Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62
doi: 10.1111/jcpt.12225
SIMMEON-Prep study: SIMulation of Medication Errors in ONcology: prevention of antineoplastic preparation errors L. Sarfati* Pharmacy Resident, F. Ranchon*† PharmD PhD, N. Vantard* PharmD, V. Schwiertz* PharmD, N. Gauthier* PharmD, S. He* PharmD, E. Kiouris* PharmD, C. Gourc-Berthod* PharmD, M. G. Guedat* PharmD, C. Alloux* PharmD, M.-P. Gustin‡§¶ PharmD PhD, B. You†** MD PhD, V. Trillet-Lenoir†** MD PhD, G. Freyer†** MD PhD and C. Rioufol*† PharmD PhD *Clinical Oncology Pharmacy Unit, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Pierre Benite, †Universite Lyon 1, EMR UCBL/HCL 3738, Lyon, ‡Departement de sante publique, Faculte de Pharmacie, Universite Lyon 1, Lyon, §H^opital Nord-Ouest Villefranche-sur-Sa^one, Lyon, ¶Hospices Civils de Lyon, Service de Biostatistique, Lyon, and **Medical Oncology Department, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Lyon, France
Received 25 June 2014, Accepted 24 September 2014
Keywords: antineoplastic drugs, cancer, medication errors, prevention, simulation
SUMMARY
WHAT IS KNOWN AND OBJECTIVE
What is known and objective: Medication errors (ME) in oncology are known to cause serious iatrogenic complications. However, MEs still occur at each step in the anticancer chemotherapy process, particularly when injections are prepared in the hospital pharmacy. This study assessed whether a ME simulation program would help prevent ME-associated iatrogenic complications. Methods: The 5-month prospective study, consisting of three phases, was undertaken in the centralized pharmaceutical unit of a university hospital of Lyon, France. During the first simulation phase, 25 instruction sheets each containing one simulated error were inserted among various instruction sheets issued to blinded technicians. The second phase consisted of activity aimed at raising pharmacy technicians’ awareness of risk of medication errors associated with antineoplastic drugs. The third phase consisted of re-enacting the error simulation process 3 months after the awareness campaign. The rate and severity of undetected medication errors were measured during the two simulation (first and third) phases. The potential seriousness of the ME was assessed using the NCC MERPâ index. Results and discussion: The rate of undetected medication errors decreased from 12 in the first simulation phase (48%) to five in the second simulation phase (20%, P = 0∙04). The number of potential deaths due to administration of a faulty preparation decreased from three to zero. Awareness of iatrogenic risk through error simulation allowed pharmacy technicians to improve their ability to identify errors. What is new and conclusion: This study is the first demonstration of the successful application of a simulation-based learning tool for reducing errors in the preparation of injectable anticancer drugs. Such a program should form part of the continuous quality improvement of risk management strategies for cancer patients.
Medical errors are the third leading cause of death in the Unites States, with an estimated 210 000 preventable adverse events per year arising from such errors.1 Prevention of medication errors (ME) remains a worldwide priority for health systems because of their high frequency, their preventability (1∙5 million preventable adverse drug events associated with a ME each year in the United States)2 and their potential severity. In oncology, ME may have particularly severe consequences because of the high toxicity of the antineoplastic drugs used, and the generally poor health status of cancer patients. The potentially life-threatening error rate varies between 2∙6%3 and 10%,4 depending on the type of error (near miss or proven error) and the reporting system. ME resulting in patient death is well documented.5–7 There is a good evidence for the effectiveness of computerized physician order entry and centralized pharmacy units for the safe handling of antineoplastic drugs.8,9 Nevertheless, pharmaceutical errors still occur and are under-documented. Several potential errors may occur during the preparation of antineoplastic agents.10 The preparation error rate ranges from 0∙23% to 6∙66%,11,12 depending on the study and the error detection method used. Preparation errors are mostly intercepted before reaching the patient, but some are not and have lead to serious harm.13 It is crucial to assess and identify the ways in which human factors contribute to the occurrence of errors.14 Teaching procedural methods is appropriate in the training of health professionals, but this is not sufficient for dealing with exceptional or accidental situations.15 Intervention studies have been shown to be effective in detecting and preventing medication errors during drug handling.16 Simulation-based learning, inspired by flight simulators, emerged in health care with the Sim-Oneâ mannequin for anaesthesia in the 1960s.17 Simulation experiments have been developed in medical education, reducing the incidence of ME in comparison with conventional learning.18 Such studies concerned patient identity errors and errors in the preparation of chemotherapy. All highlighted the importance of human factors.19,20 The SIMMEON-Prep (SIMulation of Medication Errors in ONcology) study focuses on assessing the effectiveness of a simulation-based learning program in preventing MEs during the preparation of injectable antineoplastic agents by pharmacy technicians.
Correspondence: Dr Catherine Rioufol, Hospices civils de Lyon, Groupement Hospitalier Sud, Unite de Pharmacie Clinique Oncologique, 165 chemin du grand Revoyet, 69495 Pierre Benite Cedex, France. Tel.: +33 4 78 86 43 68, fax: +33 4 78 86 43 61, e-mail: [email protected]
© 2014 John Wiley & Sons Ltd
55
L. Sarfati et al.
Simulation of medication errors in oncology
administration route errors (three of which were intrathecal instead of intravenous route), three physicochemical incompatibilities, two inappropriate infusion rates, two drug errors and one expiration date error. Medication errors probability was graded on a dedicated score combining the three following criteria: whether the error or a similar one had occurred in our hospital in the last 6 years (+5), or was described in the literature (+3), or was technically feasible in standard practice (+2); scores thus ranged from 0 to 10 (the higher the score, the more probable the error). A score greater than or equal to five was indicated a highly probable error, a score equal to three a probable error and a score strictly less than three a possible error. Simulated errors were scored by the panel of pharmacists in a brainstorming session. The thesaurus included 15 highly probable, eight probable and two possible errors. Potential ME seriousness was assessed on the NCC MERPâ index22 and on published data. Six of the 25 simulated errors would have been life-threatening or fatal if the corresponding preparations had been administered.
METHODS Setting The SIMMEON-Prep study was conducted in the centralized pharmaceutical unit of a 1000-bed University hospital involved in cancer patient care (Centre Hospitalier Lyon Sud, Hospices Civils de Lyon, France). More than 50 000 injectable antineoplastic preparations are dispensed annually. Pharmacy technicians preparing the drugs rely on instruction sheets that have been approved by the pharmacists. In addition to their initial theoretical and practical training, pharmacy technicians are trained in the detection and prevention of ME during preparation, which is regularly assessed by pharmacists with questionnaires and training sessions. Study phases The SIMMEON-Prep study was based on ME simulation. Errors were intentionally included in instruction sheets by senior pharmacists and then transmitted to 12 blinded pharmacy technicians, unaware that any study was being conducted. The 5-month prospective study took place in three phases, from May to September 2013. During the first blind simulation phase, 25 instruction sheets were randomly included among the usual (non-erroneous) instruction sheets transmitted to pharmacy technicians. When a technician detected an erroneous sheet, the corresponding drug was not delivered. The pharmacists, unlike the technicians, were aware of the transmission of erroneous sheets and therefore intercepted the corresponding preparations well before they could be administered to patients. The second phase consisted in raising pharmacy technicians’ awareness of the ME risks associated with antineoplastic drugs. During this education phase, the intercepted and non-intercepted errors were discussed, detailing the potential harm caused if the erroneous treatments had reached the patients. A questionnaire was filled out by the 12 pharmacy technicians before and after the awareness campaign to evaluate their perception of the risk of preparation errors (Appendix S1). The third phase of the study consisted in re-enacting the error simulation process 3 months after the ME risk awareness campaign. The intentionally erroneous instruction sheets were similar in the two simulation phases, and the composition of the team had not changed. The end point of the study was the number of undetected erroneous instruction sheets. This outcome was measured for the two simulation phases, before and after the pharmacy technicians awareness campaign.
Statistical analysis The rate of undetected simulated error was compared before and after the awareness campaign on McNemar’s exact test, with a binomial distribution. RESULTS AND DISCUSSION First phase Thirteen of the 25 intentionally erroneous instruction sheets were detected by the pharmacy technicians (52%). The remaining 12 (48%) comprised four dosage errors, four administration route errors, one drug error, one physico-chemical incompatibility, one infusion rate error and one erroneous expiration date. The rate of detected and undetected errors is described in Fig. 1. The potential seriousness of the undetected errors is shown in Fig. 2. Four of the 13 simulated dosage errors were not detected: two underdosages and two overdosages. The underdosages comprised one etoposide and one bevacizumab preparation, at, respectively, 1/10th and 1/30th of the usual dose. If these highly probable errors (scored 7) had reached the patients, hospital admission would have been necessary to administer an additional dose, and the probability of survival would have been impaired as full treatment would not have been administered. Two of the 10 simulated overdosages were not detected: one dacarbazine preparation corresponded to an antilymphoma instead of an antimelanoma regimen, thus using twice the needed dose. This highly probable error (scored 7) would have required primary or prolonged hospital admission. In the second case of overdosage, two identical instruction sheets were transmitted simultaneously to the pharmacy technicians, and two identical preparations were made for the same patient. This highly probable error (scored 10) would have caused a double dose (24 mg instead of 12 mg) of vinblastine to be administered, leading to neurotoxicity and hospital admission. Two instruction sheets contained erroneous infusion rates. One was not detected and would have led to 1000 mg fluorouracil diluted in a final volume of 1 L being administered in 20 min instead of 24 h, that is a 70-fold infusion rate. This probable error (scored 3) would have required primary or prolonged hospital admission for haematological and cardiovascular complications.
Description of the simulated errors A thesaurus including the intentionally erroneous instruction sheets was developed during a brainstorming session by a panel of five senior hospital pharmacists who were specialists in oncology. Each instruction sheet contained one error. Each error was designed in the light of oncology ME either described in the literature or having previously occurred in our hospital.21 The simulated errors were comparable to reported ME in typology, occurrence and potential severity. Simulated error typology is described in Table 1. There were 13 dosage errors (10 overdosages and three underdosages), four
© 2014 John Wiley & Sons Ltd
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 56
Overdosage
Underdosage
Type of error
Etoposide 20 mg administered intravenously Oxaliplatin 398 mg administered intravenously Methotrexate 1000 mg administered intrathecally Doxorubicin 307 mg administered intravenously Cisplatin 400 mg administered intravenously Methotrexate 40 mg administered intrathecally Dacarbazine 1530 mg administered intravenously Etoposide 650 mg administered intravenously Vinblastine 12 mg 9 2 administered intravenously Carboplatin 1180 mg administered intravenously Vincristine 2∙3 mg administered intravenously
Cytarabine 15 mg administered intrathecally Bevacizumab 15 mg administered intravenously
Simulated preparation
© 2014 John Wiley & Sons Ltd
57 5
5
Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization Monitoring needed
Carboplatin 800 mg administered intravenously Vincristine 2 mg administered intravenously
2
2
2
2
335
336
337
5
2
0
0
5
2
2
5
334
0
5
Etoposide 200 mg administered intravenously Vinblastine 12 mg administered intravenously
Life-threatening harm Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization
2
332
333
2
331
0
5
2
330
5
Life-threatening harm
2
0
5
Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization Death
2
0
5
2
(c) Technical feasibility (+2)
Temporary harm necessitating hospitalization
0
(b) Described in the literature (+3)
0
(a) Reported in the unit (+5)
Monitoring needed
Potential severity (NCC MERP22)
Doxorubicin 50 mg administered intravenously Cisplatin 160 mg administered intravenously Methotrexate 15 mg administered intrathecally Dacarbazine 695 mg administered intravenously
Cytarabine 40 mg administered intrathecally Bevacizumab 487∙5 mg administered intravenously Etoposide 200 mg administered intravenously Oxaliplatin 133 mg administered intravenously Methotrexate 15 mg administered intrathecally
Theoretical preparation
Table 1. Description of 25 simulated errors: type, severity, occurrence
10
10
10
7
7
5
10
10
5
10
7
7
2
Occurrence score (a+b+c)
(continued)
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Possible
Probability of occurrence
Simulation of medication errors in oncology L. Sarfati et al.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62
© 2014 John Wiley & Sons Ltd
58
Expired preparation
Physico-chemical incompatibility
Drug
Infusion rate
Administration route
Type of error
Bortezomib 0∙75 mg administered intrathecally Vincristine 1 mg administered intrathecally Vinblastine 11 mg/ 20 mL administered through the intramuscular route Vinblastine 11 mg administered intrathecally Fluorouracil 2268 mg administered via an infuser in 2 h Fluorouracil 966 mg/1 L in 20 min Paclitaxel protocol/ docetaxel preparation 133 mg Ifosfamide protocol/cyclophosphamide preparation 1305 mg Oxaliplatin 137 mg in 0∙9% sodium chloride Busulfan 0∙55 mg/ mL Methotrexate without lightproof bag Fludarabine administration after expiry date
Simulated preparation
Table 1 (continued)
Fludarabine administration before expiry date
Busulfan 0∙51 mg/ mL Methotrexate with light-proof bag
Vinblastine 11 mg administered intravenously Fluorouracil 2268 mg administered via an infuser in 48 h Fluorouracil 966 mg/1 L in 24 h Docetaxel protocol/ docetaxel preparation 133 mg Cyclophosphamide protocol/cyclophosphamide preparation 1305 mg Oxaliplatin 137 mg in 5% dextrose
Bortezomib 0∙75 mg administered intravenously Vincristine 1 mg administered intravenously Vinblastine 11 mg/ 20 mL administered intravenously
Theoretical preparation
0 5
5
Monitoring needed
Monitoring needed
0
0
2
2
0
2
340 5
0
0
339 0
Monitoring needed
Monitoring needed
Temporary harm necessitating hospitalization
0
313
0
0
338
339
0
37
0
0
Temporary harm necessitating hospitalization
0
36
0
0
Death
0
35
313
0
Irreversible adverse event
(c) Technical feasibility (+2)
(b) Described in the literature (+3)
0
0
Death
Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization
0
(a) Reported in the unit (+5)
Death
Potential severity (NCC MERP22)
7
7
0
10
3
3
3
3
3
3
3
3
Occurrence score (a+b+c)
Highly probable
Highly probable
Possible
Highly probable
Probable
Probable
Probable
Probable
Probable
Probable
Probable
Probable
Probability of occurrence
Simulation of medication errors in oncology L. Sarfati et al.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62
L. Sarfati et al.
Simulation of medication errors in oncology
Fig. 1. Detected and undetected simulated errors according to their typology, 1st simulation phase.
Fig. 2. Severity of the undetected errors during both phases.
One of the three physico-chemical incompatibilities was not detected: an oxaliplatin preparation in 0∙9% sodium chloride instead of 5% dextrose solution. This highly probable error (scored 10) would have led to inefficacy or toxicity, because of oxaliplatin degradation in the presence of chloride. The pharmacy technicians detected none of the four administration route errors, scored as probable errors. In three cases, the simulated error concerned drugs known to be fatal when administered intrathecally: two vinca alkaloid and one bortezomib preparation. The fourth administration route error concerned vinblastine by intramuscular instead of intravenous route in a volume of 20 mL. If this preparation had reached the patient, serious and irreversible adverse events would have occurred because of the unsuitable volume of solution for intramuscular administration and the necrotic effect of vinblastine. Two instruction sheets contained a discrepancy between the name of the protocol and the name of the drug; one was not intercepted, leading to a preparation of docetaxel instead of paclitaxel. This probable error (scored 3) would have led to potential lack of efficacy. The pharmacy technicians also prepared one fludarabine bag that would have expired before being administered. This highly probable error would have led to toxicity or lack of efficacy. Overall, three of the 12 undetected errors could have resulted in the patient’s death; they concerned intrathecal administration of neurotoxic drugs: 0∙75 mg bortezomib, 1 mg vincristine and 11 mg vinblastine.
Awareness and perception phase The pharmacy technicians’ awareness level following the education campaign was assessed using the perception questionnaire presented in Table 2. Third phase In the second simulation phase, 20 of 25 errors (80%) were detected by the pharmacy technicians and five not (20%), vs., respectively, 13 (52%) and 12 (48%) before the awareness campaign. Comparing these rates of undetected simulated errors showed a significant statistical difference in favour of program efficacy (P = 0∙04). The number of detected and undetected errors is shown in Fig. 3. Two of the five errors were overdosages, one an expiration error and two drug errors. One of the two instruction sheets containing drug errors, with discrepancy between the name of the protocol (ifosfamide) and the name of the drug (cyclophosphamide), was detected only in the second simulation phase. The potential seriousness of the undetected errors is shown in Fig. 2. None of the errors undetected in the second simulation phase would have resulted in the patient’s death if the preparation had been administered. Discussions The SIMMEON-Prep program described here is a novel simulation-based learning tool to prevent ME in the preparation of anticancer drugs in hospital. The preparation stage is known to be
© 2014 John Wiley & Sons Ltd
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 59
L. Sarfati et al.
Simulation of medication errors in oncology
Fig. 3. Detected and undetected simulated errors, according to their typology, 2nd simulation phase.
Table 2. Results from the Perception Questionnaire: The 11 questions answered by the pharmacy technicians Questions
Part 1 (pre-awareness)
Choice of answers
The preparation process for anticancer chemotherapy treatments within your unit seem to be Is your training appropriate for the detection of medication errors? Do you think that the pharmacy technician has detected a vinca alkaloid preparation for an intrathecal administration?
Part 2 (per-awareness)
Part 3 (post-awareness)
Which of the following proposals refer to the same administration route? □ Intrathecal □ Intraspinal □ Intramuscular □ Spinal puncture Among the 25 instruction erroneous sheets, how many do you think you have detected? How many of the 4 erroneous sheets that could have led to the patient’s death do you think you have detected? The preparation process for anticancer chemotherapy treatments within your unit seem to be What do you think about the results of this program: range from most important (1) to least important (4) a. They generate stress b. They lead to declining levels of confidence in the team/the process c. They raise awareness of the risks related to the preparation of injectable anticancer agents d. This program should be ongoing Do you think that this medication error simulation program was part of an initiative for improved quality? Do you find this approach to be effective? To you, participating in this programme for the prevention of medication errors is
prone to errors,10 and preparation errors are not systematically detected at the only further stage, that is administration. Results of this study show the impact of raising awareness of iatrogenic complications with the rate of undetected errors decreasing from 12/25 (48%) to 5/25 (20%) (P = 0∙04) in the two simulation phases. These results are all the more important as the improvement was realized under real-life conditions. The
Results
Very secure Reliable At risk Appropriate Moderately appropriate Yes No No answer Correct answers
6/12 5/12 1/12 4/12 8/12 8/12 3/12 1/12 3/12
Correct answers
7/12
Correct answer
0/12
Very secure Reliable At risk Answer a Answer b Answer c Answer d
2/12 4/12 6/12 4 3 1 2
Yes
12/12
Yes Restricting Reassuring Harmful Rewarding
12/12 3/12 12/12 1/12 7/12
simulated error thesaurus, created during a pharmacists’ brainstorming session, was based on the literature, the technical feasibility of the error and also on feedback from the hospital, as 52% of the errors had previously occurred. Between the two simulation phases, undetected errors changed in rate, type and potential harm. Three simulated errors which could have resulted in the patients’ death were only detected in the second phase. Five
© 2014 John Wiley & Sons Ltd
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 60
L. Sarfati et al.
Simulation of medication errors in oncology
errors that were not yet intercepted would not have resulted in the patients’ death. Reasons for missed interception during the second simulation phase were probably multiple: lack of technicians’ knowledge of the drugs used in the chemotherapy regimen, and lack of training or technicians’ selective concentration on potentially fatal errors. This result also highlights the complexity and the variety of anticancer regimens. More than 800 distinct antineoplastic regimens (prescription forms) are registered in our hospital, including those from 250 ongoing clinical trials. Finally, some mistakes introduced in the instruction sheets should probably be more relevant to pharmacists than technicians even if technicians did not make this point during the debriefing phase. Evaluating risk perception is necessary with proven medical errors23 as well as with errors during simulations, as previously described for surgery.24 A systemic approach was adopted in the analysis and presentation of the results, while respecting the anonymity of the participants. The study took the form of a double-blind trial with neither the pharmacy technicians nor the pharmacists knowing who were associated with the errors. The education phase has changed how pharmacy technicians perceive the risk of preparation errors, making possible the development of a culture of safety.25 The present study identified risky behaviours by analysing undetected simulated errors. Error typology highlighted not only pharmacy technicians’ lack of training but also lack of caution in reading the procedures on instruction sheets. Errors were of variable severity, and some might have seemed highly unlikely: a solution containing a neurotoxic drug administered intrathecally would most likely be fatal. The results are all the more alarming because similar fatal errors have actually been reported.5 These highlight the multifactorial origin of error and the role of human factors and behaviour. Despite the implementation of many preventative measures, the risk of human error linked to lack of vigilance remains.20 Moreover, the SIMMEON-Prep study demonstrated that error detection was unrelated to potential clinical seriousness. In addition to the usual causes related to inattention by healthcare workers (psychological factors, distraction, etc.),14 the study identified the lack of awareness on the seriousness of errors as a further risk factor. Although our university hospital provides specialist services in the care of cancer patients, with provision of initial training, on-the job training and experiential training, these are not sufficient. The implication of human factors and behaviour in the occurrence and prevention of error remains inadequately studied.26 Several simulation programs, which have proved successful in civil aviation, in the army and more recently in health care, are effective in preventing medical errors and improving patient safety.27 Simulation is defined by Ziv et al.28 as being the best way to learn from mistakes according to the healthcare professionals questioned. With simulation, Henneman et al. showed that 39% of healthcare professionals failed to intercept errors in patient identification. As with the SIMMEON-Prep design, the healthcare professionals were unaware that the focus of the study was patient identity.19
Further studies are necessary to assess the relation between occurrence of error and pharmacy technicians’ experience and workload. Psychological and management criteria should be included. Caution is required to ensure that healthcare providers do not lose self-confidence, lest they become the second victims of medical error.29 Simulation should be used by healthcare professionals involved in prevention of medication errors, to design further novel programs to mimic real-life conditions. Such studies are also necessary to assess the impact of a simulation program on the other factors, such as nurses and pharmacists, involved in the preparation and administration of antineoplastic drugs. Prospective follow-up of errors will be undertaken to assess the sustainability of the observed improvement in the safe management of antineoplastic drug regimens. Limitations of our study include the heterogeneous population of pharmacy technicians, with different levels of training and experience, although the technicians were the same individuals throughout the study, and the non-representative sample of 25 error-containing instruction sheets. Great care was necessary to prevent the faulty preparations reaching the patient. WHAT IS NEW AND CONCLUSION SIMMEON-Prep is the first simulation-based learning tool that has been shown to be effective in preventing medication errors in the preparation of anticancer drugs in hospitals. Awareness of iatrogenic risk through error simulation allowed pharmacy technicians to improve their ability to detect errors. Such a program should form part of an ongoing process of continuous quality improvement through all the steps in the medication use process and should include all other healthcare workers involved. ACKNOWLEDGEMENTS We would like to acknowledge the pharmaceutical team at the Centre Hospitalier Lyon Sud, Hospices Civils de Lyon, France, and Dr Nathalie Bleyzac. CONFLICT OF INTEREST AND SOURCE OF FUNDING No conflict of interests have been declared. No direct funding was received for the study. The authors were personally salaried by their institutions during the period of writing, although no specific salary was set aside or given for writing the paper. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Appendix S1 Questionnaire on the safety perception of injectable anticancer preparations in the Oncological Clinical Pharmacy Unit.
REFERENCES 1. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf, 2013;9:122–128. 2. Aspden P, Wolcott J, Bootman JL et al. for the Committee on Identifying and Preventing Medication Errors, Institute of Medicine,
editors. Preventing Medication Errors: Quality Chasm Series. Washington, DC: National Academies Press, 2007. 3. Ranchon F, Salles G, Sp€ath H-M et al. Chemotherapeutic errors in hospitalised
© 2014 John Wiley & Sons Ltd
cancer patients: attributable damage and extra costs. BMC Cancer, 2011;11:478. 4. Fyhr A, Akselsson R. Characteristics of medication errors with parenteral cytotoxic drugs. Eur J Cancer Care (Engl), 2012;21:606–613.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 61
L. Sarfati et al.
Simulation of medication errors in oncology
5. Gilbar PJ. Intrathecal chemotherapy: potential for medication error. Cancer Nurs, 2014;37:299–309. 6. Pongudom S, Chinthammitr Y. Inadvertent intrathecal vincristine administration: report of a fatal case despite cerebrospinal fluid lavage and a review of the literature. J Med Assoc Thai, 2011;94:258–263. 7. Barzdo M, Zydek L, Smedra-Kazmirska A, Zgoda M, Machała W, Berent J. Erroneous administration of vinblastine. Pharm World Sci, 2009;31:362–364. 8. Kaushal R, Shojania KG, Bates DW. Effects of computerized physician order entry and clinical decision support systems on medication safety: a systematic review. Arch Intern Med, 2003;163:1409–1416. 9. Nerich V, Limat S, Demarchi M et al. Computerized physician order entry of injectable antineoplastic drugs: an epidemiologic study of prescribing medication errors. Int J Med Inform, 2010;79:699–706. 10. White R, Cassano-Piche A, Fields A et al. Intravenous chemotherapy preparation errors: patient safety risks identified in a pan-Canadian exploratory study. J Oncol Pharm Pract, 2014;20:40–46. 11. Martin F, Legat C, Coutet J et al. Prevention of preparation errors of cytotoxic drugs in centralized units: from epidemiology to quality assurance. Bull Cancer, 2004;91:972–976. 12. Escoms MC, Caba~ nas MJ, Oliveras M, Hidalgo E, Barroso C. Errors evolution and analysis in antineoplastic drug preparation during one year. Pharm World Sci, 1996;18:178–181. 13. McEvilly M, Popelas C, Tremmel B. Use of uridine triacetate for the management of fluorouracil overdose. Am J Health-Syst Pharm, 2011;68:1806–1809. 14. Reason J. Understanding adverse events: human factors. Qual Health Care, 1995;4:80– 89. 15. Hollnagel E, Woods DD, Leveson N. Resilience Engineering: Concepts and Precepts. Hampshire, UK: Ashgate, 2006. 16. Niemann D, Bertsche A, Meyrath D et al. A prospective three-step intervention study to prevent medication errors in drug handling in paediatric care. J Clin Nurs, 2014 [Epub ahead of print]. 17. Denson JS, Abrahamson S. A computercontrolled patient simulator. JAMA, 1969;208:504–508.
18. Ford DG, Seybert AL, Smithburger PL, Kobulinsky LR, Samosky JT, Kane-Gill SL. Impact of simulation-based learning on medication error rates in critically ill patients. Intensive Care Med, 2010;36:1526– 1531. 19. Henneman PL, Fisher DL, Henneman EA, Pham TA, Campbell MM, Nathanson BH. Patient identification errors are common in a simulated setting. Ann Emerg Med, 2010;55:503–509. 20. Jean J, Bleyzac N, Constant H. Ameliorer la detection des erreurs par une equipe pharmaceutique lors de la fabrication centralisee des chimiotherapies. Techniques Hospitalieres, 2006;698:60–65. 21. Ranchon F, Moch C, You B et al. Predictors of prescription errors involving anticancer chemotherapy agents. Eur J Cancer, 2012;48:1192–1199. 22. National Coordinating Council for Medication Error Reporting and Prevention, USA. NCC MERP Taxonomy of Medication Errors. 1998. http://www.nccmerp.org (Accessed 3 June 2014). 23. Amin HJ, Aziz K, Halamek LP et al. Simulation-based learning combined with debriefing: trainers satisfaction with a new approach to training the trainers to teach neonatal resuscitation. BMC Res Notes, 2013;6:251. 24. Arriaga AF, Bader AM, Wong JM et al. Simulation-based trial of surgical-crisis checklists. N Engl J Med, 2013;368:246–253. 25. Occelli P, Quenon J-L, Kret M et al. Validation of the French version of the Hospital Survey on Patient Safety Culture questionnaire. Int J Qual Health Care, 2013;25:459– 468. 26. Irwin A, Mearns K, Watson M, Urquhart J. The effect of proximity, Tall Man lettering, and time pressure on accurate visual perception of drug names. Hum Factors, 2013;55:253–266. 27. Motola I, Devine LA, Chung HS, Sullivan JE, Issenberg SB. Simulation in healthcare education: a best evidence practical guide. AMEE Guide No. 82. Med Teach, 2013;35:1511–1530. 28. Ziv A, Ben-David S, Ziv M. Simulation based medical education: an opportunity to learn from errors. Med Teach, 2005;27:193– 199.
© 2014 John Wiley & Sons Ltd
29. Galam E. L’erreur Medicale, le Burnout et le Soignant. Paris: Springer, 2012. 30. Mu~ noz A, Barcel o R, Viteri A, Rubio I, De Argumedo GL, L opez-Vivanco G. Oxaliplatin overdosage successfully recovered with mild toxicities. Acta Oncol, 2006;45:621–622. 31. Finkelstein Y, Zevin S, Raikhlin-Eisenkraft B, Bentur Y. Intrathecal methotrexate neurotoxicity: clinical correlates and antidotal treatment. Environ Toxicol Pharmacol, 2005;19:721–725. 32. B€ack H, Gustavsson A, Eksborg S, R€ odjer S. Accidental doxorubicin overdosage. Acta Oncol, 1995;34:533–536. 33. Jurek T, Rorat M, Dys P, Swiatek B. Fatal cisplatin overdose in the treatment of mediastinal lymphoma with the ESHAP regimen - analysis of the causes of the adverse drug event. Onkologie, 2013;36:49–52. 34. Bradley AM, Buie LW, Kuykendal A, Voorhees PM. Successful use of intrathecal carboxypeptidase G2 for intrathecal methotrexate overdose: a case study and review of the literature. Clin Lymphoma Myeloma Leuk, 2013;13:166–170. 35. Aversa SML, Zanon S, Marino D, Canova F, Chiarion-Sileni V, Jirillo A. Overdose of Vinblastine in place of Vinorelbine during IGEV chemotherapy. Immunopharmacol Immunotoxicol, 2012;34:879–880. 36. Liem RI, Higman MA, Chen AR, Arceci RJ. Misinterpretation of a Calvert-derived formula leading to carboplatin overdose in two children. J Pediatr Hematol Oncol, 2003;25:818–821. 37. Uner A, Ozet A, Arpaci F, Unsal D. Longterm clinical outcome after accidental overdose of multiple chemotherapeutic agents. Pharmacotherapy, 2005;25:1011–1016. 38. Scalzone M, Coccia P, Cerchiara G et al. Errors involving patients receiving intrathecal chemotherapy. J Chemother, 2010;22:83– 87. 39. Schulmeister L. Look-alike, sound-alike oncology medications. Clin J Oncol Nurs, 2006;10:35–41. 40. Jerremalm E, Hedeland M, Wallin I, Bondesson U, Ehrsson H. Oxaliplatin degradation in the presence of chloride: identification and cytotoxicity of the monochloro monooxalato complex. Pharm Res, 2004;21:891–894.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 62
doi: 10.1111/jcpt.12225
SIMMEON-Prep study: SIMulation of Medication Errors in ONcology: prevention of antineoplastic preparation errors L. Sarfati* Pharmacy Resident, F. Ranchon*† PharmD PhD, N. Vantard* PharmD, V. Schwiertz* PharmD, N. Gauthier* PharmD, S. He* PharmD, E. Kiouris* PharmD, C. Gourc-Berthod* PharmD, M. G. Guedat* PharmD, C. Alloux* PharmD, M.-P. Gustin‡§¶ PharmD PhD, B. You†** MD PhD, V. Trillet-Lenoir†** MD PhD, G. Freyer†** MD PhD and C. Rioufol*† PharmD PhD *Clinical Oncology Pharmacy Unit, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Pierre Benite, †Universite Lyon 1, EMR UCBL/HCL 3738, Lyon, ‡Departement de sante publique, Faculte de Pharmacie, Universite Lyon 1, Lyon, §H^opital Nord-Ouest Villefranche-sur-Sa^one, Lyon, ¶Hospices Civils de Lyon, Service de Biostatistique, Lyon, and **Medical Oncology Department, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Lyon, France
Received 25 June 2014, Accepted 24 September 2014
Keywords: antineoplastic drugs, cancer, medication errors, prevention, simulation
SUMMARY
WHAT IS KNOWN AND OBJECTIVE
What is known and objective: Medication errors (ME) in oncology are known to cause serious iatrogenic complications. However, MEs still occur at each step in the anticancer chemotherapy process, particularly when injections are prepared in the hospital pharmacy. This study assessed whether a ME simulation program would help prevent ME-associated iatrogenic complications. Methods: The 5-month prospective study, consisting of three phases, was undertaken in the centralized pharmaceutical unit of a university hospital of Lyon, France. During the first simulation phase, 25 instruction sheets each containing one simulated error were inserted among various instruction sheets issued to blinded technicians. The second phase consisted of activity aimed at raising pharmacy technicians’ awareness of risk of medication errors associated with antineoplastic drugs. The third phase consisted of re-enacting the error simulation process 3 months after the awareness campaign. The rate and severity of undetected medication errors were measured during the two simulation (first and third) phases. The potential seriousness of the ME was assessed using the NCC MERPâ index. Results and discussion: The rate of undetected medication errors decreased from 12 in the first simulation phase (48%) to five in the second simulation phase (20%, P = 0∙04). The number of potential deaths due to administration of a faulty preparation decreased from three to zero. Awareness of iatrogenic risk through error simulation allowed pharmacy technicians to improve their ability to identify errors. What is new and conclusion: This study is the first demonstration of the successful application of a simulation-based learning tool for reducing errors in the preparation of injectable anticancer drugs. Such a program should form part of the continuous quality improvement of risk management strategies for cancer patients.
Medical errors are the third leading cause of death in the Unites States, with an estimated 210 000 preventable adverse events per year arising from such errors.1 Prevention of medication errors (ME) remains a worldwide priority for health systems because of their high frequency, their preventability (1∙5 million preventable adverse drug events associated with a ME each year in the United States)2 and their potential severity. In oncology, ME may have particularly severe consequences because of the high toxicity of the antineoplastic drugs used, and the generally poor health status of cancer patients. The potentially life-threatening error rate varies between 2∙6%3 and 10%,4 depending on the type of error (near miss or proven error) and the reporting system. ME resulting in patient death is well documented.5–7 There is a good evidence for the effectiveness of computerized physician order entry and centralized pharmacy units for the safe handling of antineoplastic drugs.8,9 Nevertheless, pharmaceutical errors still occur and are under-documented. Several potential errors may occur during the preparation of antineoplastic agents.10 The preparation error rate ranges from 0∙23% to 6∙66%,11,12 depending on the study and the error detection method used. Preparation errors are mostly intercepted before reaching the patient, but some are not and have lead to serious harm.13 It is crucial to assess and identify the ways in which human factors contribute to the occurrence of errors.14 Teaching procedural methods is appropriate in the training of health professionals, but this is not sufficient for dealing with exceptional or accidental situations.15 Intervention studies have been shown to be effective in detecting and preventing medication errors during drug handling.16 Simulation-based learning, inspired by flight simulators, emerged in health care with the Sim-Oneâ mannequin for anaesthesia in the 1960s.17 Simulation experiments have been developed in medical education, reducing the incidence of ME in comparison with conventional learning.18 Such studies concerned patient identity errors and errors in the preparation of chemotherapy. All highlighted the importance of human factors.19,20 The SIMMEON-Prep (SIMulation of Medication Errors in ONcology) study focuses on assessing the effectiveness of a simulation-based learning program in preventing MEs during the preparation of injectable antineoplastic agents by pharmacy technicians.
Correspondence: Dr Catherine Rioufol, Hospices civils de Lyon, Groupement Hospitalier Sud, Unite de Pharmacie Clinique Oncologique, 165 chemin du grand Revoyet, 69495 Pierre Benite Cedex, France. Tel.: +33 4 78 86 43 68, fax: +33 4 78 86 43 61, e-mail: [email protected]
© 2014 John Wiley & Sons Ltd
55
L. Sarfati et al.
Simulation of medication errors in oncology
administration route errors (three of which were intrathecal instead of intravenous route), three physicochemical incompatibilities, two inappropriate infusion rates, two drug errors and one expiration date error. Medication errors probability was graded on a dedicated score combining the three following criteria: whether the error or a similar one had occurred in our hospital in the last 6 years (+5), or was described in the literature (+3), or was technically feasible in standard practice (+2); scores thus ranged from 0 to 10 (the higher the score, the more probable the error). A score greater than or equal to five was indicated a highly probable error, a score equal to three a probable error and a score strictly less than three a possible error. Simulated errors were scored by the panel of pharmacists in a brainstorming session. The thesaurus included 15 highly probable, eight probable and two possible errors. Potential ME seriousness was assessed on the NCC MERPâ index22 and on published data. Six of the 25 simulated errors would have been life-threatening or fatal if the corresponding preparations had been administered.
METHODS Setting The SIMMEON-Prep study was conducted in the centralized pharmaceutical unit of a 1000-bed University hospital involved in cancer patient care (Centre Hospitalier Lyon Sud, Hospices Civils de Lyon, France). More than 50 000 injectable antineoplastic preparations are dispensed annually. Pharmacy technicians preparing the drugs rely on instruction sheets that have been approved by the pharmacists. In addition to their initial theoretical and practical training, pharmacy technicians are trained in the detection and prevention of ME during preparation, which is regularly assessed by pharmacists with questionnaires and training sessions. Study phases The SIMMEON-Prep study was based on ME simulation. Errors were intentionally included in instruction sheets by senior pharmacists and then transmitted to 12 blinded pharmacy technicians, unaware that any study was being conducted. The 5-month prospective study took place in three phases, from May to September 2013. During the first blind simulation phase, 25 instruction sheets were randomly included among the usual (non-erroneous) instruction sheets transmitted to pharmacy technicians. When a technician detected an erroneous sheet, the corresponding drug was not delivered. The pharmacists, unlike the technicians, were aware of the transmission of erroneous sheets and therefore intercepted the corresponding preparations well before they could be administered to patients. The second phase consisted in raising pharmacy technicians’ awareness of the ME risks associated with antineoplastic drugs. During this education phase, the intercepted and non-intercepted errors were discussed, detailing the potential harm caused if the erroneous treatments had reached the patients. A questionnaire was filled out by the 12 pharmacy technicians before and after the awareness campaign to evaluate their perception of the risk of preparation errors (Appendix S1). The third phase of the study consisted in re-enacting the error simulation process 3 months after the ME risk awareness campaign. The intentionally erroneous instruction sheets were similar in the two simulation phases, and the composition of the team had not changed. The end point of the study was the number of undetected erroneous instruction sheets. This outcome was measured for the two simulation phases, before and after the pharmacy technicians awareness campaign.
Statistical analysis The rate of undetected simulated error was compared before and after the awareness campaign on McNemar’s exact test, with a binomial distribution. RESULTS AND DISCUSSION First phase Thirteen of the 25 intentionally erroneous instruction sheets were detected by the pharmacy technicians (52%). The remaining 12 (48%) comprised four dosage errors, four administration route errors, one drug error, one physico-chemical incompatibility, one infusion rate error and one erroneous expiration date. The rate of detected and undetected errors is described in Fig. 1. The potential seriousness of the undetected errors is shown in Fig. 2. Four of the 13 simulated dosage errors were not detected: two underdosages and two overdosages. The underdosages comprised one etoposide and one bevacizumab preparation, at, respectively, 1/10th and 1/30th of the usual dose. If these highly probable errors (scored 7) had reached the patients, hospital admission would have been necessary to administer an additional dose, and the probability of survival would have been impaired as full treatment would not have been administered. Two of the 10 simulated overdosages were not detected: one dacarbazine preparation corresponded to an antilymphoma instead of an antimelanoma regimen, thus using twice the needed dose. This highly probable error (scored 7) would have required primary or prolonged hospital admission. In the second case of overdosage, two identical instruction sheets were transmitted simultaneously to the pharmacy technicians, and two identical preparations were made for the same patient. This highly probable error (scored 10) would have caused a double dose (24 mg instead of 12 mg) of vinblastine to be administered, leading to neurotoxicity and hospital admission. Two instruction sheets contained erroneous infusion rates. One was not detected and would have led to 1000 mg fluorouracil diluted in a final volume of 1 L being administered in 20 min instead of 24 h, that is a 70-fold infusion rate. This probable error (scored 3) would have required primary or prolonged hospital admission for haematological and cardiovascular complications.
Description of the simulated errors A thesaurus including the intentionally erroneous instruction sheets was developed during a brainstorming session by a panel of five senior hospital pharmacists who were specialists in oncology. Each instruction sheet contained one error. Each error was designed in the light of oncology ME either described in the literature or having previously occurred in our hospital.21 The simulated errors were comparable to reported ME in typology, occurrence and potential severity. Simulated error typology is described in Table 1. There were 13 dosage errors (10 overdosages and three underdosages), four
© 2014 John Wiley & Sons Ltd
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 56
Overdosage
Underdosage
Type of error
Etoposide 20 mg administered intravenously Oxaliplatin 398 mg administered intravenously Methotrexate 1000 mg administered intrathecally Doxorubicin 307 mg administered intravenously Cisplatin 400 mg administered intravenously Methotrexate 40 mg administered intrathecally Dacarbazine 1530 mg administered intravenously Etoposide 650 mg administered intravenously Vinblastine 12 mg 9 2 administered intravenously Carboplatin 1180 mg administered intravenously Vincristine 2∙3 mg administered intravenously
Cytarabine 15 mg administered intrathecally Bevacizumab 15 mg administered intravenously
Simulated preparation
© 2014 John Wiley & Sons Ltd
57 5
5
Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization Monitoring needed
Carboplatin 800 mg administered intravenously Vincristine 2 mg administered intravenously
2
2
2
2
335
336
337
5
2
0
0
5
2
2
5
334
0
5
Etoposide 200 mg administered intravenously Vinblastine 12 mg administered intravenously
Life-threatening harm Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization
2
332
333
2
331
0
5
2
330
5
Life-threatening harm
2
0
5
Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization Death
2
0
5
2
(c) Technical feasibility (+2)
Temporary harm necessitating hospitalization
0
(b) Described in the literature (+3)
0
(a) Reported in the unit (+5)
Monitoring needed
Potential severity (NCC MERP22)
Doxorubicin 50 mg administered intravenously Cisplatin 160 mg administered intravenously Methotrexate 15 mg administered intrathecally Dacarbazine 695 mg administered intravenously
Cytarabine 40 mg administered intrathecally Bevacizumab 487∙5 mg administered intravenously Etoposide 200 mg administered intravenously Oxaliplatin 133 mg administered intravenously Methotrexate 15 mg administered intrathecally
Theoretical preparation
Table 1. Description of 25 simulated errors: type, severity, occurrence
10
10
10
7
7
5
10
10
5
10
7
7
2
Occurrence score (a+b+c)
(continued)
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Highly probable
Possible
Probability of occurrence
Simulation of medication errors in oncology L. Sarfati et al.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62
© 2014 John Wiley & Sons Ltd
58
Expired preparation
Physico-chemical incompatibility
Drug
Infusion rate
Administration route
Type of error
Bortezomib 0∙75 mg administered intrathecally Vincristine 1 mg administered intrathecally Vinblastine 11 mg/ 20 mL administered through the intramuscular route Vinblastine 11 mg administered intrathecally Fluorouracil 2268 mg administered via an infuser in 2 h Fluorouracil 966 mg/1 L in 20 min Paclitaxel protocol/ docetaxel preparation 133 mg Ifosfamide protocol/cyclophosphamide preparation 1305 mg Oxaliplatin 137 mg in 0∙9% sodium chloride Busulfan 0∙55 mg/ mL Methotrexate without lightproof bag Fludarabine administration after expiry date
Simulated preparation
Table 1 (continued)
Fludarabine administration before expiry date
Busulfan 0∙51 mg/ mL Methotrexate with light-proof bag
Vinblastine 11 mg administered intravenously Fluorouracil 2268 mg administered via an infuser in 48 h Fluorouracil 966 mg/1 L in 24 h Docetaxel protocol/ docetaxel preparation 133 mg Cyclophosphamide protocol/cyclophosphamide preparation 1305 mg Oxaliplatin 137 mg in 5% dextrose
Bortezomib 0∙75 mg administered intravenously Vincristine 1 mg administered intravenously Vinblastine 11 mg/ 20 mL administered intravenously
Theoretical preparation
0 5
5
Monitoring needed
Monitoring needed
0
0
2
2
0
2
340 5
0
0
339 0
Monitoring needed
Monitoring needed
Temporary harm necessitating hospitalization
0
313
0
0
338
339
0
37
0
0
Temporary harm necessitating hospitalization
0
36
0
0
Death
0
35
313
0
Irreversible adverse event
(c) Technical feasibility (+2)
(b) Described in the literature (+3)
0
0
Death
Temporary harm necessitating hospitalization Temporary harm necessitating hospitalization
0
(a) Reported in the unit (+5)
Death
Potential severity (NCC MERP22)
7
7
0
10
3
3
3
3
3
3
3
3
Occurrence score (a+b+c)
Highly probable
Highly probable
Possible
Highly probable
Probable
Probable
Probable
Probable
Probable
Probable
Probable
Probable
Probability of occurrence
Simulation of medication errors in oncology L. Sarfati et al.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62
L. Sarfati et al.
Simulation of medication errors in oncology
Fig. 1. Detected and undetected simulated errors according to their typology, 1st simulation phase.
Fig. 2. Severity of the undetected errors during both phases.
One of the three physico-chemical incompatibilities was not detected: an oxaliplatin preparation in 0∙9% sodium chloride instead of 5% dextrose solution. This highly probable error (scored 10) would have led to inefficacy or toxicity, because of oxaliplatin degradation in the presence of chloride. The pharmacy technicians detected none of the four administration route errors, scored as probable errors. In three cases, the simulated error concerned drugs known to be fatal when administered intrathecally: two vinca alkaloid and one bortezomib preparation. The fourth administration route error concerned vinblastine by intramuscular instead of intravenous route in a volume of 20 mL. If this preparation had reached the patient, serious and irreversible adverse events would have occurred because of the unsuitable volume of solution for intramuscular administration and the necrotic effect of vinblastine. Two instruction sheets contained a discrepancy between the name of the protocol and the name of the drug; one was not intercepted, leading to a preparation of docetaxel instead of paclitaxel. This probable error (scored 3) would have led to potential lack of efficacy. The pharmacy technicians also prepared one fludarabine bag that would have expired before being administered. This highly probable error would have led to toxicity or lack of efficacy. Overall, three of the 12 undetected errors could have resulted in the patient’s death; they concerned intrathecal administration of neurotoxic drugs: 0∙75 mg bortezomib, 1 mg vincristine and 11 mg vinblastine.
Awareness and perception phase The pharmacy technicians’ awareness level following the education campaign was assessed using the perception questionnaire presented in Table 2. Third phase In the second simulation phase, 20 of 25 errors (80%) were detected by the pharmacy technicians and five not (20%), vs., respectively, 13 (52%) and 12 (48%) before the awareness campaign. Comparing these rates of undetected simulated errors showed a significant statistical difference in favour of program efficacy (P = 0∙04). The number of detected and undetected errors is shown in Fig. 3. Two of the five errors were overdosages, one an expiration error and two drug errors. One of the two instruction sheets containing drug errors, with discrepancy between the name of the protocol (ifosfamide) and the name of the drug (cyclophosphamide), was detected only in the second simulation phase. The potential seriousness of the undetected errors is shown in Fig. 2. None of the errors undetected in the second simulation phase would have resulted in the patient’s death if the preparation had been administered. Discussions The SIMMEON-Prep program described here is a novel simulation-based learning tool to prevent ME in the preparation of anticancer drugs in hospital. The preparation stage is known to be
© 2014 John Wiley & Sons Ltd
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 59
L. Sarfati et al.
Simulation of medication errors in oncology
Fig. 3. Detected and undetected simulated errors, according to their typology, 2nd simulation phase.
Table 2. Results from the Perception Questionnaire: The 11 questions answered by the pharmacy technicians Questions
Part 1 (pre-awareness)
Choice of answers
The preparation process for anticancer chemotherapy treatments within your unit seem to be Is your training appropriate for the detection of medication errors? Do you think that the pharmacy technician has detected a vinca alkaloid preparation for an intrathecal administration?
Part 2 (per-awareness)
Part 3 (post-awareness)
Which of the following proposals refer to the same administration route? □ Intrathecal □ Intraspinal □ Intramuscular □ Spinal puncture Among the 25 instruction erroneous sheets, how many do you think you have detected? How many of the 4 erroneous sheets that could have led to the patient’s death do you think you have detected? The preparation process for anticancer chemotherapy treatments within your unit seem to be What do you think about the results of this program: range from most important (1) to least important (4) a. They generate stress b. They lead to declining levels of confidence in the team/the process c. They raise awareness of the risks related to the preparation of injectable anticancer agents d. This program should be ongoing Do you think that this medication error simulation program was part of an initiative for improved quality? Do you find this approach to be effective? To you, participating in this programme for the prevention of medication errors is
prone to errors,10 and preparation errors are not systematically detected at the only further stage, that is administration. Results of this study show the impact of raising awareness of iatrogenic complications with the rate of undetected errors decreasing from 12/25 (48%) to 5/25 (20%) (P = 0∙04) in the two simulation phases. These results are all the more important as the improvement was realized under real-life conditions. The
Results
Very secure Reliable At risk Appropriate Moderately appropriate Yes No No answer Correct answers
6/12 5/12 1/12 4/12 8/12 8/12 3/12 1/12 3/12
Correct answers
7/12
Correct answer
0/12
Very secure Reliable At risk Answer a Answer b Answer c Answer d
2/12 4/12 6/12 4 3 1 2
Yes
12/12
Yes Restricting Reassuring Harmful Rewarding
12/12 3/12 12/12 1/12 7/12
simulated error thesaurus, created during a pharmacists’ brainstorming session, was based on the literature, the technical feasibility of the error and also on feedback from the hospital, as 52% of the errors had previously occurred. Between the two simulation phases, undetected errors changed in rate, type and potential harm. Three simulated errors which could have resulted in the patients’ death were only detected in the second phase. Five
© 2014 John Wiley & Sons Ltd
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 60
L. Sarfati et al.
Simulation of medication errors in oncology
errors that were not yet intercepted would not have resulted in the patients’ death. Reasons for missed interception during the second simulation phase were probably multiple: lack of technicians’ knowledge of the drugs used in the chemotherapy regimen, and lack of training or technicians’ selective concentration on potentially fatal errors. This result also highlights the complexity and the variety of anticancer regimens. More than 800 distinct antineoplastic regimens (prescription forms) are registered in our hospital, including those from 250 ongoing clinical trials. Finally, some mistakes introduced in the instruction sheets should probably be more relevant to pharmacists than technicians even if technicians did not make this point during the debriefing phase. Evaluating risk perception is necessary with proven medical errors23 as well as with errors during simulations, as previously described for surgery.24 A systemic approach was adopted in the analysis and presentation of the results, while respecting the anonymity of the participants. The study took the form of a double-blind trial with neither the pharmacy technicians nor the pharmacists knowing who were associated with the errors. The education phase has changed how pharmacy technicians perceive the risk of preparation errors, making possible the development of a culture of safety.25 The present study identified risky behaviours by analysing undetected simulated errors. Error typology highlighted not only pharmacy technicians’ lack of training but also lack of caution in reading the procedures on instruction sheets. Errors were of variable severity, and some might have seemed highly unlikely: a solution containing a neurotoxic drug administered intrathecally would most likely be fatal. The results are all the more alarming because similar fatal errors have actually been reported.5 These highlight the multifactorial origin of error and the role of human factors and behaviour. Despite the implementation of many preventative measures, the risk of human error linked to lack of vigilance remains.20 Moreover, the SIMMEON-Prep study demonstrated that error detection was unrelated to potential clinical seriousness. In addition to the usual causes related to inattention by healthcare workers (psychological factors, distraction, etc.),14 the study identified the lack of awareness on the seriousness of errors as a further risk factor. Although our university hospital provides specialist services in the care of cancer patients, with provision of initial training, on-the job training and experiential training, these are not sufficient. The implication of human factors and behaviour in the occurrence and prevention of error remains inadequately studied.26 Several simulation programs, which have proved successful in civil aviation, in the army and more recently in health care, are effective in preventing medical errors and improving patient safety.27 Simulation is defined by Ziv et al.28 as being the best way to learn from mistakes according to the healthcare professionals questioned. With simulation, Henneman et al. showed that 39% of healthcare professionals failed to intercept errors in patient identification. As with the SIMMEON-Prep design, the healthcare professionals were unaware that the focus of the study was patient identity.19
Further studies are necessary to assess the relation between occurrence of error and pharmacy technicians’ experience and workload. Psychological and management criteria should be included. Caution is required to ensure that healthcare providers do not lose self-confidence, lest they become the second victims of medical error.29 Simulation should be used by healthcare professionals involved in prevention of medication errors, to design further novel programs to mimic real-life conditions. Such studies are also necessary to assess the impact of a simulation program on the other factors, such as nurses and pharmacists, involved in the preparation and administration of antineoplastic drugs. Prospective follow-up of errors will be undertaken to assess the sustainability of the observed improvement in the safe management of antineoplastic drug regimens. Limitations of our study include the heterogeneous population of pharmacy technicians, with different levels of training and experience, although the technicians were the same individuals throughout the study, and the non-representative sample of 25 error-containing instruction sheets. Great care was necessary to prevent the faulty preparations reaching the patient. WHAT IS NEW AND CONCLUSION SIMMEON-Prep is the first simulation-based learning tool that has been shown to be effective in preventing medication errors in the preparation of anticancer drugs in hospitals. Awareness of iatrogenic risk through error simulation allowed pharmacy technicians to improve their ability to detect errors. Such a program should form part of an ongoing process of continuous quality improvement through all the steps in the medication use process and should include all other healthcare workers involved. ACKNOWLEDGEMENTS We would like to acknowledge the pharmaceutical team at the Centre Hospitalier Lyon Sud, Hospices Civils de Lyon, France, and Dr Nathalie Bleyzac. CONFLICT OF INTEREST AND SOURCE OF FUNDING No conflict of interests have been declared. No direct funding was received for the study. The authors were personally salaried by their institutions during the period of writing, although no specific salary was set aside or given for writing the paper. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Appendix S1 Questionnaire on the safety perception of injectable anticancer preparations in the Oncological Clinical Pharmacy Unit.
REFERENCES 1. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf, 2013;9:122–128. 2. Aspden P, Wolcott J, Bootman JL et al. for the Committee on Identifying and Preventing Medication Errors, Institute of Medicine,
editors. Preventing Medication Errors: Quality Chasm Series. Washington, DC: National Academies Press, 2007. 3. Ranchon F, Salles G, Sp€ath H-M et al. Chemotherapeutic errors in hospitalised
© 2014 John Wiley & Sons Ltd
cancer patients: attributable damage and extra costs. BMC Cancer, 2011;11:478. 4. Fyhr A, Akselsson R. Characteristics of medication errors with parenteral cytotoxic drugs. Eur J Cancer Care (Engl), 2012;21:606–613.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 61
L. Sarfati et al.
Simulation of medication errors in oncology
5. Gilbar PJ. Intrathecal chemotherapy: potential for medication error. Cancer Nurs, 2014;37:299–309. 6. Pongudom S, Chinthammitr Y. Inadvertent intrathecal vincristine administration: report of a fatal case despite cerebrospinal fluid lavage and a review of the literature. J Med Assoc Thai, 2011;94:258–263. 7. Barzdo M, Zydek L, Smedra-Kazmirska A, Zgoda M, Machała W, Berent J. Erroneous administration of vinblastine. Pharm World Sci, 2009;31:362–364. 8. Kaushal R, Shojania KG, Bates DW. Effects of computerized physician order entry and clinical decision support systems on medication safety: a systematic review. Arch Intern Med, 2003;163:1409–1416. 9. Nerich V, Limat S, Demarchi M et al. Computerized physician order entry of injectable antineoplastic drugs: an epidemiologic study of prescribing medication errors. Int J Med Inform, 2010;79:699–706. 10. White R, Cassano-Piche A, Fields A et al. Intravenous chemotherapy preparation errors: patient safety risks identified in a pan-Canadian exploratory study. J Oncol Pharm Pract, 2014;20:40–46. 11. Martin F, Legat C, Coutet J et al. Prevention of preparation errors of cytotoxic drugs in centralized units: from epidemiology to quality assurance. Bull Cancer, 2004;91:972–976. 12. Escoms MC, Caba~ nas MJ, Oliveras M, Hidalgo E, Barroso C. Errors evolution and analysis in antineoplastic drug preparation during one year. Pharm World Sci, 1996;18:178–181. 13. McEvilly M, Popelas C, Tremmel B. Use of uridine triacetate for the management of fluorouracil overdose. Am J Health-Syst Pharm, 2011;68:1806–1809. 14. Reason J. Understanding adverse events: human factors. Qual Health Care, 1995;4:80– 89. 15. Hollnagel E, Woods DD, Leveson N. Resilience Engineering: Concepts and Precepts. Hampshire, UK: Ashgate, 2006. 16. Niemann D, Bertsche A, Meyrath D et al. A prospective three-step intervention study to prevent medication errors in drug handling in paediatric care. J Clin Nurs, 2014 [Epub ahead of print]. 17. Denson JS, Abrahamson S. A computercontrolled patient simulator. JAMA, 1969;208:504–508.
18. Ford DG, Seybert AL, Smithburger PL, Kobulinsky LR, Samosky JT, Kane-Gill SL. Impact of simulation-based learning on medication error rates in critically ill patients. Intensive Care Med, 2010;36:1526– 1531. 19. Henneman PL, Fisher DL, Henneman EA, Pham TA, Campbell MM, Nathanson BH. Patient identification errors are common in a simulated setting. Ann Emerg Med, 2010;55:503–509. 20. Jean J, Bleyzac N, Constant H. Ameliorer la detection des erreurs par une equipe pharmaceutique lors de la fabrication centralisee des chimiotherapies. Techniques Hospitalieres, 2006;698:60–65. 21. Ranchon F, Moch C, You B et al. Predictors of prescription errors involving anticancer chemotherapy agents. Eur J Cancer, 2012;48:1192–1199. 22. National Coordinating Council for Medication Error Reporting and Prevention, USA. NCC MERP Taxonomy of Medication Errors. 1998. http://www.nccmerp.org (Accessed 3 June 2014). 23. Amin HJ, Aziz K, Halamek LP et al. Simulation-based learning combined with debriefing: trainers satisfaction with a new approach to training the trainers to teach neonatal resuscitation. BMC Res Notes, 2013;6:251. 24. Arriaga AF, Bader AM, Wong JM et al. Simulation-based trial of surgical-crisis checklists. N Engl J Med, 2013;368:246–253. 25. Occelli P, Quenon J-L, Kret M et al. Validation of the French version of the Hospital Survey on Patient Safety Culture questionnaire. Int J Qual Health Care, 2013;25:459– 468. 26. Irwin A, Mearns K, Watson M, Urquhart J. The effect of proximity, Tall Man lettering, and time pressure on accurate visual perception of drug names. Hum Factors, 2013;55:253–266. 27. Motola I, Devine LA, Chung HS, Sullivan JE, Issenberg SB. Simulation in healthcare education: a best evidence practical guide. AMEE Guide No. 82. Med Teach, 2013;35:1511–1530. 28. Ziv A, Ben-David S, Ziv M. Simulation based medical education: an opportunity to learn from errors. Med Teach, 2005;27:193– 199.
© 2014 John Wiley & Sons Ltd
29. Galam E. L’erreur Medicale, le Burnout et le Soignant. Paris: Springer, 2012. 30. Mu~ noz A, Barcel o R, Viteri A, Rubio I, De Argumedo GL, L opez-Vivanco G. Oxaliplatin overdosage successfully recovered with mild toxicities. Acta Oncol, 2006;45:621–622. 31. Finkelstein Y, Zevin S, Raikhlin-Eisenkraft B, Bentur Y. Intrathecal methotrexate neurotoxicity: clinical correlates and antidotal treatment. Environ Toxicol Pharmacol, 2005;19:721–725. 32. B€ack H, Gustavsson A, Eksborg S, R€ odjer S. Accidental doxorubicin overdosage. Acta Oncol, 1995;34:533–536. 33. Jurek T, Rorat M, Dys P, Swiatek B. Fatal cisplatin overdose in the treatment of mediastinal lymphoma with the ESHAP regimen - analysis of the causes of the adverse drug event. Onkologie, 2013;36:49–52. 34. Bradley AM, Buie LW, Kuykendal A, Voorhees PM. Successful use of intrathecal carboxypeptidase G2 for intrathecal methotrexate overdose: a case study and review of the literature. Clin Lymphoma Myeloma Leuk, 2013;13:166–170. 35. Aversa SML, Zanon S, Marino D, Canova F, Chiarion-Sileni V, Jirillo A. Overdose of Vinblastine in place of Vinorelbine during IGEV chemotherapy. Immunopharmacol Immunotoxicol, 2012;34:879–880. 36. Liem RI, Higman MA, Chen AR, Arceci RJ. Misinterpretation of a Calvert-derived formula leading to carboplatin overdose in two children. J Pediatr Hematol Oncol, 2003;25:818–821. 37. Uner A, Ozet A, Arpaci F, Unsal D. Longterm clinical outcome after accidental overdose of multiple chemotherapeutic agents. Pharmacotherapy, 2005;25:1011–1016. 38. Scalzone M, Coccia P, Cerchiara G et al. Errors involving patients receiving intrathecal chemotherapy. J Chemother, 2010;22:83– 87. 39. Schulmeister L. Look-alike, sound-alike oncology medications. Clin J Oncol Nurs, 2006;10:35–41. 40. Jerremalm E, Hedeland M, Wallin I, Bondesson U, Ehrsson H. Oxaliplatin degradation in the presence of chloride: identification and cytotoxicity of the monochloro monooxalato complex. Pharm Res, 2004;21:891–894.
Journal of Clinical Pharmacy and Therapeutics, 2015, 40, 55–62 62
