Preventable adverse drug events

Transcription

1 note Effect of bar-code-assisted medication administration on medication error rates in an adult medical intensive care unit Jaculin L. DeYoung, Marie E. VanderKooi, and Jeffrey F. Barletta Purpose. The effect of bar-code-assisted medication administration (BCMA) on the rate of medication errors in adult patients in a medical intensive care unit (ICU) was studied. Methods. Medication errors were identified in a community teaching hospital medical ICU using a direct observation technique whereby nurses were observed administering medications. Observations occurred for four consecutive 24-hour periods one month before and four months after the implementation of BCMA. Errors in the following categories were recorded: wrong drug, wrong administration time, wrong route, wrong dose, omission, administration of a drug with no order, and documentation error. Two evaluators reviewed all errors for accuracy. Medication error rates were calculated and compared by determining the number of medication errors identified per number of medications administered (observed) preimplementation and postimplementation of BCMA. Statisti- Preventable adverse drug events (ADEs) or medication errors are a leading cause of medical injuries in hospitalized patients and a primary area of focus for quality-improvement initiatives within health care institutions. 1 Knowledge about ADEs has grown substantially over the past decade, largely due to a 1995 report by Bates et al., 2 who found that 6.1 ADEs occurred per 100 admissions and 11.5 ADEs occurred per 1000 patientdays. Of these ADEs, 28% were judged to be preventable. Interestingly, the incidence of ADEs in the medical intensive care unit (ICU) was highest (19.4 per 1000 patientdays), followed by the general medical units (10.6 per 1000 patient-days), the surgical ICU (10.5 per 1000 patientdays), and the general surgical units (8.9 per 1000 patient-days). A more recent report estimated that ADEs occurred in 3.1% of all inpatient stays, and 8.6% of these ADEs were deemed preventable. 3 Several studies have specifically examined medical errors in the critical care setting There is, however, great variability in the error rate reported, primarily attributable to the error-detection method used (i.e., cal analyses were conducted to determine significance. Results. A total of 1465 medication administrations were observed (775 preimplementation and 690 postimplementation) for 92 patients (45 preimplementation and 47 postimplementation). The medication error rate was reduced by 56% after the implementation of BCMA (19.7% versus 8.7%, p < 0.001). This benefit was related to a reduction associated with errors of wrong administration time. Wrong administration time errors decreased from 18.8% during preimplementation to 7.5% postimplementation (p < 0.001). There were no significant differences in other error types. Conclusion. The implementation of BCMA significantly reduced the number of wrong administration time errors in an adult medical ICU. Index terms: Codes; Drug administration; Errors, medication; Hospitals Am J Health-Syst Pharm. 2009; 66: direct observation versus voluntary reporting or medical chart review). Direct observation is considered more accurate than voluntary report- Jaculin L. DeYoung, Pharm.D., BCNSP, is Clinical Pharmacy Specialist, Medication Safety; Marie E. VanderKooi, M.S.N., B.S.N., RN, is Clinical Nurse Specialist; and Jeffrey F. Ba r l e t ta, Pharm.D., FCCM, is Clinical Pharmacy Specialist, Critical Care, Spectrum Health, Grand Rapids, MI. Address correspondence to Dr. DeYoung at the Department of Quality, MC 1C001, Spectrum Health, 100 Michigan NE, Grand Rapids, MI (jaci.deyoung@spectrum-health.org). Supported by grants from the Department of Pharmacy, Nursing Education and Research Foundation, Grand Rapids, and Spectrum Health Foundation, Grand Rapids. The authors have declared no potential conflicts of interest. Copyright 2009, American Society of Health-System Pharmacists, Inc. All rights reserved /09/ $ DOI /ajhp Am J Health-Syst Pharm Vol 66 Jun 15, 2009

2 ing or medical chart review due to issues with underreporting and poor documentation, respectively. 11 Even with direct observation, reported error rates differ. These differences could be related to how medication errors are defined, the phase of the medication-use process studied, the denominator used to calculate the error rate (e.g., total doses administered versus 1000 patient-days), and the specific ICU population evaluated (e.g., medical ICU, surgical ICU, pediatric ICU). One phase of the medication-use process associated with high rates of medication errors is the administration phase. One study revealed that 44% of medication errors experienced by ICU patients occurred in the administration phase. 5 In another study, approximately 61% of all serious medical errors were related to a medication error during the medication ordering or administration phase of treatment. 10 Another study, conducted in a pediatric ICU, found that 36% of errors occurred in the administration phase. 8 Furthermore, of the seven preventable, clinically important errors, five occurred during the medication administration phase. Bar-code-assisted medication administration (BCMA) is intended to reduce errors that occur during the administration phase of the medication-use process by verifying and documenting medication administration at the point of care (i.e., patient s bedside). A BCMA system consists of hardware (scanner and computer), software, and bar-coded unit dose medication packages. When a nurse scans a patient s wristband using a handheld scanning device, the BCMA software program opens, allowing the nurse to review the electronic medication administration record (EMAR) and determine which medications are due for administration. After selecting and preparing the medications for administration, the nurse scans the bar code on the unit dose medication package. If the scanned medication matches the medication on the profile, the nurse completes the verification process and administers the medication. However, if the drug is not on the patient s medication profile, if the dose is too high or too low, if the dosage form is incorrect for the intended route of administration, or if the administration time is too early or too late, an alert is generated to warn the nurse of a potential medication error. Approximately 24% of hospitals in the United States have implemented BCMA technology, a rate that has been steadily increasing (1.5% in 2002 to 9.4% in 2005) Despite increasing usage of BCMA, few studies have evaluated the effect of BCMA on medication error rates, and none have exclusively examined these rates in the ICU setting. 1 While these studies have found a reduction in medication errors with BCMA, it is unknown if these results can be extrapolated to the ICU setting, given the increased number of medications used, the increased use of i.v. medications, the use of infusions requiring dosage calculation, and the increased level of acuity or stress compared with the non-icu setting. This study was conducted to determine the effect of BCMA on the rates of medication administration errors in adult patients in a medical ICU. Methods The study was conducted in a 38- bed adult medical ICU located in a 744-bed community teaching hospital in Grand Rapids, MI. The medical ICU follows a semiclosed model, whereby patients who require ventilatory support are managed by a pulmonary intensivist service. Patients not requiring ventilatory support are managed by the pulmonary intensivist service or a nonintensivist service (e.g., hospitalist). Multidisciplinary rounds occur each morning and include physicians, physician extenders (i.e., nurse practitioners, physician assistants), a clinical pharmacy specialist, a respiratory therapist, and a dietitian. Medication orders are handwritten or preprinted on order forms and faxed to a decentralized pharmacy. A pharmacist enters the medication orders into the computerized pharmacy system. These orders then appear on the EMAR, which is accessible by all clinicians. Standard administration times are used for medication administration (e.g., medications ordered for daily are administered at 9:00 a.m.). The average nurse:patient ratio is 1:2. Data collection. Observations occurred 24 hours a day for four consecutive days (7:00 a.m. Tuesday through 7:00 a.m. Saturday) one month before and four months after the implementation of BCMA in January Other hospital procedures or initiatives targeting medication administration were not implemented during the time period between BCMA preimplementation and postimplementation data collection periods (December 2006 May 2007). Medication administration errors were categorized as follows: wrong drug, wrong administration time, wrong route, wrong dose, administration of a drug with no order, omission, and documentation error. The definitions of these categories were adapted from a previous study 11 and were designed to be consistent with institutional policy and the alerts generated by the bar-coding program (appendix). The drug class associated with each error was also recorded. Data collectors utilized a direct observation technique similar to that previously described 18 and consisted of a small group of pharmacy residents, pharmacy specialists, and a nurse specialist. Training for data collectors consisted of a lecture, description and demonstration of the direct observation technique, simulated medication administration scenarios with practice data Am J Health-Syst Pharm Vol 66 Jun 15,

3 collection, and a written examination. At least one data collector was available during all hours, and multiple data collectors were available at 9:00 a.m. because the majority of medications are administered at this time. The data collectors randomly approached nurses and asked if they could observe the nurses administer medications. Nurses were informed that the purpose of the study was to determine the effect of bar-code technology on medication safety. The term medication errors was purposefully avoided. Nurses were also informed that the observer could not answer any medication-related questions and were referred to the satellite pharmacy or pharmacy specialist if they asked medicationrelated questions. If the nurse agreed, the data collector would observe the nurse administering medications and compare the medications to be administered with the patient s medication profile. All medication administrations were included except metered-dose inhalers and nebulizer treatments (administered by respiratory therapists who were not using BCMA at the time of the study), fentanyl infusions (due to an automated dispensing cabinet and EMAR charting interface limitation), and total parenteral nutrition (including fat emulsions administered separately). Data collectors only intervened if an error was imminent. Other data collected included the nurse:patient ratio, number of years practicing as a nurse, number of years as an ICU nurse in the medical ICU, and number of line items (e.g., the total number of medications) on the EMAR. This study was approved by the health system s institutional review board. The institutional review board waived the requirement for informed consent due to the noninterventional nature of this study. Data analysis. All medication errors were independently reviewed by two senior evaluators. If a disagreement arose regarding the classification of an observation as a medication error, a third evaluator reviewed the observation and provided a recommendation. Interrater reliability was assessed using the k statistic. Medication error rates were calculated and compared by determining the number of medication errors identified per number of medications administered (observed) for the BCMA preimplementation and postimplementation groups. A secondary analysis, which was planned before data collection, was conducted excluding those errors classified as documentation errors. Pearson s chi-square test was used to compare error rates between groups and other dichotomous data. To compare continuous data, the Student s t test and Mann-Whitney U test were used as appropriate. A p value of <0.05 was considered significant. Sample-size analysis revealed that a total of 1200 observations (600 observations preimplementation and 600 observations postimplementation) would be required to detect a difference in the medication error rate from 10% to 5% with 90% power. Data analysis was performed using SPSS, version 13.0, for Windows (SPSS, Inc., Chicago, IL). Results A total of 1465 medication administrations were observed in 92 patients (775 administrations in 47 patients before BCMA implementation, 690 administrations in 45 patients after BCMA implementation). Characteristics of observed nurses and EMARs are displayed in Table 1. The frequency of medication errors before and after BCMA implementation was 19.7% (153 of 775) and 8.7% (60 of 690), respectively (p < 0.001). This translates to a relative risk reduction of 56% after the implementation of BCMA. This benefit was primarily attributable to a reduction in wrong-time errors. Wrong administration time and omission errors were the most common error types in both the preimplementation and postimplementation groups (Table 2). There was a high level of interrater reliability for classifying observations as medication errors (k = 0.88). When documentation errors were removed from the analysis, the Table 1. Characteristics of Observed Nurses and EMARs Before and After Implementation of BCMA a Characteristic Preimplementation Postimplementation p No. patients per nurse, mean ± S.D. 2.3 ± ± Years practicing as a nurse, median (range) 2.5 (0.1 35) 4 (0.8 35) b Years practicing as ICU nurse, median (range) 1.5 (0.1 23) 2 (0.1 35) c No. line items on EMAR, mean ± S.D. 21 ± ± 7.2 <0.001 No. medications administered at time of observation, median (range) 5 (1 20) 4 (1 15) a EMAR = electronic medication administration record, BCMA = bar-code-assisted medication administration, ICU = intensive care unit. b p = when adjusted for the five months that elapsed between the preimplementation and postimplementation periods. c p = when adjusted for the five months that elapsed between the preimplementation and postimplementation periods Am J Health-Syst Pharm Vol 66 Jun 15, 2009

4 frequency of medication errors in the preimplementation and postimplementation groups was 17.2% and 6.4%, respectively (p < 0.001). Examples of documentation errors included holding a b-blocker for bradycardia, holding insulin for low blood glucose level, holding a stool softener for diarrhea, and not administering an oral medication because of loss of enteral access. In all cases, documentation was not provided stating that the dose omission was clinically appropriate based on the specific patient scenario; thus, these errors appeared as missed doses. Table 2. Frequency of Errors by Type of Medication Administration Error a Type of Error Preimplementation (n = 153) No. (%) Errors Postimplementation (n = 60) Wrong administration time 146 (95.4) 52 (86.7) b Omission 6 (3.9) 7 (11.7) Wrong drug 1 (0.7) 2 (3.3) Wrong dose 0 2 (3.3) Wrong route 0 1 (1.7) Drug without an order 1 (0.7) 1 (1.7) Documentation error 20 (13.1) 16 (26.7) a Some administrations were associated with more than one type of error; therefore, the sum of the errors listed by error type does not equal the total error rate reported. b p < Table 3. Frequency of Medication Administration Errors by Drug Class a Drug Class Total (n = 213) The medication classes most frequently associated with medication errors were gastrointestinal agents and cardiovascular agents (Table 3). Within those classes, wrong administration time errors accounted for a majority of the errors (gastrointestinal agents, preimplementation = 100%, postimplementation = 91%; cardiovascular agents, preimplementation = 100%, postimplementation = 83%). Discussion We found a 56% reduction in medication errors after BCMA implementation in an adult medical No. (%) Errors During All Administrations Preimplementation (n = 153) Postimplementation (n = 60) Gastrointestinal agents 62 (29.1) 39 (25.5) 23 (38.3) Cardiovascular agents 35 (16.4) 29 (19.0) 6 (10.0) Electrolytes or vitamins 22 (10.3) 20 (13.0) 2 (3.3) Anticoagulants 17 (8.0) 14 (9.2) 3 (5.0) Neurologic agents 16 (7.5) 12 (7.8) 4 (6.7) Antibiotics 14 (6.6) 10 (6.5) 4 (6.7) Insulins 9 (4.2) 9 (5.9) 0 Pain relievers or narcotics 3 (1.4) 3 (2.0) 0 a The drug classes most commonly associated with medication errors are listed; miscellaneous medications (e.g., topical powders and creams, flushes) are not included. ICU. This finding is similar to that found by Paoletti and colleagues 15 in a non-icu setting. The authors, who also used direct observation, found a 54% reduction (20.6% versus 10%, p = 0.045) in medication errors after implementing EMAR and BCMA systems in a medical surgical unit. Others have reported a reduction in medication error rates with BCMA based on voluntary reporting, but this method largely underestimates the frequency of medication errors compared with direct observation. 16,17 While the rate of medication errors found in the current study before BMCA implementation is consistent with the rates reported in other published studies using direct observation, there are several differences that must be addressed. Tissot and colleagues 6 evaluated administration errors in a medical ICU and reported an error rate of 6.6%. This rate is somewhat lower than the findings of the current study and could be due to the differences in observation length (6 hours/day versus 24 hours/ day) and their exclusion of nights and weekends. In contrast, Van den Bemt et al. 7 reported a rate of administration errors of 44.6% in a mixed medical surgical ICU. Kopp and colleagues 9 evaluated all phases of the drug delivery process in a mixed medical surgical ICU and reported a medication error rate of 20%. Of those errors, 36% were attributed to the prescribing stage, 4% to the transcription phase, 28% to the dispensing phase, and 32% to the administration phase. Similarly, Buckley and colleagues 8 studied a pediatric ICU and reported a medication error rate of 19.8%. Of those errors, 31% occurred in the prescribing phase, 12% in the transcription phase, 21% in the dispensing phase, and 36% in the administration phase. While the error rate reported in the current study (19.6%) is similar, it represents only the administration phase of the medication-use process. Am J Health-Syst Pharm Vol 66 Jun 15,

5 BCMA was most effective in reducing administration errors categorized as wrong administration time errors. There are two potential reasons for this finding. As part of routine quality assurance, feedback is provided to nurses when a medication discrepancy or ADE is detected. Before BCMA implementation, discrepancies related to medications administered at the wrong time were difficult to capture. BCMA allows the exact time a medication was administered to be documented; thus, any discrepancy would be notably evident. This could have resulted in an increased awareness by the bedside nurse, since these errors would now be easily recognized. Further, the low rate of error types other than wrong administration time errors may have limited the ability to detect a significant difference with BCMA. Wrong administration time was the predominant category of error reported in another study (approximately 40%), but other error types were still common. 7 The fact that we did not find more types of errors could be explained by differences in the use of other technologies designed to prevent medication errors. Our institution utilizes automated dispensing cabinets along with an EMAR, which can reduce errors classified as wrong drug, wrong route, or other transcription-related errors. Although wrong administration time errors are often considered less serious than other types of errors, they should not be overlooked. The drug classes most commonly associated with medication errors were gastrointestinal agents and cardiovascular agents. While there are minimal adverse clinical sequelae associated with the late administration of an acid-suppressive agent, uneven dosing schedules of cardiovascular medications could lead to fluctuations in hemodynamic values (e.g., blood pressure, heart rate). In addition, there were many instances where the data collectors observed the administration of insulin occurring at unscheduled or inappropriate times, resulting in fluctuating blood glucose levels. This issue highlighted the importance of documenting reasons for held medications as a communication tool and in an effort to avoid further errors. Despite the low number of errors other than wrong administration time errors detected in this study, the value of BCMA remains. A study that evaluated the severity of errors found that, though a small percentage of errors resulted in patient harm, a significant risk to patient safety exists due to the magnitude of medication use. 19 These investigators estimated that less than 10% of errors resulted in moderate or severe adverse effects but represented 17,000 potential events. This highlights the significant impact of BCMA on medication safety. Limitations to this study included those associated with direct observation and the possibility of the Hawthorne effect. However, previous studies have demonstrated a negligible effect on the observed party through direct observation. 8,9,11 To assess for this, we compared the medication error rates between the first day and the last day of data collection and did not detect a significant difference. Nonetheless, the Hawthorne effect may have occurred. We do acknowledge that nurses were eager to share their experiences with bar-code technology and asked questions about the functionality during the data collection period postimplementation. Another limitation was the single-center design and applicability to other institutions that have different processes for medication delivery, utilize other forms of technology to prevent medication errors (e.g., computerized prescriber order entry), or do not have a clinical pharmacist available. Finally, the administration of some medications was not observed because they were administered by someone other than a nurse (i.e., respiratory therapist) or technical issues with the bar-code system (i.e., fentanyl). Conclusion The implementation of BCMA significantly reduced the number of wrong administration time errors in an adult medical ICU. References 1. Leape LL, Brennan TA, Laird N et al. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med. 1991; 324: Bates DW, Cullen DJ, Laird N et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. JAMA. 1995; 274: Elixhauser A, Owens P. Adverse drug events in U.S. hospitals, (accessed 2008 Sep 12). 4. Beckmann U, Bohringer C, Carless R et al. Evaluation of two methods for quality improvement in intensive care: facilitated incident monitoring and retrospective medical chart review. Crit Care Med. 2003; 31: Cullen DJ, Sweitzer BJ, Bates DW et al. Preventable adverse drug events in hospitalized patients: a comparative study of intensive care and general care units. Crit Care Med. 1997; 25: Tissot E, Cornette C, Demoly P et al. Medication errors at the administration stage in an intensive care unit. Intensive Care Med. 1999; 25: Van den Bemt PM, Fijn R, van der Voort PH et al. Frequency and determinants of drug administration errors in the intensive care unit. Crit Care Med. 2002; 30: Buckley MS, Erstad BL, Kopp BJ et al. Direct observation approach for detecting medication errors and adverse drug events in a pediatric intensive care unit. Pediatr Crit Care Med. 2007; 8: Kopp BJ, Erstad BL, Allen ME et al. Medication errors and adverse drug events in an intensive care unit: direct observation approach for detection. Crit Care Med. 2006; 34: Rothschild JM, Landrigan CP, Cronin JW et al. The Critical Care Safety Study: the incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med. 2005; 33: 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: Pedersen CA, Gumpper KF. ASHP national survey on informatics: assessment of the adoption and use of pharmacy 1114 Am J Health-Syst Pharm Vol 66 Jun 15, 2009

6 informatics in U.S. hospitals Am J Health-Syst Pharm. 2008; 65: Pedersen CA, Schneider PJ, Scheckelhoff DJ. ASHP national survey of pharmacy practice in hospital settings: dispensing and administration Am J Health- Syst Pharm. 2006; 63: Pedersen CA. Trends in pharmacy staff involvement in the emergency department, medication reconciliation, ambulatory services, benchmarking facility operations and patient safety technologies. Paper presented at ASHP Midyear Clinical Meeting. Anaheim, CA; 2006 Dec. 15. Paoletti RD, Suess TM, Lesko MG et al. Using bar-code technology and medication observation methodology for safer medication administration. Am J Health- Syst Pharm. 2007; 64: Puckett F. Medication-management component of a point-of-care information system. Am J Health-Syst Pharm. 1995; 52: Johnson CL, Carlson RA, Tucker CL et al. Using BCMA software to improve patient safety in Veterans Administration Medical Centers. J Healthc Inf Manag. 2002; 16: Allan EL, Barker KN. Fundamentals of medication error research. Am J Hosp Pharm. 1990; 47: Sakowski J, Newman JM, Dozier K. Severity of medication administration errors detected by bar-code medication administration system. Am J Health-Syst Pharm. 2008; 65: Appendix Definitions of types of medication administration errors Administration of drug with no order: An attempt to administer a medication that was not on a patient s profile and did not have an order for that medication pending. Documentation error: An error of omission that was considered appropriate based on the clinical scenario but without documentation of the reason for not administering. Omission: Not administering a medication before the next dose of the same medication was due without a documented reason. Wrong administration time: Administration of a medication more than 60 minutes before or after the scheduled administration time without a documented reason. Wrong dose: Administration of a dose not consistent with the intended dose as specified on the electronic medication administration record (EMAR). Wrong drug: Administration of a medication that was listed on the patient s active profile but was not scheduled for administration during the observation period. Wrong route: An attempt to administer a drug via a route not consistent with the intended route for that medication as specified on the EMAR. Am J Health-Syst Pharm Vol 66 Jun 15,