system. These individuals were certified medication observers trained by the AU MEDS specialist. Observers witnessed nurses in the selected units providing standard patient care; observers did not interfere with any activities and were not involved with patient care in any way. On the other hand, if at any time an error could potentially result in patient harm, the research pharmacist intervened as discreetly as possible to correct the issue. The observer introduced the study to the nurses, recorded drug preparation and administration, reviewed the original medication orders, and compared the orders with what was administered to determine if any errors occurred. Observers’ notes were reconciled with the original physicians’ orders to identify any discrepancies between what was written and what was observed. Discrepancies were classified by the type of error that had occurred: wrong time, wrong route, wrong technique, omission, wrong form, extra dose, and unauthorized drug.
These error types are defined in the appendix. All doses observed during the study period were entered into the AU MEDS software program for analysis. Data collected by all observers were maintained in a secure master database. Observers did not count doses as opportunities for error if a drug was left at the patient’s bedside for selfadministration and the administration was not actually witnessed by the observer, if a dose was associated with an uninterpretable written order, or if the observer did not witness the entire process of administration. Any deviation between the order and what was observed was recorded as an error. After examining all doses witnessed, the observer tallied all omitted doses. The medication administration error rate was calculated by dividing the number of observed errors by the sum of all doses witnessed and all omitted doses. Definition of terms. A medication administration error was any discrepancy between a prescriber’s interpretable medication order and what was administered to a patient. Drug administration more than 60 minutes before or after the scheduled time was considered a wrong-time error. For routine doses, the administration schedule listed in each hospital’s medication administration record was used to determine whether a dose was administered within the acceptable time frame. Target errors were errors that should have been prevented by the BCMA system—wrong dose, wrong form, extra dose, unauthorized drug, and omission. An opportunity for error is a measure used as the basic unit of data in observational error studies.10 In this study, opportunity for error was defined as any dose that was ordered and either administered or omitted. Any administered dose was designated as either correct or incorrect (error or no error), which meant the error rate could not exceed 100%.
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
211
PRACTICE REPORTS Barcode technology
The total opportunities for error was the combined total of the number of doses given plus the number of omissions. Total errors was defined as the total of all errors that were observed, including omission, unauthorized drug, extra dose, and wrong route, form, technique, dose, or time. Accuracy rate was the percentage of doses administered correctly and calculated as follows: (total opportunities for error – total errors)/(total opportunities for error × 100). Observed doses were those for which the observer witnessed both nursing preparation (e.g., drawing up medication from vial) and administration of the medication to the patient.10 Electronic scanning was the process of using an electronic device to read a medication barcode and interpret data using software to validate medication accuracy against defined information such as patient identification, medication identification, and medication strength while documenting the actual administration and defined variables such as nurse name, date, and time. An averted event was defined as a near-miss event, meaning that an error occurred but was intercepted, preventing it from reaching the patient. Data analysis. Observer reports were reviewed by the research pharmacist to ensure that error definitions had been applied accurately. Any questionable error was discussed by the observer and the research pharmacist or with the AU MEDS specialist to determine if the error was actually an error and the issue was resolved. Data from the emergency department and outpatient oncology unit of hospital 1 were not included in the calculation of changes in target errors because of important differences between these and other units in the drug distribution system. Chi-square analysis with Yates correction was used to compare 212
phases 1 and 3 to determine whether the BCMA-eMAR system was associated with accurate medication administration in each patient care unit. The a priori level of significance was 0.05. The level of power achieved was calculated, with a goal of 0.80; small-effect sample size20 was used to conservatively assess the effect of the BCMA-eMAR system. Confidence intervals were calculated for all means. Data not collected by direct observation. When a nurse scanned a medication before administering it, one of the following warnings could have appeared on the computer screen: medication is not on current eMAR, medication is for a different patient, abnormal laboratory test results, or allergy. At this point, the nurse can decide not to administer the medication. When the nurse was warned before medication administration and decided not to administer the medication, this was considered an averted event. When a potential error was intercepted by the system after the nurse scanned the medication or the patient’s wristband, the BCMA-eMAR system recorded the event and tabulated the number of times the nurse decided to administer or not administer a dose of medication. Voluntarily reported data were recorded for five time periods: three before BCMA-eMAR implementation and two after BCMA-eMAR implementation. The voluntary reporting process was performed using a Web-based reporting and analysis software system (Quantros Safety Event Manager, version 5.13, Milpitas, CA). This system allowed the reporter to provide a narrative description of the event and included specific fields for recording the date and time of the event, patient-specific demographics, location, type of error or event, cause, name of medication, severity of error or harm, and strategies to prevent
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
reccurrence. The software provided a number of search and reporting options that permitted the tabulation of recorded data. Results Medication administration accuracy. The total opportunities for error and the accuracy rates by patient care unit during phases 1 and 3 are listed in Table 1. Electronic scanning percentages exceeded 90% during phase 3. The occurrence rates for all error types are shown in Table 2. The overall accuracy rate at hospital 1 changed significantly, from 89% in phase 1 to 90% in phase 3 (p = 0.0015). If wrong-time errors are excluded, the accuracy rate improved from 92% in phase 1 to 96% in phase 3 (p = 0.000008). The overall accuracy rate did not change significantly from phase 1 to phase 3 at hospital 2; when wrong-time errors were excluded from consideration, the accuracy rate improved from 93% in phase 1 to 96% in phase 3 (p = 0.015). Target-error analysis. The results of the target-error analysis in study phases 1 and 3 were compared (Table 2). The number of target errors at both facilities did decrease from phase 1 to phase 3; for hospital 1, this analysis was performed with the emergency department and outpatient oncology units excluded, as these were considered special interest units. Special care units. The medication accuracy rate decreased significantly after BCMA implementation in the ICU at hospital 2 (accuracy rate, 94% and 83% in phases 1 and 3, respectively; p = 0.004). The analysis showed a large increase in technique errors (e.g., failure to use a filter straw to remove medication contents from an ampul, failure to have the patient rinse his or her mouth after using an inhaler)—1 in phase 1 and 13 in phase 3. The accuracy rate for hospital 1’s outpatient oncology unit remained
PRACTICE REPORTS Barcode technology
Table 1.
Total Opportunities for Error and Accuracy Rate Before (Phase 1) and After (Phase 3) Implementation of BCMA with Electronic MARa
Unit and Study Phase
Total Opportunities for Error
Accuracy Rate (%)
Accuracy Rate Excluding Wrong-Time Errors (%)
Hospital 1 Medical–surgical Phase 1 534 88 93 Phase 3 101 93 96 Telemetry Phase 1 310 94 94 Phase 3 161 88b 94 Rehabilitation Phase 1 531 86 90 Phase 3 744 85 94c Emergency Phase 1 205 86 87 Phase 3 237 95d 99e Inpatient oncology Phase 1 310 89 94 Phase 3 87 94 98 Outpatient oncology Phase 1 202 97 97 Phase 3 247 97 98g Total Phase 1 2092 89 92 Phase 3 1577 90d 96f Hospital 2 Medical–surgical Phase 1 569 89 94 Phase 3 129 92 93 Telemetry Phase 1 465 89 93 Phase 3 254 89 98g Rehabilitation Phase 1 692 87 92 Phase 3 325 94h 97i Intensive care Phase 1 335 94 96 Phase 3 65 83c 88j Total Phase 1 2061 89 93 Phase 3 773 91 96k a Chi-square analysis with Yates correction was conducted for comparisons between phases 1 and 3 in the same study units. BCMA = barcode-assisted medication administration, MAR = medication administration record. b p = 0.04. c p = 0.004. d p = 0.0015. e p = 0.000002. f p = 0.000008. g p = 0.006. h p = 0.0005. i p = 0.002. j p = 0.003. k p = 0.015.
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
213
214
0 1 45 13 54 6 7 2
86 36
0 9 13 11 69 19 6 2
66 83
0 9 36 13 67 94 71 22 6 3
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
a Excludes the emergency department and outpatient oncology unit because of differences in their drug distribution systems. The exclusion was for target errors only. With the exclusions, there were 115 total target errors and 1570 no-target error doses during phase 1 and 37 total target errors and 1056 no-error doses during phase 3 (p = 0.0001, chi-square analysis with Yates correction). b There were 92 total target errors and 1969 no-error doses during phase 1 and 17 total target errors and 756 no-error doses during phase 3 (p = 0.002, chi-square analysis with Yates correction).
Wrong Form Wrong Dose Study Phase
Hospital 1—all units Phase 1 40 4 4 Phase 3 16 0 1 Hospital 1—units excludeda Phase 1 32 4 4 Phase 3 15 0 1 Hospital 2—all unitsb Phase 1 20 4 7 Phase 3 9 0 0
Unauthorized Drug
Omission
Wrong Time
Other Errors Wrong Technique Target Errors Extra Dose
Types of Errors Observed in Study Hospitals Before (Phase 1) and After (Phase 3) Implementation of Barcode-Assisted Medication Administration with Electronic Medication Administration Record
Table 2.
Wrong Route
PRACTICE REPORTS Barcode technology
fairly consistent, even when wrongtime errors were excluded. Because most chemotherapy and support medications administered are for a single dose, wrong-time errors are not as relevant in this unit as in other practice settings. The accuracy rate increased from 97% in phase 1 to 98% in phase 3. In this unit, there were a total of 4 errors in phase 3, including 2 technique errors involving the administration rate of support medications (not chemotherapy), 1 unauthorized-drug error involving a support medication that was not indicated on the physician’s order, and 1 wrong dose involving a support medication. In the emergency department at hospital 1, the accuracy rate increased from 86% to 95% (p = 0.0015) from phase 1 to phase 3. When wrongtime errors were excluded, the accuracy rate improved—from 87% to 99% (p = 0.000002). The number of wrong-dose errors decreased from 8 to 0 in phase 3. Errors averted by BCMA-eMAR and voluntarily reported errors. As demonstrated in the observational data, some of the medication errors identified were not detected or prevented by BCMA-eMAR. Electronic data collected from the BCMA-eMAR system and voluntarily reported data were compared with the data obtained via direct observation. Table 3 shows the number of doses of medication administered in the study hospitals, the number of errors reported in the voluntary reporting system, the number of observed errors during the study for all observation periods, and the number of events averted by BCMA-eMAR on the observational units. Medication errors consistently declined after BCMA-eMAR was implemented, with the number of doses administered showing little change (Figure 1). The number of averted events far exceeded both voluntarily reported and directly observed medication er-
PRACTICE REPORTS Barcode technology
Table 3.
Errors Reported Voluntarily and Detected by Direct Observation Before and After Implementation of Barcode-Assisted Medication Administration (BCMA) No. Voluntarily Reported Errorsa
Total No. Doses Givena
Date Range
No. Errors Observed Directly in Study Units
No. BCMA Units Studied
No. Errors Averted by BCMA in Study Units
Before BCMA Dec 2006–Mar 2007 663,785 101 4 451 Dec 2007–Mar 2008 697,830 81 2 87 Jan–May 2008 907,441 73 4 152 After BCMA Feb–May 2009 724,068 26 8 182 Sep 2010–Dec 2010 671,834 33 4 58
NA NA NA 1,121 1,234
For all units (not just study units) of both study hospitals combined. NA = not applicable.
a
Figure 1. Number of medication errors occurring in hospital 1 (A) and hospital 2 (B) after implementation of barcode-assisted medication administration with electronic medication administration record. Bars represent doses administered. 35
31
29
30
200000
24
22 150000
18 15
100000
16 16
16
12
15
20
16 16 12
14
11 11 12
13 10
50000
25
9
8
7
7
9
11 11 8
10
10
8
5 0 Dec
Oct
Nov
Sep
Jul
Aug
Jun
May
Apr
Feb
Mar
11 Jan
Dec
Oct
Nov
Sep
Apr
May
Feb Mar 09
Apr
May
Mar
Jan
Feb
Mar
Dec
Dec
Feb
40
37
35
33 200000
30 25 19 16 12
12 9 8
5
7
6
0
Dec
Nov
May
3 Apr
Nov
Oct
Sep
Apr
6
4
May
Mar 09
Apr
May
Mar
Feb
Jan 08
Mar
Dec
Feb
Dec
0
Feb
3
Sep
4
15
12
10
Aug
9
9
Jul
8
50000
14 11
Jun
11
20
17
15 14
Mar
14 15
Feb
100000 20
11 Jan
20
Oct
24
Dec
150000
Jan 07
No. Doses Administered
250000
Jan 07
0
B
15
No. Errors
No. Doses Administered
250000
No. Errors
A
No. errors Trend line
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
215
PRACTICE REPORTS Barcode technology
rors. The majority of averted events were related to abnormal laboratory results, in which case the nurse chose to withhold the dose based on the laboratory test results. There was a general decline in voluntarily reported errors over the entire study period even though the number of administered doses remained fairly constant. Discussion The results of this study demonstrated that BCMA-eMAR was associated with significant reductions in target errors. In addition, many adverse events associated with giving a medication when a patient had a contraindicating laboratory test result may have been averted by using BCMA-eMAR. Improvements in medication accuracy rates were seen in adult inpatient units. These changes were greater when wrong-time errors were eliminated from the comparisons. All of the patient care units studied were high-volume medication administration environments. Results from the medical–surgical and rehabilitation units from both hospitals were similar. However, the accuracy rate, when wrong-time errors were excluded, in the telemetry unit of hospital 2 improved significantly, but there was no change in the telemetry unit of hospital 1. There was a difference in the number of beds in these units (38 beds versus 14 beds, respectively). In addition, patients in the telemetry unit of hospital 1 were predominantly medical–cardiology patients, while patients in the telemetry unit of hospital 2 were a mixture of postprocedure interventional patients, including patients who had undergone angioplasty or coronary artery bypass graft surgery as well as medical–cardiology patients. Finally, patients in the telemetry unit in hospital 1 had a longer mean length of stay than did those in hospital 2. In addition to observations in adult inpatient units, this study eval216
uated BCMA-eMAR in three special care environments. The number of medication errors in the emergency department, most of which were technique errors, significantly decreased from phase 1 to phase 3 (p = 0.0015; when excluding wrong-time errors, p = 0.000002). However, there were no reductions in accuracy rates in the outpatient oncology infusion clinic (hospital 1) and the ICU (hospital 2). The outpatient oncology infusion clinic incorporates multiple double checks with nurses and pharmacists, which likely explains the high accuracy rates before and after BCMA-eMAR implementation. The increase in observed errors in the ICU was the result of multiple medication technique misadventures. Technique errors are not typically identified and prevented using BCMA. In this case, education of the nursing staff is key to ensuring that staff members understand the relationships between medication dosage forms, clinical effects, and routes of administration. An institution may implement pop-up reminders on the eMAR to remind nurses of proper medicationhandling techniques. The use of BCMA-eMAR was accompanied by a reduction in the number of voluntarily reported medication errors. BCMA-eMAR and direct observation are more effective than voluntary reporting programs for intercepting and recording errors and preventing them from reaching patients. Voluntary reporting is typically rich in detail and often allows a more thorough analysis of events, possibly resulting in substantive improvement opportunities. Direct observation of medication administration successfully identifies medication errors but requires more resources to perform. BCMA-eMAR systems electronically capture events and provide reports that can be used to develop improvement plans. When used together, these systems provide rich information for process improvements.
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
Some medication errors may not be prevented by the use of BCMAeMAR, such as prescribing the wrong medication for a condition, failing to appropriately tailor the dosage for the patient, prescribing the wrong route of administration, failing to implement an order, and failing to change an i.v. fluid or initiate new medications when an order is changed. If a medication order is incorrectly entered on a patient’s profile by a pharmacist, a nurse reviewing the order must detect the error or the BCMA-eMAR system will allow the wrong medication to be administered. BCMA-eMAR does not prevent a medication that should be administered intravenously from being administered orally. I.V. pump programming errors will not be prevented by BCMA-eMAR unless barcode technology is used to manage i.v. pump manipulations. The observational data collected revealed a large number of wrongtime medication administration errors despite the liberal two-hour administration window. While these errors may have less importance in some cases, they may be critical in certain patients, such as those who have diabetes mellitus and may have had their meal but whose insulin was delayed or given too early, resulting in blood sugar abnormalities. BCMA-eMAR made little difference in improving the timeliness of medication administration in the two hospitals studied. In fact, wrong-time errors increased in hospital 1 but decreased in hospital 2 after BCMAeMAR implementation. Importantly, this study found that BCMA-eMAR did not introduce new types of errors into the medication administration process. Although nurses initially believed that a requirement to scan barcodes on medications, patients, and themselves would dramatically slow the medication administration process, this effect was not observed as they became efficient with the use of the system.
PRACTICE REPORTS Barcode technology
The results of this study expand and further support the positive effects and limitations of BCMA on reducing medication errors. In addition, this study included observations from patient care units for which there are no other data in the literature (emergency department, ICU, and oncology) and provided comparative data from different methodologies of gathering medication errors (voluntarily reported versus directly observed). Conclusion Implementation of BCMAeMAR in two hospitals was associated with significant increases in total medication accuracy rates in most study units and did not introduce new types of error into the medication administration process. Accuracy rates further improved when wrong-time errors were excluded from analysis. The frequency of errors preventable by BCMA-eMAR decreased significantly in both hospitals after implementation of that technology. BCMA-eMAR and direct observation were more effective than voluntary reporting programs at intercepting and recording errors and preventing them from reaching patients. References 1. Aspden P, Wolcott JA, Bootman JL, Cronenwett LR, eds. Preventing medication errors: quality chasm series. Washington, DC: National Academies Press; 2007. 2. Bates DW, Cullen DJ, Laird N et al. Incidence of adverse drug events and potential adverse events. Implications for prevention. JAMA. 1995; 274:29-34. 3. Pedersen CA, Schneider PJ, Scheckelhoff DJ. ASHP national survey of pharmacy practice in hospital settings: monitoring and patient education—2012. Am J Health-Syst Pharm. 2013; 70:787-803. 4. Flynn EA, Barker KN, Pepper GA et al. Comparison of methods for detecting medication errors in 36 hospitals and skilled-nursing facilities. Am J Health-Syst Pharm. 2002; 59:436-46. 5. 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; 5:305-14.
6. Maddox R, Danello S, Williams GK, Fields M. Intravenous infusion safety initiative: collaboration, evidence-based best practices and “smart” technology help avert high-risk adverse drug events and improve patient outcomes. www.ncbi.nlm.nih.gov/books/NBK43752 (accessed 2013 Oct 21). 7. Helmons PJ, Wargel LN, Daniels CE. Effect of bar-code-assisted medication administration on medication administration errors and accuracy in multiple patient care areas. Am J Health-Syst Pharm. 2009; 66:1202-10. 8. Ammenwerth E, Schnell-Inderst P, Machan C et al. The effect of electronic prescribing on medication errors and adverse drug events: a systematic review. J Am Med Inform Assoc. 2008; 15:585-600. 9. Medication errors: incidence and cost. In: Aspden P, Wolcott JA, Bootman JL, Cronen Wett LR, eds. Preventing medication errors: quality chasm series. Washington, DC: National Academies Press; 2007:105-42. 10. Patterson ES, Rogers ML, Chapman RJ et al. Compliance with intended use of bar code medication administration in acute and long-term care: an observational study. Hum Factors. 2006; 48:15-22. 11. Poon EG, Cina JL, Churchill W et al. Medication dispensing errors and potential adverse drug events before and after implementing bar code technology in the pharmacy. Ann Int Med. 2006; 145:426-34. 12. Young J, Slebodnik M, Sands L. Bar code technology and medication administration error. J Patient Saf. 2010; 6:115-20. 13. Poon EG, Keohane CA, Yoon CS et al. Effect of bar-code technology on the safety of medication administration. N Engl J Med. 2010; 362:1698-707 14. DeYoung JL, Vanderkooi ME, Barletta JF. Effect of bar-code-assisted medication administration on medication error rates in an adult medical intensive care unit. Am J Health-Syst Pharm. 2009; 66:1110-5. 15. Hassink JJ, Essenberg MD, Roukema JA, van den Bemt PM. Effect of bar-codeassisted medication administration on medication administration errors. Am J Health-Syst Pharm. 2013; 70:572-3. Letter. 16. American Hospital Association, Health Research and Educational Trust, Institute for Safe Medication Practices. Pathways for medication safety, assessing bedside bar-code readiness. www.ismp.org/tools/ pathwaysection3.pdf (accessed 2013 Oct 21). 17. Barker KN, Flynn EA, Pepper GA et al. Medication errors observed in 36 health care facilities. Arch Intern Med. 2002; 162:1897-903. 18. Barker KN, Flynn EA, Pepper GA. Observation method of detecting medication errors. Am J Health-Syst Pharm. 2002; 59:2314-6. 19. MedAccuracy. AU MEDS: how it was developed and how it works. www.med
a c c u r a c y. c o m / AU % 2 0 M E D S . h t m (accessed 2013 Oct 21). 20. Cohen J. Statistical power analysis for the behavioral sciences, 2nd ed. Hillsdale, NJ: Lawrence Erlbaum; 1988:235.
Appendix—Error category definitions Unauthorized drug: Administration of a dose of medication that was not ordered for the patient; may also be referred to as an unauthorized or wrong-drug error. Extra dose: Any dose given in excess of the total number of times ordered by the physician, such as a dose given on the basis of an expired order, after a drug has been discontinued, or after a drug has been put on hold. If a physician ordered a drug to be given every morning and the nurse gave it in the evening, the error is placed in this category. Omission: Failure to give an ordered dose. If the patient refuses the medication, an opportunity for error is not counted if the nurse responsible for administering the dose tried to give it. If no attempt was made to administer the dose, then an omission error is counted. Doses withheld according to policies calling for the withholding of medication doses, such as nothing by mouth before surgery, are not counted as errors or opportunities for errors. The observer will detect omissions by comparing the medications administered at a given time with doses that should have been given at that time based on the prescriber’s written orders. Wrong route: Medication administered to a patient using a different route than ordered (e.g., oral administration of a drug ordered for intramuscular use). Also included in this category are doses given in the wrong site, such as the right eye instead of the left eye. Wrong form: The administration of a dose in a different form than ordered by the physician when the physician specified or implied a specific dose form. Giving a tablet when a suspension was ordered is an example. If one of the following dosage forms is crushed, a wrong-form error is counted: extended-release products, enteric-coated drugs, sublingual medications, and effervescent tablets. Wrong technique: An incorrect or omitted action by the nurse during the preparation or administration of a dose that does not result in another type of error. For example, if the wrong rate of infusion is used and the patient receives the correct dose, a wrong-technique error has occurred. If a heart rate is not measured prior to drug administration, a wrong-technique error has occurred. If the heart rate is measured and the rate is too low for the dose to be given but the nurse still administers the drug, an extra-dose error has occurred Wrong dose: Any dose that contained the wrong number of preformed dosage units (such as tablets) or is, in the judgment of the observer, ±17% the correct oral dosage. In judging dosage, measuring devices and graduations
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
217
PRACTICE REPORTS Barcode technology are those provided for routine use by the institution: graduations on the syringe for injections, on medicine cups for oral liquids, and drops for the dropper provided. Any dose of an injectable product that is ±10% or more of the correct dosage is considered an error of this type. Wrong-dose errors are counted for ointments, topical solutions, and similar medications when the dose is quantitatively specified by the physician (e.g., in inches of ointment). Wrong time: Administration of a dose more than 60 minutes before or after the scheduled administration time, unless there is a valid reason. Valid reasons include situations where the physician has ordered that the patient not consume anything by mouth, the patient is off the floor for a diagnostic test, or in surgery. As-needed doses should be administered only as frequently as ordered—the time of the previous dose’s administration should be determined from the medication administration record. The first dose given according to the standard administration schedule is considered to establish that the schedule and subsequent doses on the same day can then be examined for wrong-time errors.
218
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
PRACTICE REPORTS
Effect of barcode technology with electronic medication administration record on medication accuracy rates Heather H. Seibert, Ray R. Maddox, Elizabeth A. Flynn, and Carolyn K. Williams
M
edication errors and consequent deaths continue to escalate, costing health care systems billions of dollars each year.1 It is estimated that at least 1 medication error occurs per day per hospitalized patient. An estimated 450,000 adverse drug events—medication errors that result in patient harm— occur annually, approximately 25% of which are preventable.2 Various technologies have been introduced to help improve the accuracy of medication administration, including automated dispensing cabinets, computerized prescriber order entry (CPOE), “smart” (computerized) i.v. infusion pumps, barcodeassisted medication administration (BCMA) systems, electronic medication administration records (eMARs), and wireless connectivity and integration with hospital information technology. A recent report on the adoption rates of medication-safety technologies revealed that health care organizations have implemented smart pumps more than other technologies.3 BCMA is the second most
Purpose. The effect of barcode-assisted medication administration (BCMA) with electronic medication administration record (eMAR) technology on the occurrence of medication administration errors was evaluated. Methods. A pretest–posttest nonequivalent comparison group was used to investigate the effect of BCMA-eMAR on the medication administration accuracy rates at two community-based hospitals. Patient care units included three matched pairs in the two hospitals—two medical– surgical, two telemetry, and two rehabilitation units—plus a medical–surgical intensive care unit, an emergency department, and both an inpatient oncology unit and an outpatient oncology service at one of the hospitals. Medication administration accuracy rates were observed and recorded before (phase 1) and approximately 6 and 12 months after (phases 2 and 3, respectively) the implementation of BCMA-eMAR. Results. The overall accuracy rate at hospital 1 increased significantly from phase 1 (89%) to phase 3 (90%) (p = 0.0015); if
commonly implemented technology, followed by CPOE with decision support.
Heather H. Seibert, Pharm.D., M.B.A., is Manager and Clinical Pharmacy Specialist, Centers for Medication Management; and Ray R. Maddox, Pharm.D., FASHP, is Director, Clinical Pharmacy, Research and Pulmonary Medicine, St. Joseph’s/Candler Health System, Savannah, GA. Elizabeth A. Flynn, Ph.D., is Independent Research Consultant, Artesia, NM. Carolyn Williams, B.S.Pharm., is Medication Safety Specialist, Clinical Pharmacy, St. Joseph’s/ Candler Health System.
wrong-time errors are excluded, the accuracy rate improved from 92% in phase 1 to 96% in phase 3 (p = 0.000008). The overall accuracy rate did not change significantly from phase 1 to phase 3 at hospital 2; when wrong-time errors were excluded from consideration, the accuracy rate improved from 93% in phase 1 to 96% in phase 3 (p = 0.015). Conclusion. Implementation of BCMAeMAR in two hospitals was associated with significant increases in total medication accuracy rates in most study units and did not introduce new types of error into the medication administration process. Accuracy rates further improved when wrongtime errors were excluded from analysis. The frequency of errors preventable by BCMA-eMAR decreased significantly in both hospitals after implementation of that technology. BCMA-eMAR and direct observation were more effective than voluntary reporting programs at intercepting and recording errors and preventing them from reaching patients. Am J Health-Syst Pharm. 2014; 71:209-18
Medication errors can be identified and quantified through at least four different processes. The most
Address correspondence to Dr. Seibert ([email protected]). The authors have declared no potential conflicts of interest. Copyright © 2014, American Society of Health-System Pharmacists, Inc. All rights reserved. 1079-2082/14/0201-0209$06.00. DOI 10.2146/ajhp130332
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
209
PRACTICE REPORTS Barcode technology
common way of identifying a medication error is through the voluntary reporting of events recognized during the medication-use process by clinicians caring for the patient. However, this method vastly underreports the number, frequency, and outcomes of medication errors. 4 Other commonly used methods involve computerized monitoring and manual chart review, though computerized monitoring may identify fewer errors than chart surveillance.5 A more-thorough method of identifying medication errors is direct observation of medication administration by caregivers in patient care areas. Direct observation is resource intensive and seldom used but is recognized as the gold standard of identifying and documenting medication errors.4 Lastly, data from various medication-safety technologies (smart pumps, BCMA, CPOE) that intervene and prevent errors from occurring present a picture of the “potential” medication error rate in health systems where these technologies are deployed.6-8 Studies have shown that the percentage of medication misadventures attributable to errors in drug administration ranges from 2.4% to 11.1% but may be as high as 34–49%, according to some international evaluations.9 Few of these errors are intercepted before reaching the patient. For this reason, the use of BCMA to help improve the accuracy of medication administration at the point of care has seemed particularly promising. However, the implementation of BCMA has proved challenging, with fewer than half of nonfederal hospitals having adopted this technology.3 Published research substantiating the efficacy of BCMA in decreasing the frequency of medication errors is limited.8,10-15 An Institute of Medicine (IOM) report cited a lack of solid evidence demonstrating the effect of technology on medication errors. 1 The IOM Committee on Identifying and Preventing Medication 210
Errors recognized the potential value of bedside BCMA verification but noted that data from observational studies of medication administration are needed.1 There is also a need for evidence comparing the accuracy of such electronic systems to the recognized gold-standard error-detection method (direct observation).4 This study evaluated the effects of a BCMA-eMAR system on the rate of medication administration errors in both regular and special care units in two community, nonteaching hospitals and analyzed the differences in event rates from voluntary reporting, interventions caught by BCMA technology, and interventions identified through direct observation. Methods Setting. St. Joseph’s/Candler Health System comprises two tertiary care, community hospitals totaling 644 beds, with an annual patient volume of 22,807. The hospital staff includes 455 community-based, private practice physicians, 1,245 nurses, and 53 pharmacists. Although most patients in these hospitals are adults, one hospital is a high-volume provider of obstetric services. Technology used for BCMAeMAR. At the time of BCMA-eMAR implementation, the health system used health information technology that was integrated across both hospitals (Meditech Magic 5.61, Westwood, MA). During the course of the BCMA-eMAR implementation, an upgrade to a higher version of the software (Magic 5.64) was completed, but no significant changes to the BCMA-eMAR component were made. Mobile medication carts with thin client computers and tethered scanners were initially provided for each staff member who administered medications. Within a few months, the tethered scanners were replaced with wireless scanners. Shortly after the system went live, the decision was made to install inroom terminals
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
with wireless scanners in the critical care areas and negative-pressure isolation rooms. The change in hardware resulted in significant improvements in staff acceptance, the scanning rate, and system use. BCMA-eMAR implementation. In 2006, the medication safety team evaluated the state of BCMA technology at both hospitals, completed the American Hospital Association (AHA) and Institute for Safe Medication Practices (ISMP) Readiness Plan for BCMA,16 and identified infrastructure improvements that would be needed before implementation could proceed. Improvements identified as a result of the AHA/ISMP assessment included the need for all medications to have a machinereadable barcode, new patient wristbands that contain machine-readable barcodes, significant changes in the computer database, and equipment and training for pharmacy, nursing, and respiratory staff. Completion of the project was divided over three years, allowing the health system to spread the cost of equipment over several budget cycles. Implementation of BCMA-eMAR began in fall 2007. Before the BCMA-eMAR system went live, the decision was made to conduct a research study on implementation to assess whether medication errors were prevented or new errors were introduced through the implementation of the BCMA-eMAR system. Data collection. A pretest–posttest nonequivalent comparison group design was used to investigate the effect of BCMA used in conjunction with an eMAR on the rate of medication administration errors. Observations of medication administration errors were made before (phase 1) and approximately 6 and 12 months after (phases 2 and 3, respectively) implementation of the BCMA-eMAR system. Postimplementation data were collected via direct observation after the study unit staff were fully trained, the system was operational
PRACTICE REPORTS Barcode technology
for at least 6 months, and study unit nurses achieved an electronic scanning rate of at least 80%. Postimplementation data were not collected at the same time for all study units. Study units were reevaluated approximately 12 months after BCMA-eMAR implementation. The study population comprised randomly observed nurses who administered at least one medication to adult patients on the selected units during the observation times. Nurse subjects were those working in the selected patient care units when observation occurred. In order to maintain nurse confidentiality and due to the complexity of trying to collect sufficient doses, no attempt was made to match the preimplementation and postimplementation observations for individual nurses. Instead, observations were made by patient care unit. To facilitate extrapolation of the results to other institutions, data were collected on patient care units that had a diversified adult patient population to whom a wide variety of medications was administered. Patient care units were selected for study inclusion if their use of a medication distribution system was deemed either typical (i.e., similar to like units in other hospitals, such as internal medicine and critical care) or included special interest units (i.e., emergency department and outpatient chemotherapy infusion center). Patient care units included three matched pairs in the two hospitals— two medical–surgical, two telemetry, and two rehabilitation units—plus a medical–surgical intensive care unit (ICU), an emergency department, and both an inpatient oncology unit and outpatient oncology service at one of the hospitals. Investigators estimated the sample size (number of doses to be observed) needed to achieve 90% confidence that the true medication administration error rate was being measured. Based on similar studies,17 a 10%
medication administration error rate was used to estimate the sample size needed for 90% confidence. The direct observation method of Barker et al.18 was used to collect data. Direct observation is a scientifically validated technique for measuring medication errors and provides an accurate description of how many errors occur and insight into how errors may be prevented. This method of collecting errors is nonjudgmental and nonpunitive. Errors were documented and tabulated in the AU MEDS system (MedAccuracy LLC, Lenexa, KS). This system is a nationally standardized method for monitoring medication error rates based on the direct observation of nurses preparing and administering medications.19 Data were collected by 15 licensed health care professionals (7 pharmacists, 8 nurses) and the research pharmacist, all of whom were employed by the health system. These individuals were certified medication observers trained by the AU MEDS specialist. Observers witnessed nurses in the selected units providing standard patient care; observers did not interfere with any activities and were not involved with patient care in any way. On the other hand, if at any time an error could potentially result in patient harm, the research pharmacist intervened as discreetly as possible to correct the issue. The observer introduced the study to the nurses, recorded drug preparation and administration, reviewed the original medication orders, and compared the orders with what was administered to determine if any errors occurred. Observers’ notes were reconciled with the original physicians’ orders to identify any discrepancies between what was written and what was observed. Discrepancies were classified by the type of error that had occurred: wrong time, wrong route, wrong technique, omission, wrong form, extra dose, and unauthorized drug.
These error types are defined in the appendix. All doses observed during the study period were entered into the AU MEDS software program for analysis. Data collected by all observers were maintained in a secure master database. Observers did not count doses as opportunities for error if a drug was left at the patient’s bedside for selfadministration and the administration was not actually witnessed by the observer, if a dose was associated with an uninterpretable written order, or if the observer did not witness the entire process of administration. Any deviation between the order and what was observed was recorded as an error. After examining all doses witnessed, the observer tallied all omitted doses. The medication administration error rate was calculated by dividing the number of observed errors by the sum of all doses witnessed and all omitted doses. Definition of terms. A medication administration error was any discrepancy between a prescriber’s interpretable medication order and what was administered to a patient. Drug administration more than 60 minutes before or after the scheduled time was considered a wrong-time error. For routine doses, the administration schedule listed in each hospital’s medication administration record was used to determine whether a dose was administered within the acceptable time frame. Target errors were errors that should have been prevented by the BCMA system—wrong dose, wrong form, extra dose, unauthorized drug, and omission. An opportunity for error is a measure used as the basic unit of data in observational error studies.10 In this study, opportunity for error was defined as any dose that was ordered and either administered or omitted. Any administered dose was designated as either correct or incorrect (error or no error), which meant the error rate could not exceed 100%.
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
211
PRACTICE REPORTS Barcode technology
The total opportunities for error was the combined total of the number of doses given plus the number of omissions. Total errors was defined as the total of all errors that were observed, including omission, unauthorized drug, extra dose, and wrong route, form, technique, dose, or time. Accuracy rate was the percentage of doses administered correctly and calculated as follows: (total opportunities for error – total errors)/(total opportunities for error × 100). Observed doses were those for which the observer witnessed both nursing preparation (e.g., drawing up medication from vial) and administration of the medication to the patient.10 Electronic scanning was the process of using an electronic device to read a medication barcode and interpret data using software to validate medication accuracy against defined information such as patient identification, medication identification, and medication strength while documenting the actual administration and defined variables such as nurse name, date, and time. An averted event was defined as a near-miss event, meaning that an error occurred but was intercepted, preventing it from reaching the patient. Data analysis. Observer reports were reviewed by the research pharmacist to ensure that error definitions had been applied accurately. Any questionable error was discussed by the observer and the research pharmacist or with the AU MEDS specialist to determine if the error was actually an error and the issue was resolved. Data from the emergency department and outpatient oncology unit of hospital 1 were not included in the calculation of changes in target errors because of important differences between these and other units in the drug distribution system. Chi-square analysis with Yates correction was used to compare 212
phases 1 and 3 to determine whether the BCMA-eMAR system was associated with accurate medication administration in each patient care unit. The a priori level of significance was 0.05. The level of power achieved was calculated, with a goal of 0.80; small-effect sample size20 was used to conservatively assess the effect of the BCMA-eMAR system. Confidence intervals were calculated for all means. Data not collected by direct observation. When a nurse scanned a medication before administering it, one of the following warnings could have appeared on the computer screen: medication is not on current eMAR, medication is for a different patient, abnormal laboratory test results, or allergy. At this point, the nurse can decide not to administer the medication. When the nurse was warned before medication administration and decided not to administer the medication, this was considered an averted event. When a potential error was intercepted by the system after the nurse scanned the medication or the patient’s wristband, the BCMA-eMAR system recorded the event and tabulated the number of times the nurse decided to administer or not administer a dose of medication. Voluntarily reported data were recorded for five time periods: three before BCMA-eMAR implementation and two after BCMA-eMAR implementation. The voluntary reporting process was performed using a Web-based reporting and analysis software system (Quantros Safety Event Manager, version 5.13, Milpitas, CA). This system allowed the reporter to provide a narrative description of the event and included specific fields for recording the date and time of the event, patient-specific demographics, location, type of error or event, cause, name of medication, severity of error or harm, and strategies to prevent
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
reccurrence. The software provided a number of search and reporting options that permitted the tabulation of recorded data. Results Medication administration accuracy. The total opportunities for error and the accuracy rates by patient care unit during phases 1 and 3 are listed in Table 1. Electronic scanning percentages exceeded 90% during phase 3. The occurrence rates for all error types are shown in Table 2. The overall accuracy rate at hospital 1 changed significantly, from 89% in phase 1 to 90% in phase 3 (p = 0.0015). If wrong-time errors are excluded, the accuracy rate improved from 92% in phase 1 to 96% in phase 3 (p = 0.000008). The overall accuracy rate did not change significantly from phase 1 to phase 3 at hospital 2; when wrong-time errors were excluded from consideration, the accuracy rate improved from 93% in phase 1 to 96% in phase 3 (p = 0.015). Target-error analysis. The results of the target-error analysis in study phases 1 and 3 were compared (Table 2). The number of target errors at both facilities did decrease from phase 1 to phase 3; for hospital 1, this analysis was performed with the emergency department and outpatient oncology units excluded, as these were considered special interest units. Special care units. The medication accuracy rate decreased significantly after BCMA implementation in the ICU at hospital 2 (accuracy rate, 94% and 83% in phases 1 and 3, respectively; p = 0.004). The analysis showed a large increase in technique errors (e.g., failure to use a filter straw to remove medication contents from an ampul, failure to have the patient rinse his or her mouth after using an inhaler)—1 in phase 1 and 13 in phase 3. The accuracy rate for hospital 1’s outpatient oncology unit remained
PRACTICE REPORTS Barcode technology
Table 1.
Total Opportunities for Error and Accuracy Rate Before (Phase 1) and After (Phase 3) Implementation of BCMA with Electronic MARa
Unit and Study Phase
Total Opportunities for Error
Accuracy Rate (%)
Accuracy Rate Excluding Wrong-Time Errors (%)
Hospital 1 Medical–surgical Phase 1 534 88 93 Phase 3 101 93 96 Telemetry Phase 1 310 94 94 Phase 3 161 88b 94 Rehabilitation Phase 1 531 86 90 Phase 3 744 85 94c Emergency Phase 1 205 86 87 Phase 3 237 95d 99e Inpatient oncology Phase 1 310 89 94 Phase 3 87 94 98 Outpatient oncology Phase 1 202 97 97 Phase 3 247 97 98g Total Phase 1 2092 89 92 Phase 3 1577 90d 96f Hospital 2 Medical–surgical Phase 1 569 89 94 Phase 3 129 92 93 Telemetry Phase 1 465 89 93 Phase 3 254 89 98g Rehabilitation Phase 1 692 87 92 Phase 3 325 94h 97i Intensive care Phase 1 335 94 96 Phase 3 65 83c 88j Total Phase 1 2061 89 93 Phase 3 773 91 96k a Chi-square analysis with Yates correction was conducted for comparisons between phases 1 and 3 in the same study units. BCMA = barcode-assisted medication administration, MAR = medication administration record. b p = 0.04. c p = 0.004. d p = 0.0015. e p = 0.000002. f p = 0.000008. g p = 0.006. h p = 0.0005. i p = 0.002. j p = 0.003. k p = 0.015.
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
213
214
0 1 45 13 54 6 7 2
86 36
0 9 13 11 69 19 6 2
66 83
0 9 36 13 67 94 71 22 6 3
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
a Excludes the emergency department and outpatient oncology unit because of differences in their drug distribution systems. The exclusion was for target errors only. With the exclusions, there were 115 total target errors and 1570 no-target error doses during phase 1 and 37 total target errors and 1056 no-error doses during phase 3 (p = 0.0001, chi-square analysis with Yates correction). b There were 92 total target errors and 1969 no-error doses during phase 1 and 17 total target errors and 756 no-error doses during phase 3 (p = 0.002, chi-square analysis with Yates correction).
Wrong Form Wrong Dose Study Phase
Hospital 1—all units Phase 1 40 4 4 Phase 3 16 0 1 Hospital 1—units excludeda Phase 1 32 4 4 Phase 3 15 0 1 Hospital 2—all unitsb Phase 1 20 4 7 Phase 3 9 0 0
Unauthorized Drug
Omission
Wrong Time
Other Errors Wrong Technique Target Errors Extra Dose
Types of Errors Observed in Study Hospitals Before (Phase 1) and After (Phase 3) Implementation of Barcode-Assisted Medication Administration with Electronic Medication Administration Record
Table 2.
Wrong Route
PRACTICE REPORTS Barcode technology
fairly consistent, even when wrongtime errors were excluded. Because most chemotherapy and support medications administered are for a single dose, wrong-time errors are not as relevant in this unit as in other practice settings. The accuracy rate increased from 97% in phase 1 to 98% in phase 3. In this unit, there were a total of 4 errors in phase 3, including 2 technique errors involving the administration rate of support medications (not chemotherapy), 1 unauthorized-drug error involving a support medication that was not indicated on the physician’s order, and 1 wrong dose involving a support medication. In the emergency department at hospital 1, the accuracy rate increased from 86% to 95% (p = 0.0015) from phase 1 to phase 3. When wrongtime errors were excluded, the accuracy rate improved—from 87% to 99% (p = 0.000002). The number of wrong-dose errors decreased from 8 to 0 in phase 3. Errors averted by BCMA-eMAR and voluntarily reported errors. As demonstrated in the observational data, some of the medication errors identified were not detected or prevented by BCMA-eMAR. Electronic data collected from the BCMA-eMAR system and voluntarily reported data were compared with the data obtained via direct observation. Table 3 shows the number of doses of medication administered in the study hospitals, the number of errors reported in the voluntary reporting system, the number of observed errors during the study for all observation periods, and the number of events averted by BCMA-eMAR on the observational units. Medication errors consistently declined after BCMA-eMAR was implemented, with the number of doses administered showing little change (Figure 1). The number of averted events far exceeded both voluntarily reported and directly observed medication er-
PRACTICE REPORTS Barcode technology
Table 3.
Errors Reported Voluntarily and Detected by Direct Observation Before and After Implementation of Barcode-Assisted Medication Administration (BCMA) No. Voluntarily Reported Errorsa
Total No. Doses Givena
Date Range
No. Errors Observed Directly in Study Units
No. BCMA Units Studied
No. Errors Averted by BCMA in Study Units
Before BCMA Dec 2006–Mar 2007 663,785 101 4 451 Dec 2007–Mar 2008 697,830 81 2 87 Jan–May 2008 907,441 73 4 152 After BCMA Feb–May 2009 724,068 26 8 182 Sep 2010–Dec 2010 671,834 33 4 58
NA NA NA 1,121 1,234
For all units (not just study units) of both study hospitals combined. NA = not applicable.
a
Figure 1. Number of medication errors occurring in hospital 1 (A) and hospital 2 (B) after implementation of barcode-assisted medication administration with electronic medication administration record. Bars represent doses administered. 35
31
29
30
200000
24
22 150000
18 15
100000
16 16
16
12
15
20
16 16 12
14
11 11 12
13 10
50000
25
9
8
7
7
9
11 11 8
10
10
8
5 0 Dec
Oct
Nov
Sep
Jul
Aug
Jun
May
Apr
Feb
Mar
11 Jan
Dec
Oct
Nov
Sep
Apr
May
Feb Mar 09
Apr
May
Mar
Jan
Feb
Mar
Dec
Dec
Feb
40
37
35
33 200000
30 25 19 16 12
12 9 8
5
7
6
0
Dec
Nov
May
3 Apr
Nov
Oct
Sep
Apr
6
4
May
Mar 09
Apr
May
Mar
Feb
Jan 08
Mar
Dec
Feb
Dec
0
Feb
3
Sep
4
15
12
10
Aug
9
9
Jul
8
50000
14 11
Jun
11
20
17
15 14
Mar
14 15
Feb
100000 20
11 Jan
20
Oct
24
Dec
150000
Jan 07
No. Doses Administered
250000
Jan 07
0
B
15
No. Errors
No. Doses Administered
250000
No. Errors
A
No. errors Trend line
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
215
PRACTICE REPORTS Barcode technology
rors. The majority of averted events were related to abnormal laboratory results, in which case the nurse chose to withhold the dose based on the laboratory test results. There was a general decline in voluntarily reported errors over the entire study period even though the number of administered doses remained fairly constant. Discussion The results of this study demonstrated that BCMA-eMAR was associated with significant reductions in target errors. In addition, many adverse events associated with giving a medication when a patient had a contraindicating laboratory test result may have been averted by using BCMA-eMAR. Improvements in medication accuracy rates were seen in adult inpatient units. These changes were greater when wrong-time errors were eliminated from the comparisons. All of the patient care units studied were high-volume medication administration environments. Results from the medical–surgical and rehabilitation units from both hospitals were similar. However, the accuracy rate, when wrong-time errors were excluded, in the telemetry unit of hospital 2 improved significantly, but there was no change in the telemetry unit of hospital 1. There was a difference in the number of beds in these units (38 beds versus 14 beds, respectively). In addition, patients in the telemetry unit of hospital 1 were predominantly medical–cardiology patients, while patients in the telemetry unit of hospital 2 were a mixture of postprocedure interventional patients, including patients who had undergone angioplasty or coronary artery bypass graft surgery as well as medical–cardiology patients. Finally, patients in the telemetry unit in hospital 1 had a longer mean length of stay than did those in hospital 2. In addition to observations in adult inpatient units, this study eval216
uated BCMA-eMAR in three special care environments. The number of medication errors in the emergency department, most of which were technique errors, significantly decreased from phase 1 to phase 3 (p = 0.0015; when excluding wrong-time errors, p = 0.000002). However, there were no reductions in accuracy rates in the outpatient oncology infusion clinic (hospital 1) and the ICU (hospital 2). The outpatient oncology infusion clinic incorporates multiple double checks with nurses and pharmacists, which likely explains the high accuracy rates before and after BCMA-eMAR implementation. The increase in observed errors in the ICU was the result of multiple medication technique misadventures. Technique errors are not typically identified and prevented using BCMA. In this case, education of the nursing staff is key to ensuring that staff members understand the relationships between medication dosage forms, clinical effects, and routes of administration. An institution may implement pop-up reminders on the eMAR to remind nurses of proper medicationhandling techniques. The use of BCMA-eMAR was accompanied by a reduction in the number of voluntarily reported medication errors. BCMA-eMAR and direct observation are more effective than voluntary reporting programs for intercepting and recording errors and preventing them from reaching patients. Voluntary reporting is typically rich in detail and often allows a more thorough analysis of events, possibly resulting in substantive improvement opportunities. Direct observation of medication administration successfully identifies medication errors but requires more resources to perform. BCMA-eMAR systems electronically capture events and provide reports that can be used to develop improvement plans. When used together, these systems provide rich information for process improvements.
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
Some medication errors may not be prevented by the use of BCMAeMAR, such as prescribing the wrong medication for a condition, failing to appropriately tailor the dosage for the patient, prescribing the wrong route of administration, failing to implement an order, and failing to change an i.v. fluid or initiate new medications when an order is changed. If a medication order is incorrectly entered on a patient’s profile by a pharmacist, a nurse reviewing the order must detect the error or the BCMA-eMAR system will allow the wrong medication to be administered. BCMA-eMAR does not prevent a medication that should be administered intravenously from being administered orally. I.V. pump programming errors will not be prevented by BCMA-eMAR unless barcode technology is used to manage i.v. pump manipulations. The observational data collected revealed a large number of wrongtime medication administration errors despite the liberal two-hour administration window. While these errors may have less importance in some cases, they may be critical in certain patients, such as those who have diabetes mellitus and may have had their meal but whose insulin was delayed or given too early, resulting in blood sugar abnormalities. BCMA-eMAR made little difference in improving the timeliness of medication administration in the two hospitals studied. In fact, wrong-time errors increased in hospital 1 but decreased in hospital 2 after BCMAeMAR implementation. Importantly, this study found that BCMA-eMAR did not introduce new types of errors into the medication administration process. Although nurses initially believed that a requirement to scan barcodes on medications, patients, and themselves would dramatically slow the medication administration process, this effect was not observed as they became efficient with the use of the system.
PRACTICE REPORTS Barcode technology
The results of this study expand and further support the positive effects and limitations of BCMA on reducing medication errors. In addition, this study included observations from patient care units for which there are no other data in the literature (emergency department, ICU, and oncology) and provided comparative data from different methodologies of gathering medication errors (voluntarily reported versus directly observed). Conclusion Implementation of BCMAeMAR in two hospitals was associated with significant increases in total medication accuracy rates in most study units and did not introduce new types of error into the medication administration process. Accuracy rates further improved when wrong-time errors were excluded from analysis. The frequency of errors preventable by BCMA-eMAR decreased significantly in both hospitals after implementation of that technology. BCMA-eMAR and direct observation were more effective than voluntary reporting programs at intercepting and recording errors and preventing them from reaching patients. References 1. Aspden P, Wolcott JA, Bootman JL, Cronenwett LR, eds. Preventing medication errors: quality chasm series. Washington, DC: National Academies Press; 2007. 2. Bates DW, Cullen DJ, Laird N et al. Incidence of adverse drug events and potential adverse events. Implications for prevention. JAMA. 1995; 274:29-34. 3. Pedersen CA, Schneider PJ, Scheckelhoff DJ. ASHP national survey of pharmacy practice in hospital settings: monitoring and patient education—2012. Am J Health-Syst Pharm. 2013; 70:787-803. 4. Flynn EA, Barker KN, Pepper GA et al. Comparison of methods for detecting medication errors in 36 hospitals and skilled-nursing facilities. Am J Health-Syst Pharm. 2002; 59:436-46. 5. 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; 5:305-14.
6. Maddox R, Danello S, Williams GK, Fields M. Intravenous infusion safety initiative: collaboration, evidence-based best practices and “smart” technology help avert high-risk adverse drug events and improve patient outcomes. www.ncbi.nlm.nih.gov/books/NBK43752 (accessed 2013 Oct 21). 7. Helmons PJ, Wargel LN, Daniels CE. Effect of bar-code-assisted medication administration on medication administration errors and accuracy in multiple patient care areas. Am J Health-Syst Pharm. 2009; 66:1202-10. 8. Ammenwerth E, Schnell-Inderst P, Machan C et al. The effect of electronic prescribing on medication errors and adverse drug events: a systematic review. J Am Med Inform Assoc. 2008; 15:585-600. 9. Medication errors: incidence and cost. In: Aspden P, Wolcott JA, Bootman JL, Cronen Wett LR, eds. Preventing medication errors: quality chasm series. Washington, DC: National Academies Press; 2007:105-42. 10. Patterson ES, Rogers ML, Chapman RJ et al. Compliance with intended use of bar code medication administration in acute and long-term care: an observational study. Hum Factors. 2006; 48:15-22. 11. Poon EG, Cina JL, Churchill W et al. Medication dispensing errors and potential adverse drug events before and after implementing bar code technology in the pharmacy. Ann Int Med. 2006; 145:426-34. 12. Young J, Slebodnik M, Sands L. Bar code technology and medication administration error. J Patient Saf. 2010; 6:115-20. 13. Poon EG, Keohane CA, Yoon CS et al. Effect of bar-code technology on the safety of medication administration. N Engl J Med. 2010; 362:1698-707 14. DeYoung JL, Vanderkooi ME, Barletta JF. Effect of bar-code-assisted medication administration on medication error rates in an adult medical intensive care unit. Am J Health-Syst Pharm. 2009; 66:1110-5. 15. Hassink JJ, Essenberg MD, Roukema JA, van den Bemt PM. Effect of bar-codeassisted medication administration on medication administration errors. Am J Health-Syst Pharm. 2013; 70:572-3. Letter. 16. American Hospital Association, Health Research and Educational Trust, Institute for Safe Medication Practices. Pathways for medication safety, assessing bedside bar-code readiness. www.ismp.org/tools/ pathwaysection3.pdf (accessed 2013 Oct 21). 17. Barker KN, Flynn EA, Pepper GA et al. Medication errors observed in 36 health care facilities. Arch Intern Med. 2002; 162:1897-903. 18. Barker KN, Flynn EA, Pepper GA. Observation method of detecting medication errors. Am J Health-Syst Pharm. 2002; 59:2314-6. 19. MedAccuracy. AU MEDS: how it was developed and how it works. www.med
a c c u r a c y. c o m / AU % 2 0 M E D S . h t m (accessed 2013 Oct 21). 20. Cohen J. Statistical power analysis for the behavioral sciences, 2nd ed. Hillsdale, NJ: Lawrence Erlbaum; 1988:235.
Appendix—Error category definitions Unauthorized drug: Administration of a dose of medication that was not ordered for the patient; may also be referred to as an unauthorized or wrong-drug error. Extra dose: Any dose given in excess of the total number of times ordered by the physician, such as a dose given on the basis of an expired order, after a drug has been discontinued, or after a drug has been put on hold. If a physician ordered a drug to be given every morning and the nurse gave it in the evening, the error is placed in this category. Omission: Failure to give an ordered dose. If the patient refuses the medication, an opportunity for error is not counted if the nurse responsible for administering the dose tried to give it. If no attempt was made to administer the dose, then an omission error is counted. Doses withheld according to policies calling for the withholding of medication doses, such as nothing by mouth before surgery, are not counted as errors or opportunities for errors. The observer will detect omissions by comparing the medications administered at a given time with doses that should have been given at that time based on the prescriber’s written orders. Wrong route: Medication administered to a patient using a different route than ordered (e.g., oral administration of a drug ordered for intramuscular use). Also included in this category are doses given in the wrong site, such as the right eye instead of the left eye. Wrong form: The administration of a dose in a different form than ordered by the physician when the physician specified or implied a specific dose form. Giving a tablet when a suspension was ordered is an example. If one of the following dosage forms is crushed, a wrong-form error is counted: extended-release products, enteric-coated drugs, sublingual medications, and effervescent tablets. Wrong technique: An incorrect or omitted action by the nurse during the preparation or administration of a dose that does not result in another type of error. For example, if the wrong rate of infusion is used and the patient receives the correct dose, a wrong-technique error has occurred. If a heart rate is not measured prior to drug administration, a wrong-technique error has occurred. If the heart rate is measured and the rate is too low for the dose to be given but the nurse still administers the drug, an extra-dose error has occurred Wrong dose: Any dose that contained the wrong number of preformed dosage units (such as tablets) or is, in the judgment of the observer, ±17% the correct oral dosage. In judging dosage, measuring devices and graduations
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
217
PRACTICE REPORTS Barcode technology are those provided for routine use by the institution: graduations on the syringe for injections, on medicine cups for oral liquids, and drops for the dropper provided. Any dose of an injectable product that is ±10% or more of the correct dosage is considered an error of this type. Wrong-dose errors are counted for ointments, topical solutions, and similar medications when the dose is quantitatively specified by the physician (e.g., in inches of ointment). Wrong time: Administration of a dose more than 60 minutes before or after the scheduled administration time, unless there is a valid reason. Valid reasons include situations where the physician has ordered that the patient not consume anything by mouth, the patient is off the floor for a diagnostic test, or in surgery. As-needed doses should be administered only as frequently as ordered—the time of the previous dose’s administration should be determined from the medication administration record. The first dose given according to the standard administration schedule is considered to establish that the schedule and subsequent doses on the same day can then be examined for wrong-time errors.
218
Am J Health-Syst Pharm—Vol 71 Feb 1, 2014
