62 Original Article
Drug Handling in a Paediatric Intensive Care Unit – Can Errors be Prevented by a Three-Step Intervention? Arzneimittelanwendung auf einer pädiatrischen Intensivstation – können Fehler durch eine dreistufige Intervention reduziert werden?
Affiliations
Key words ▶ drug administration ● ▶ medication error ● ▶ paediatric intensive care unit ● ▶ intervention study ● ▶ healthcare quality assurance ● Schlüsselwörter ▶ Arzneimittelanwendung ● ▶ Medikationsfehler ● ▶ pädiatrische Intensivstation ● ▶ Interventionsstudie ● ▶ Qualitätssicherung ●
Bibliography DOI http://dx.doi.org/ 10.1055/s-0033-1364030 Klin Padiatr 2014; 226: 62–67 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0300-8630 Correspondence Prof. Thilo Bertsche Department of Clinical Pharmacy University of Leipzig Eilenburger Straße 15a 04317 Leipzig Germany Tel.: + 49/341/973 66 00 Fax: + 49/341/973 66 09 thilo.bertsche@medizin. uni-leipzig.de
D. Niemann1, 2, A. Bertsche3, 4, D. Meyrath2, S. Oelsner2, A. L. Ewen2, B. Pickardt2, T. Henhapl4, G. Hoffmann4, J. Meyburg4, T. Bertsche1, 2 1
Department of Clinical Pharmacy, University of Leipzig, Leipzig, Germany Department of Clinical Pharmacology and Pharmacoepidemiology, University of Heidelberg, Heidelberg, Germany 3 Hospital for Children and Adolescents, Centre for Pediatric Research, University of Leipzig, Leipzig, Germany 4 University Children's Hospital, University of Heidelberg, Heidelberg, Germany 2
Abstract
Zusammenfassung
Background: Drug handling in paediatric intensive care units (PICU) is prone to medication errors. We aimed to identify type and prevalence of those errors and to assess preventative interventions. Methods: Prospective intervention study investigating a 3-step intervention for preventing errors in drug handling in a 10-bed PICU of a university hospital. Nurses’ drug handling was monitored in daily routine to identify the number of patients affected by errors and overall prevalence and types of errors in drug handling. We implemented a comprehensive intervention consisting of an information handout, a training course, and a 76-page reference book tailored to reduce the prevalence. Results: The prevalence of errors in drug handling decreased from 83 % (555 errors/668 processes) to 63 % (554/883; p < 0.001) after the intervention. The number of affected patients remained unchanged (95 % vs. 89 %, p = 0.370). Peroral (PO) drugs (1.33 errors/process) were more error-prone than intravenous (IV) drugs (0.64), despite being used less frequently (27 % vs. 73 % of all processes, p < 0.001). The interventions decreased the prevalence to 0.77 errors/process (p < 0.001) in PO and to 0.52 in IV drugs (p = 0.025). Conclusion: Errors in drug handling were alarmingly frequent. PO drugs were frequently subject to errors, even though being used less frequently. The implementation of a comprehensive intervention succeeded in reducing the prevalence of errors. Yet further refinements are necessary to decrease also the number of affected patients.
Hintergrund: Die Arzneimittelanwendung auf pädiatrischen Intensivstationen (ITS) ist anfällig für Medikationsfehler. Ziel der Studie war es, Art und Prävalenz solcher Fehler zu identifizieren und präventive Maßnahmen zu untersuchen. Methode: Prospektive Interventionsstudie mit einer 3-stufigen Intervention zur Prävention von Fehlern bei der Arzneimittelanwendung auf einer pädiatrischen ITS (10 Betten) eines Universitätsklinikums. Das Pflegepersonal wurde bei der täglichen Arzneimittelanwendung beobachtet, um die Anzahl von Patienten mit fehlerhafter Arzneimittelanwendung sowie die Prävalenz und Art der Fehler zu identifizieren. Zur Prävention wurde eine 3-stufige Intervention bestehend aus einem Informationshandout, einer Schulung sowie einem 76-seitigen Nachschlagehandbuch entwickelt. Ergebnisse: Die Fehlerprävalenz bei der Arzneimittelanwendung nahm von 83 % (555 Fehler in 668 Prozessen) auf 63 % (554/883; p < 0.001) nach der Intervention ab. Die Anzahl der Patienten mit mindestens einem Fehler blieb unverändert (95 % vs. 89 %, p = 0.370). Perorale (PO) Arzneimittel (1.33 Fehler/Prozess) waren fehleranfälliger als intravenöse (IV) Arzneimittel (0.64), obwohl sie seltener angewendet wurden (27 % vs. 73 % von allen Prozessen, p < 0.001). Die Intervention reduzierte die Prävalenz auf 0.77 Fehler/Prozess (p < 0.001) bei PO und auf 0.52 bei IV Arzneimitteln (p = 0.025). Schlussfolgerung: Fehler bei der Arzneimittelanwendung auf einer pädiatrischen ITS sind alarmierend häufig. PO Arzneimittel sind besonders fehleranfällig, obwohl sie seltener angewendet werden. Die Implementierung einer 3-stufigen Intervention reduzierte die Prävalenz der Fehler. Eine Weiterentwicklung ist jedoch notwendig, um auch die Anzahl der betroffenen Patienten zu reduzieren.
▼
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
▼
Downloaded by: University of Pittsburgh. Copyrighted material.
Authors
Introduction
▼
Children, in particular, are at risk to be affected by adverse drug events (ADEs) resulting from medication errors [8, 25]. Particularly in intensive care settings, the implementation of drug therapies is complex due to the treatment of often rare and life-threatening diseases such as coma diabeticum, which may lead to a continuous adjustment of dosing and drug therapy [29]. Furthermore, the high number of prescribed drugs as well as the frequent use of off-label and critical dose drugs increase the complexity of drug therapy [12, 17, 21]. Nurses require special skills in drug handling, including in drug preparation and administration aimed at error prevention. In young and critically ill children, drug handling itself requires particular attention by nurses to ensure administrating correct doses [26], proper handling of internal feeding tubes [30], and also facilitating for children to accept both their medicine and the medication’s dosage form [12, 17]. The use of excessive amounts of active ingredients and inappropriate dosage forms for children necessitate additional handling steps such as an opening of capsules, a crushing or splitting of tablets, or a diluting of stock solutions and a creation of new opportunities for errors in drug handling [26]. The risk of overdosing may be increased during the dilution of IV drugs with potentially fatal outcomes for patients [16, 25, 26]. Earlier studies demonstrated, that medication errors are frequent and that pharmaceutical services and electronic decision support decreased errors in drug therapy and the resulting ADEs [4–6]. We have now identified medication errors in PO and IV drug handling in a paediatric intensive care unit (PICU) and developed a 3-step intervention program aimed at preventing those errors.
Methods
▼
Setting Following the approval by the responsible Ethics Committee, we performed a prospective intervention study in a 10-bed interdisciplinary PICU, that is embedded in a children’s university hospital with 177 beds and treats patients of all ages from premature infants to young adults.
Study protocol We aimed to identify and prevent medication errors in drug handling, including processes of drug preparation and administration performed by nurses in daily routine in a 10-bed PICU through implementing a comprehensive intervention program. The study consisted of a baseline assessment of prevalent medication errors in drug handling followed by 3 intervention steps. All nurses in the ward were invited to participate in the study. During 26 working days prior to and following each intervention step, trained clinical pharmacists (n = 5) monitored and documented drug handling processes that were carried out by nurses in the regular morning hours of drug administration (7–9 a.m. and 11 a.m.-1 p.m.). Monitors were obliged to intercept any process including an error that may potentially cause serious ADEs such as administration of an incorrect drug (emergency intervention). A senior clinical pharmacist trained the monitors in a 2-day theoretical and practical course and supervised the monitoring and the process documentation. Furthermore, the senior pharmacist
examined each documented process and evaluated the processes regarding potential predefined medication errors in drug handling. An expert panel consisting of paediatricians, nurses, clinical pharmacists and clinical pharmacologists predefined 24 subcategories of medication errors in drug handling, considering internal or external guidelines, manufacturers’ recommendations, the summary of product characteristics as well as former studies of our group. Furthermore, the expert panel (n = 21) rated in an independent survey the clinical relevance of each error subcategory on a scale with four levels: 1 = no, 2 = low, 3 = middle, and 4 = high clinical relevance. The most frequently detected subcategories of errors including their definitions and ▶ Table 1. Furtherthe median clinical relevance are listed in ● more, we performed a questionnaire survey to detect misconceptions and knowledge deficits potentially leading to medication errors in drug handling. The questionnaire included questions about the appropriateness of routine drug handling, knowledge about PO and IV drugs. The content of the questions was tailored to frequently used drugs and processes in the PICU based on data of the baseline assessment. Nurses were given 2 weeks to complete the questionnaire survey. Incorrect or missing answers were defined as knowledge deficit. In order to prevent medication errors in drug handling, we developed and implemented an intervention programme consisting of a 3-page handout for nurses, a 60-min training course (3 appointments in 7 weeks), and a comprehensive 76-page reference book.
Outcomes We assessed the number of patients with medication errors in drug handling processes, the overall prevalence of medication errors concerning the number of processes and the number of errors per patient.
Statistics and data analysis Data is presented as a median with first and third quartile (Q25/ Q75). Comparisons of groups were analysed by Chi2-test, Fisher’s exact test, or 2-sided Mann-Whitney U test, as appropriate and a p-value ≤ 0.05 was considered significant. Calculations were conducted using KyPlot Version 2.0 beta 15 (32 bit) by KyensLab Incorporated, Tokyo, Japan.
Results
▼
Participants In our baseline assessment prior to carrying out the intervention programme, we monitored 28 nurses (82 % female), who were also invited to complete the questionnaire survey, resulting in 16 out of 28 questionnaires being returned to us (response rate: 57 %). After the 3-step intervention we monitored 29 nurses (93 % female). Drug handling was monitored for 38 (42 % female) patients in our assessment prior to the start of the intervention programme, and for a total of 47 (26 % female, p = 0.106) patients following the completion of the full intervention programme. The age of patients ranged from 3 days–21.7 years (median age 2.3 years) for the former group of patients and from 1 day–20.5 years (median age 0.3 years, p = 0.021) for the latter. The errorrate per patient, however, was independent from the age of the patient. The median (Q25/Q75) of monitoring days per patient was 2 (1/4.75) in the baseline assessment and 2 (1/4.5, p = 0.608) after the 3-step intervention.
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
Downloaded by: University of Pittsburgh. Copyrighted material.
Original Article 63
64 Original Article
Table 1 Definition and prevalence of errors in the most frequently detected subcategories of medication errors in drug handling in the PICU. Definition
Clinical rele-
Prevalence
vance [median
[ % (n/total)]
(Q25/Q75)]
PO drugs: withdrawing PO fluids with a syringe directly from the bottle pouring excess fluid back into the bottle incorrect dissolving/ suspending of drugs for PO administration no shaking of suspensions mixing drugs incorrect administration of drugs via feeding tube IV drugs: incorrect solvent for IV drugs incorrect volume of solvent for IV drugs incorrect IV drug administration panning a 24 h-infusion
PO or IV drugs: wrong or missing labelling no self-protection (occupational exposure risk)
neglected hygiene
p-value before vs. after intervention
Before
After
intervention
intervention
withdrawing PO fluids with a syringe directly from the bottle without using a special adapter or a cup. withdrawing too much of a PO liquid and pouring excess amounts back into the bottle. dissolving/suspending tablets or capsules for PO administration deviating from guidelines, or summary of product characteristics. shaking of suspensions (PO liquids) was not done before use. mixing drugs by crushing or suspending before administration administering PO drugs via a feeding tube deviating from guidelines, standards, or summary of product characteristics.
2 (2/2)
91 % (42/46)
70 % (66/94)
0.005
2 (2/3)
11 % (5/46)
18 % (17/94)
0.271
4 (3/4)
42 % (19/45)
10 % (9/86)
< 0.001
4 (3/4)
86 % (21/31)
26 % (11/42)
< 0.001
3 (3/4)
14 % (18/128)
1 % (3/304)
< 0.001
3 (3/4)
96 % (26/27)
66 % (38/58)
0.002
dissolving an IV drug in a solvent that deviates from the manufacturers’ recommendation. dissolving an IV drug in a volume that deviates from the manufacturers’ recommendation. administering IV drugs in a way deviating from guidelines. 24 h-infusions in syringes were not brandished after preparation following to an inconsistent concentration in the syringe.
3 (3/4)
6 % (16/252)
3 % (8/245)
0.108
3 (3/4)
9 % (22/252)
7 % (18/245)
0.571
3 (3/4)
31 % (15/49)
10 % (7/73)
< 0.001
3 (3/3)
51 % (65/128)
71 % (84/118)
< 0.001
no or incorrect labelling of prepared drugs. self-protection, e. g. wearing gloves, was not done while handling or administering immunosuppressive, carcinogenic, mutagenic, or teratogenic drugs. not following guidelines and standards of hygiene
4 (4/4) 3 (3/4)
15 % (77/514) 82 % (31/38)
17 % (118/698) 38 % (5/13)
3 (3/4)
20 % (132/668)
11 % (98/883)
0.367 0.003
< 0.001
The subcategories of errors were rated by the expert panel regarding clinical relevance on a scale with four levels: 1 = no, 2 = low, 3 = middle, and 4 = high clinical relavance
Prevalence of medication errors in drug handling In 2 752 drug handling processes we identified 1 913 medication errors. In the baseline assessment, the overall prevalence of errors was 83 % (555 errors in 668 processes) and 95 % (36/38) of the patients were affected by at least one error. Following the implementation of the full intervention programme, the overall error prevalence decreased to 63 % (554 errors in 883 processes; p < 0.001), whereas the number of patients with at least one error remained unchanged (89 %; 42/47; p = 0.370). The median (Q25/Q75) number of errors per patient with 4.5 (2/10.75) before and 6 (1/12.5) after the intervention was not changed (p = 0.860). The frequency of error-free processes raised from 43 % (284/668) in the baseline assessment to 50 % (438/883, p = 0.006) after the intervention. In the baseline assessment 9 % (47/536) of the errors had a clinical relevance at low level, 65 % (351/536) at middle-level, and 26 % (138/536) had high level. After the intervention the frequency of errors with a minor clinical relevance was with 16 % (84/534) increased in comparison to
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
the baseline assessment (p < 0.001). The frequency of errors middle-level decreased after the intervention to 54 % (290/534, p < 0.001). The frequency of error with high clinical relevance was with 30 % (160/534) unchanged (p = 0.124) after the implementation of the intervention. Errors in the subcategory “Others” (19/555 in the baseline assessment and 20/554 after the intervention) were not analysed regarding clinical relevance. PO drugs were less frequently used than IV drugs (baseline assessment 27 % PO drugs vs. 73 % IV drugs) but were more error-prone than IV drugs (p < 0.001). We identified 1.33 errors/drug handling process (241/181) concerning PO drugs compared to 0.64 errors/drug handling process (313/486) in IV drugs (p < 0.001). Through the intervention, this rate decreased to 0.77 errors/ process (290/376; p < 0.001) involving PO drugs and 0.52 errors/ ▶ Table 1 lists process (260/502; p = 0.025) involving IV drugs. ● the predefined subcategories of medication errors in the PICU. The drug groups involved in medication errors and the fre▶ Table 2. quency of administration of those drugs are shown in ●
Downloaded by: University of Pittsburgh. Copyrighted material.
Subcategory
Original Article 65
Table 2 Impact of a 3-step intervention on drug groups. Before intervention (B)
After intervention (I)
Frequency of adminis-
Errors/drug
Frequency of admin-
Errors/drug
p-Value errors/
tration [ %; total drugs
[n/total]
istration [ %; total
[n/total]
drug (B vs. I)
electrolytes antiinfective agents antithrombotic agents immunosuppressive agents (incl. glucocorticoids) antihypertensive agents analgesics micronutrients/vitamins/amino acids gastrointestinal agents cardiotonic agents miscellaneous agents anticonvulsants hypnotics and sedatives expectorants and bronchodilator agents caffeine thyroid hormones
26 16 9 8
0.46 (89/193) 0.74 (86/116) 1.00 (67/67) 1.34 (83/62)
19 17 3 8
0.35 (69/200) 0.5 (89/179) 1.06 (37/35) 0.62 (50/81)
0.052 0.009 0.476 0.004
8 7 7 5 3 3 2 2 1 1 1
0.82 (46/56) 0.87 (47/54) 1.20 (61/51) 0.85 (34/40) 0.76 (16/21) 0.60 (12/20) 1.00 (18/18) 0.33 (6/18) 0.86 (6/7) 1.17 (7/6) 0.60 (3/5)
6 4 9 7 2 5 4 7 2 6 3
0.65 (43/66) 0.82 (31/38) 0.76 (69/96) 0.55 (41/74) 0.88 (14/16) 0.66 (35/53) 0.36 (16/45) 0.76 (58/76) 0.35 (6/17) 1.17 (72/62) 0.56 (15/27)
0.242 0.577 0.001 0.028 0.688 0.120 0.001 0.028 n. a. n. a. n. a.
n. a.: not assessable
Knowledge deficits in the questionnaire survey In the questionnaire survey, we identified knowledge deficits in 25 % of the responses (157 wrong or missing answers in a total of 640 questions). Nurses were less familiar with IV drug administration [29 % (99/336) knowledge deficits] than with PO drug administration [20 % (58/288, p = 0.007)]. There was no correlation between the knowledge deficit of the nurses and their work experience in years (correlation coefficient R2 = 0.016, likelihood 95 %).
Emergency interventions In total, the PICU monitoring team intervened 7 times, thus preventing in one case a mix-up of ganciclovir and meropenem, in another a mix-up of cefuroxime and cefotaxime, and in 2 cases each an administration of expired caffeine solutions, a preparation of an incorrect dose of a continuous heparin infusion and a caffeine solution, and a preparation of a 10 % saline solution with was not prescribed to the patient.
Discussion
▼
Medication errors in drug handling in the PICU were rather frequent affecting 4 in 5 drug handling processes. Our intervention programme consisting of 3 intervention steps reduced the overall relative risk of medication errors in drug handling by nearly 25 %. The number of intensive care patients, however, with at least one error in drug handling did not decrease significantly, due most likely to the persistence of a considerable number of errors. Furthermore, the intervention was variably successful across dosage forms and subcategories of medication errors. As expected, less frequently handled drugs such as PO drugs were more error-prone than the more common IV drugs. In comparison to published data in paediatric wards [13, 22, 23], we observed a notably high prevalence of 83 % medication errors in the PICU. This is likely due to differences in the observation processes across studies. We decided to use monitoring to detect medication errors in daily routine. Monitoring was conducted by several monitors to avoid fatigue in objective process docu-
mentation as well as to minimise the influence of monitors’ expectation on the study data (Rosenthal effect). Furthermore, comparisons are made often with studies in adult patients that used similar observation methods and comparable error descriptions [4]. Our intervention programme, in contrast, decreased the prevalence to 63 %, albeit with variable success across dosage forms and predefined subcategories of medication errors. Although IV drug handling is considered particularly errorprone [1, 20, 27], in our PICU medication errors with IV drugs were less frequent than with PO drugs (0.64 vs. 1.33 errors/process). In our PICU, PO drugs were less frequently administered than IV drugs while nurses had less knowledge deficits in PO drugs than in IV drugs. Therefore, we assume that the staff was less familiar with the actual handling or had a reduced awareness of potential errors and potential resulting serious ADEs [7]. We also should bear in mind, that it is easier to reduce excessive error-rates such as the error-rate in PO drugs than comparatively low error-rates in the baseline assessment such as in IV drugs. This is in line with our experience that the 3-step intervention prevented mainly medication errors in PO drug handling while the rates in IV drugs were not influenced to the same extent. Particularly in young and critically ill children, additional handling steps become necessary such as the splitting or crushing of tablets. These steps provide new occasions for errors to emerge and should hence be looked at in detail. This way, nurses’ awareness of potential ADEs resulting from incorrect drug handling also in less frequently administered dosage forms such as PO drugs was likely raised by our intervention. Furthermore, in the interdisciplinary setting of the PICU a lot of different drugs including high-risk drugs were administered [12]. Strategies to improve quality frequently prioritize critical dose drugs, drugs that were administered most frequently or drugs known as error-prone [4]. Our data, however, suggest the necessity also to consider rarely administered as well as non high-risk drug groups such as electrolytes or gastrointestinal drugs, which also bear a certain risk for ADEs when handled incorrectly [28]. For instance, ranitidine may lead to nausea and vomiting when administered undiluted and should be handled with special care in children and newborns [24]. Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
Downloaded by: University of Pittsburgh. Copyrighted material.
drugs n = 1065]
n = 734]
66 Original Article
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
A programme, even addressing multiple factors of error prevention may not lead to its full success if not repeated for each new nurse. In our study, the response rate in the questionnaire survey was low with only 57 %. This indicates that for further steps in our (and other) setting incentives are required for the nurses to take part in such a programme.
Limitations
▼
(i) We focused on medication errors in drug handling without attempting to identify actual clinical ADEs that may result from these errors. (ii) A potential observation bias, possibly induced by the presence of a monitor (Hawthorne effect) may limit our results and may have caused an underestimation of the extent of the problem. (iii) In the observation, we did not include weekend, evening, and night shifts, in which the prevalence of errors may have been different.
Conclusion
▼
Drug handling errors in paediatric intensive care are highly frequent. The implementation of a comprehensive intervention programme decreased the error-prevalence. However, its success varied across dosage forms and subcategories of errors. We propose to monitor drug handling processes and to develop intervention programmes, including potentially a special course aimed at new nurses and continuous training in drug handling processes, as a means to prevent medication errors effectively.
Acknowledgements
▼
We would like to thank all participating nurses and physicians for their kind support. We thank Prof. Dr. Walter E. Haefeli for his helpful remarks in developing the manuscript and Prof. Dr. Thomas Weber for language editing of the manuscript.
Conflict of interest: The authors have no conflict of interest to disclose. References 1 Barber N, Taxis K. Incidence and severity of intravenous drug errors in a German hospital. Eur J Clin Pharmacol 2004; 59: 815–817 2 Bates DW. Preventing medication errors: A summary. Am J Health Syst Pharm 2007; 64 (14 Supplement 9): S3–S9 3 Berdot S, Sabatier B, Gillaizeau F et al. Evaluation of drug administration errors in a teaching hospital. BMC Health Serv Res 2012; 12: 60 4 Bertsche T, Niemann D, Mayer Y et al. Prioritising the prevention of medication handling errors. Pharm World Sci 2008; 30: 907–915 5 Bertsche T, Bertsche A, Krieg EM et al. Prospective pilot intervention study to prevent medication errors in drugs administered to children by mouth or gastric tube: a programme for nurses, physicians and parents. Qual Saf Health Care 2010; 19: e26 (1–5) 6 Bertsche T, Pfaff J, Schiller P et al. Prevention of adverse drug reactions in intensive care patients by personal intervention based on an electronic clinical decision support system. Intensive Care Med 2010; 36: 665–672 7 Cornish P. “Avoid the crush”: hazards of medication administration in patients with dysphagia or a feeding tube. CMAJ 2005; 172: 871–872 8 Dabliz R, Levine S. Medication Safety in Neonates. Am J Perinatol 2011; 29: 49–56 9 Davis L, Ware R, McCann D et al. Evaluation of contextual influences on the medication administration practice of paediatric nurses. J Adv Nurs 2009; 65: 1293–1299
Downloaded by: University of Pittsburgh. Copyrighted material.
In line with previous studies, we confirm in our study that antiinfective agents were administered with a high prevalence and were frequently involved in errors. Antiinfective agents are often in the focus of interventions such as an Antibiotic Stewardship Initiative that aimed to restrict and control the administration of antibiotics to prevent antimicrobial resistance as well as to minimise costs [15]. In these programmes, however, drug handling aspects itself are frequently neglected. The prevention of medication errors varied in their success rate across the predefined subcategories of errors. While the intervention decreased errors in the subcategory “Incorrect IV drug administration” from 31 to 6 %, it failed to improve medication errors of the subcategory “Not panning a 24 h-infusion”. This subcategory of medication errors, however, was mentioned by each intervention strategy but surprisingly the frequency of errors in this subcategory increased after the intervention. The missing error-reduction in this category is in line with literature stressing that nurses are not always aware of the necessity of following rules [9]. Other surprisingly frequent errors, including PO drugs with apparently self-explanatory steps such as shaking of suspensions (68 %) were responsive to the intervention. According to the literature, the health care team appears to lack awareness of the consequences [2, 18] such as under- and overdosage of the drug and changes of the concentration in the bottle, thereby putting further patients at risk for dosing deviations. Furthermore, our intervention was successful in preventing errors in subcategories such as “Incorrect dissolving/suspending of drugs for PO administration” (42 % reduced to 10 %), or “mixing drugs” (14 % reduced to 1 %) after the intervention programme. The error-prevention in these subcategories stands in contrast to previous studies, which often highlighted drugrelated errors such as wrong-time [3, 22, 23], wrong-administration-technique [23], or patient misidentification [8]. We, however, could not confirm such a high prevalence of those errors in the mentioned categories. This study underlines the different involvement of error subcategories in drug handling processes. Furthermore, it identified many ward-specific areas such as preparing 24 h infusion pumps and subcategories of medication errors, which might be helpful in future error prevention strategies. Although the number of affected patients did not decrease significantly, the overall prevalence of medication errors decreased from 83 % to 63 %. One explanation for the only moderate success of the intervention may be the work load [10, 14] and psychological stress of the nurses [19] in an intensive care unit. In comparison to general wards, in the PICU nursing care is more complex and laborious and drug handling, which can take up to 15 min [11], is only a small part of the overall nursing programme. Additionally, in our PICU fluctuations in the nursing team were frequent. Some nurses stayed only for a few months in the unit because they took part in a specialisation programme for paediatric intensive care and normally worked in another hospital. Standards in nursing care and drug handling may vary between hospitals, whereby drug handling processes may be conducted heterogeneously in the PICU. For further studies, we would, therefore, recommend to design a beginners’ programme for each new nurse, e. g. including a questionnaire survey about standard drug handling processes in the PICU, a subsequent training course by a pharmacist, and the issuance of a wardbased reference book. Continuous training in correct drug handling for all nurses may be a further solution to prevent medication errors in drug handling effectively and sustainably.
10 Endacott R. The continuing imperative to measure workload in ICU: impact on patient safety and staff well-being. Intensive Care Med 2012; 38: 1415–1417 11 Elganzouri ES, Standish CA, Androwich I. Medication Administration Time Study (MATS): nursing staff performance of medication administration. J Nurs Adm 2009; 39: 204–210 12 Franke HA, Woods DM, Holl JL. High-alert medications in the pediatric intensive care unit. Pediatr Crit Care Med 2009; 10: 85–90 13 Ghaleb MA, Barber N, Franklin BD et al. The incidence and nature of prescribing and medication administration errors in paediatric inpatients. Arch Dis Child 2010; 95: 113–118 14 Hiemisch A, Kiess W, Brähler E. Mental Job Strain in an University Children’s Hospital – A Study on Stress Experience and the Resulting Employee Health. Klin Padiatr 2011; 223: 236–241 15 Huebner J, Rack-Hoch AI, Pear A et al. Pilot Project of an Pediatric Antibiotic Stewardshio Initiative at the Hauner Children’s Hospital. Klin Padiatr 2013; 225: 223–229 16 Koren G, Haslam RH. Pediatric medication errors: predicting and preventing tenfold disasters. J Clin Pharmacol 1994; 34: 1043–1045 17 McDonnell C, Hum S, Frndova H et al. Pharmacotherapy in pediatric critical illness: a prospective observational study. Paediatr Drugs 2009; 11: 323–331 18 Merry AF, Anderson BJ. Medication errors – new approaches to prevention. Pediatr Anesth 2011; 21: 743–753 19 Pardoe P. Psychological support for nurses on paediatric intensive care units. Nurs Child Young People 2011; 23: 27–29 20 Parshuram CS, To T, Seto W et al. Systematic evaluation of errors occurring during the preparation of intravenous medication. CMAJ 2008; 178: 42–48
21 Prandstretter C, Tamesberger M, Wagner O et al. Medical prescriptions to premature and newborn infants in an Austrian Neonatal Intensive Care Unit. Klin Padiatr 2009; 221: 312–317 22 Prot S. Drug administration errors and their determinants in pediatric in-patients. Int J Qual Health Care 2005; 17: 381–389 23 Schneider MP, Cotting J, Pannatier A. Evaluation of nurses’ errors associated in the preparation and administration of medication in a pediatric intensive care unit. Pharm World Sci 1998; 20: 178–182 24 Terrin G, Passariello A, De Curtis M et al. Ranitidine is Associated With Infections, Necrotizing Enterocolitis, and Fatal Outcome in Newborns. Pediatrics 2011; 129: e40–e45 25 Tzimenatos L, Bond GR. Severe injury or death in young children from therapeutic errors: a summary of 238 cases from the American Association of Poison Control Centers. Clin Toxicol 2009; 47: 348–354 26 Uppal N, Yasseen B, Seto W et al. Drug formulations that require less than 0.1 mL of stock solution to prepare doses for infants and children. CMAJ 2011; 183: E246–E248 27 Westbrook JI, Woods A, Rob MI et al. Association of interruptions with an increased risk and severity of medication administration errors. Arch Intern Med 2010; 170: 683–690 28 Wetherton AR, Corey TS, Buchino JJ et al. Fatal intravenous injection of potassium in hospitalized patients. Am J Forensic Med Pathol 2003; 24: 128–131 29 Wetter J, Gohlke S, Stutte S et al. Pediatric Hyperglycemic Hyperosmolar Coma Diabeticum: Diagnostic Evaluation and Therapeutic Concept. Klin Padiatr 2012; 224: 26–31 30 Williams NT. Medication administration through enteral feeding tubes. Am J Health Syst Pharm 2008; 65: 2347–2357
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
Downloaded by: University of Pittsburgh. Copyrighted material.
Original Article 67
Drug Handling in a Paediatric Intensive Care Unit – Can Errors be Prevented by a Three-Step Intervention? Arzneimittelanwendung auf einer pädiatrischen Intensivstation – können Fehler durch eine dreistufige Intervention reduziert werden?
Affiliations
Key words ▶ drug administration ● ▶ medication error ● ▶ paediatric intensive care unit ● ▶ intervention study ● ▶ healthcare quality assurance ● Schlüsselwörter ▶ Arzneimittelanwendung ● ▶ Medikationsfehler ● ▶ pädiatrische Intensivstation ● ▶ Interventionsstudie ● ▶ Qualitätssicherung ●
Bibliography DOI http://dx.doi.org/ 10.1055/s-0033-1364030 Klin Padiatr 2014; 226: 62–67 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0300-8630 Correspondence Prof. Thilo Bertsche Department of Clinical Pharmacy University of Leipzig Eilenburger Straße 15a 04317 Leipzig Germany Tel.: + 49/341/973 66 00 Fax: + 49/341/973 66 09 thilo.bertsche@medizin. uni-leipzig.de
D. Niemann1, 2, A. Bertsche3, 4, D. Meyrath2, S. Oelsner2, A. L. Ewen2, B. Pickardt2, T. Henhapl4, G. Hoffmann4, J. Meyburg4, T. Bertsche1, 2 1
Department of Clinical Pharmacy, University of Leipzig, Leipzig, Germany Department of Clinical Pharmacology and Pharmacoepidemiology, University of Heidelberg, Heidelberg, Germany 3 Hospital for Children and Adolescents, Centre for Pediatric Research, University of Leipzig, Leipzig, Germany 4 University Children's Hospital, University of Heidelberg, Heidelberg, Germany 2
Abstract
Zusammenfassung
Background: Drug handling in paediatric intensive care units (PICU) is prone to medication errors. We aimed to identify type and prevalence of those errors and to assess preventative interventions. Methods: Prospective intervention study investigating a 3-step intervention for preventing errors in drug handling in a 10-bed PICU of a university hospital. Nurses’ drug handling was monitored in daily routine to identify the number of patients affected by errors and overall prevalence and types of errors in drug handling. We implemented a comprehensive intervention consisting of an information handout, a training course, and a 76-page reference book tailored to reduce the prevalence. Results: The prevalence of errors in drug handling decreased from 83 % (555 errors/668 processes) to 63 % (554/883; p < 0.001) after the intervention. The number of affected patients remained unchanged (95 % vs. 89 %, p = 0.370). Peroral (PO) drugs (1.33 errors/process) were more error-prone than intravenous (IV) drugs (0.64), despite being used less frequently (27 % vs. 73 % of all processes, p < 0.001). The interventions decreased the prevalence to 0.77 errors/process (p < 0.001) in PO and to 0.52 in IV drugs (p = 0.025). Conclusion: Errors in drug handling were alarmingly frequent. PO drugs were frequently subject to errors, even though being used less frequently. The implementation of a comprehensive intervention succeeded in reducing the prevalence of errors. Yet further refinements are necessary to decrease also the number of affected patients.
Hintergrund: Die Arzneimittelanwendung auf pädiatrischen Intensivstationen (ITS) ist anfällig für Medikationsfehler. Ziel der Studie war es, Art und Prävalenz solcher Fehler zu identifizieren und präventive Maßnahmen zu untersuchen. Methode: Prospektive Interventionsstudie mit einer 3-stufigen Intervention zur Prävention von Fehlern bei der Arzneimittelanwendung auf einer pädiatrischen ITS (10 Betten) eines Universitätsklinikums. Das Pflegepersonal wurde bei der täglichen Arzneimittelanwendung beobachtet, um die Anzahl von Patienten mit fehlerhafter Arzneimittelanwendung sowie die Prävalenz und Art der Fehler zu identifizieren. Zur Prävention wurde eine 3-stufige Intervention bestehend aus einem Informationshandout, einer Schulung sowie einem 76-seitigen Nachschlagehandbuch entwickelt. Ergebnisse: Die Fehlerprävalenz bei der Arzneimittelanwendung nahm von 83 % (555 Fehler in 668 Prozessen) auf 63 % (554/883; p < 0.001) nach der Intervention ab. Die Anzahl der Patienten mit mindestens einem Fehler blieb unverändert (95 % vs. 89 %, p = 0.370). Perorale (PO) Arzneimittel (1.33 Fehler/Prozess) waren fehleranfälliger als intravenöse (IV) Arzneimittel (0.64), obwohl sie seltener angewendet wurden (27 % vs. 73 % von allen Prozessen, p < 0.001). Die Intervention reduzierte die Prävalenz auf 0.77 Fehler/Prozess (p < 0.001) bei PO und auf 0.52 bei IV Arzneimitteln (p = 0.025). Schlussfolgerung: Fehler bei der Arzneimittelanwendung auf einer pädiatrischen ITS sind alarmierend häufig. PO Arzneimittel sind besonders fehleranfällig, obwohl sie seltener angewendet werden. Die Implementierung einer 3-stufigen Intervention reduzierte die Prävalenz der Fehler. Eine Weiterentwicklung ist jedoch notwendig, um auch die Anzahl der betroffenen Patienten zu reduzieren.
▼
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
▼
Downloaded by: University of Pittsburgh. Copyrighted material.
Authors
Introduction
▼
Children, in particular, are at risk to be affected by adverse drug events (ADEs) resulting from medication errors [8, 25]. Particularly in intensive care settings, the implementation of drug therapies is complex due to the treatment of often rare and life-threatening diseases such as coma diabeticum, which may lead to a continuous adjustment of dosing and drug therapy [29]. Furthermore, the high number of prescribed drugs as well as the frequent use of off-label and critical dose drugs increase the complexity of drug therapy [12, 17, 21]. Nurses require special skills in drug handling, including in drug preparation and administration aimed at error prevention. In young and critically ill children, drug handling itself requires particular attention by nurses to ensure administrating correct doses [26], proper handling of internal feeding tubes [30], and also facilitating for children to accept both their medicine and the medication’s dosage form [12, 17]. The use of excessive amounts of active ingredients and inappropriate dosage forms for children necessitate additional handling steps such as an opening of capsules, a crushing or splitting of tablets, or a diluting of stock solutions and a creation of new opportunities for errors in drug handling [26]. The risk of overdosing may be increased during the dilution of IV drugs with potentially fatal outcomes for patients [16, 25, 26]. Earlier studies demonstrated, that medication errors are frequent and that pharmaceutical services and electronic decision support decreased errors in drug therapy and the resulting ADEs [4–6]. We have now identified medication errors in PO and IV drug handling in a paediatric intensive care unit (PICU) and developed a 3-step intervention program aimed at preventing those errors.
Methods
▼
Setting Following the approval by the responsible Ethics Committee, we performed a prospective intervention study in a 10-bed interdisciplinary PICU, that is embedded in a children’s university hospital with 177 beds and treats patients of all ages from premature infants to young adults.
Study protocol We aimed to identify and prevent medication errors in drug handling, including processes of drug preparation and administration performed by nurses in daily routine in a 10-bed PICU through implementing a comprehensive intervention program. The study consisted of a baseline assessment of prevalent medication errors in drug handling followed by 3 intervention steps. All nurses in the ward were invited to participate in the study. During 26 working days prior to and following each intervention step, trained clinical pharmacists (n = 5) monitored and documented drug handling processes that were carried out by nurses in the regular morning hours of drug administration (7–9 a.m. and 11 a.m.-1 p.m.). Monitors were obliged to intercept any process including an error that may potentially cause serious ADEs such as administration of an incorrect drug (emergency intervention). A senior clinical pharmacist trained the monitors in a 2-day theoretical and practical course and supervised the monitoring and the process documentation. Furthermore, the senior pharmacist
examined each documented process and evaluated the processes regarding potential predefined medication errors in drug handling. An expert panel consisting of paediatricians, nurses, clinical pharmacists and clinical pharmacologists predefined 24 subcategories of medication errors in drug handling, considering internal or external guidelines, manufacturers’ recommendations, the summary of product characteristics as well as former studies of our group. Furthermore, the expert panel (n = 21) rated in an independent survey the clinical relevance of each error subcategory on a scale with four levels: 1 = no, 2 = low, 3 = middle, and 4 = high clinical relevance. The most frequently detected subcategories of errors including their definitions and ▶ Table 1. Furtherthe median clinical relevance are listed in ● more, we performed a questionnaire survey to detect misconceptions and knowledge deficits potentially leading to medication errors in drug handling. The questionnaire included questions about the appropriateness of routine drug handling, knowledge about PO and IV drugs. The content of the questions was tailored to frequently used drugs and processes in the PICU based on data of the baseline assessment. Nurses were given 2 weeks to complete the questionnaire survey. Incorrect or missing answers were defined as knowledge deficit. In order to prevent medication errors in drug handling, we developed and implemented an intervention programme consisting of a 3-page handout for nurses, a 60-min training course (3 appointments in 7 weeks), and a comprehensive 76-page reference book.
Outcomes We assessed the number of patients with medication errors in drug handling processes, the overall prevalence of medication errors concerning the number of processes and the number of errors per patient.
Statistics and data analysis Data is presented as a median with first and third quartile (Q25/ Q75). Comparisons of groups were analysed by Chi2-test, Fisher’s exact test, or 2-sided Mann-Whitney U test, as appropriate and a p-value ≤ 0.05 was considered significant. Calculations were conducted using KyPlot Version 2.0 beta 15 (32 bit) by KyensLab Incorporated, Tokyo, Japan.
Results
▼
Participants In our baseline assessment prior to carrying out the intervention programme, we monitored 28 nurses (82 % female), who were also invited to complete the questionnaire survey, resulting in 16 out of 28 questionnaires being returned to us (response rate: 57 %). After the 3-step intervention we monitored 29 nurses (93 % female). Drug handling was monitored for 38 (42 % female) patients in our assessment prior to the start of the intervention programme, and for a total of 47 (26 % female, p = 0.106) patients following the completion of the full intervention programme. The age of patients ranged from 3 days–21.7 years (median age 2.3 years) for the former group of patients and from 1 day–20.5 years (median age 0.3 years, p = 0.021) for the latter. The errorrate per patient, however, was independent from the age of the patient. The median (Q25/Q75) of monitoring days per patient was 2 (1/4.75) in the baseline assessment and 2 (1/4.5, p = 0.608) after the 3-step intervention.
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
Downloaded by: University of Pittsburgh. Copyrighted material.
Original Article 63
64 Original Article
Table 1 Definition and prevalence of errors in the most frequently detected subcategories of medication errors in drug handling in the PICU. Definition
Clinical rele-
Prevalence
vance [median
[ % (n/total)]
(Q25/Q75)]
PO drugs: withdrawing PO fluids with a syringe directly from the bottle pouring excess fluid back into the bottle incorrect dissolving/ suspending of drugs for PO administration no shaking of suspensions mixing drugs incorrect administration of drugs via feeding tube IV drugs: incorrect solvent for IV drugs incorrect volume of solvent for IV drugs incorrect IV drug administration panning a 24 h-infusion
PO or IV drugs: wrong or missing labelling no self-protection (occupational exposure risk)
neglected hygiene
p-value before vs. after intervention
Before
After
intervention
intervention
withdrawing PO fluids with a syringe directly from the bottle without using a special adapter or a cup. withdrawing too much of a PO liquid and pouring excess amounts back into the bottle. dissolving/suspending tablets or capsules for PO administration deviating from guidelines, or summary of product characteristics. shaking of suspensions (PO liquids) was not done before use. mixing drugs by crushing or suspending before administration administering PO drugs via a feeding tube deviating from guidelines, standards, or summary of product characteristics.
2 (2/2)
91 % (42/46)
70 % (66/94)
0.005
2 (2/3)
11 % (5/46)
18 % (17/94)
0.271
4 (3/4)
42 % (19/45)
10 % (9/86)
< 0.001
4 (3/4)
86 % (21/31)
26 % (11/42)
< 0.001
3 (3/4)
14 % (18/128)
1 % (3/304)
< 0.001
3 (3/4)
96 % (26/27)
66 % (38/58)
0.002
dissolving an IV drug in a solvent that deviates from the manufacturers’ recommendation. dissolving an IV drug in a volume that deviates from the manufacturers’ recommendation. administering IV drugs in a way deviating from guidelines. 24 h-infusions in syringes were not brandished after preparation following to an inconsistent concentration in the syringe.
3 (3/4)
6 % (16/252)
3 % (8/245)
0.108
3 (3/4)
9 % (22/252)
7 % (18/245)
0.571
3 (3/4)
31 % (15/49)
10 % (7/73)
< 0.001
3 (3/3)
51 % (65/128)
71 % (84/118)
< 0.001
no or incorrect labelling of prepared drugs. self-protection, e. g. wearing gloves, was not done while handling or administering immunosuppressive, carcinogenic, mutagenic, or teratogenic drugs. not following guidelines and standards of hygiene
4 (4/4) 3 (3/4)
15 % (77/514) 82 % (31/38)
17 % (118/698) 38 % (5/13)
3 (3/4)
20 % (132/668)
11 % (98/883)
0.367 0.003
< 0.001
The subcategories of errors were rated by the expert panel regarding clinical relevance on a scale with four levels: 1 = no, 2 = low, 3 = middle, and 4 = high clinical relavance
Prevalence of medication errors in drug handling In 2 752 drug handling processes we identified 1 913 medication errors. In the baseline assessment, the overall prevalence of errors was 83 % (555 errors in 668 processes) and 95 % (36/38) of the patients were affected by at least one error. Following the implementation of the full intervention programme, the overall error prevalence decreased to 63 % (554 errors in 883 processes; p < 0.001), whereas the number of patients with at least one error remained unchanged (89 %; 42/47; p = 0.370). The median (Q25/Q75) number of errors per patient with 4.5 (2/10.75) before and 6 (1/12.5) after the intervention was not changed (p = 0.860). The frequency of error-free processes raised from 43 % (284/668) in the baseline assessment to 50 % (438/883, p = 0.006) after the intervention. In the baseline assessment 9 % (47/536) of the errors had a clinical relevance at low level, 65 % (351/536) at middle-level, and 26 % (138/536) had high level. After the intervention the frequency of errors with a minor clinical relevance was with 16 % (84/534) increased in comparison to
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
the baseline assessment (p < 0.001). The frequency of errors middle-level decreased after the intervention to 54 % (290/534, p < 0.001). The frequency of error with high clinical relevance was with 30 % (160/534) unchanged (p = 0.124) after the implementation of the intervention. Errors in the subcategory “Others” (19/555 in the baseline assessment and 20/554 after the intervention) were not analysed regarding clinical relevance. PO drugs were less frequently used than IV drugs (baseline assessment 27 % PO drugs vs. 73 % IV drugs) but were more error-prone than IV drugs (p < 0.001). We identified 1.33 errors/drug handling process (241/181) concerning PO drugs compared to 0.64 errors/drug handling process (313/486) in IV drugs (p < 0.001). Through the intervention, this rate decreased to 0.77 errors/ process (290/376; p < 0.001) involving PO drugs and 0.52 errors/ ▶ Table 1 lists process (260/502; p = 0.025) involving IV drugs. ● the predefined subcategories of medication errors in the PICU. The drug groups involved in medication errors and the fre▶ Table 2. quency of administration of those drugs are shown in ●
Downloaded by: University of Pittsburgh. Copyrighted material.
Subcategory
Original Article 65
Table 2 Impact of a 3-step intervention on drug groups. Before intervention (B)
After intervention (I)
Frequency of adminis-
Errors/drug
Frequency of admin-
Errors/drug
p-Value errors/
tration [ %; total drugs
[n/total]
istration [ %; total
[n/total]
drug (B vs. I)
electrolytes antiinfective agents antithrombotic agents immunosuppressive agents (incl. glucocorticoids) antihypertensive agents analgesics micronutrients/vitamins/amino acids gastrointestinal agents cardiotonic agents miscellaneous agents anticonvulsants hypnotics and sedatives expectorants and bronchodilator agents caffeine thyroid hormones
26 16 9 8
0.46 (89/193) 0.74 (86/116) 1.00 (67/67) 1.34 (83/62)
19 17 3 8
0.35 (69/200) 0.5 (89/179) 1.06 (37/35) 0.62 (50/81)
0.052 0.009 0.476 0.004
8 7 7 5 3 3 2 2 1 1 1
0.82 (46/56) 0.87 (47/54) 1.20 (61/51) 0.85 (34/40) 0.76 (16/21) 0.60 (12/20) 1.00 (18/18) 0.33 (6/18) 0.86 (6/7) 1.17 (7/6) 0.60 (3/5)
6 4 9 7 2 5 4 7 2 6 3
0.65 (43/66) 0.82 (31/38) 0.76 (69/96) 0.55 (41/74) 0.88 (14/16) 0.66 (35/53) 0.36 (16/45) 0.76 (58/76) 0.35 (6/17) 1.17 (72/62) 0.56 (15/27)
0.242 0.577 0.001 0.028 0.688 0.120 0.001 0.028 n. a. n. a. n. a.
n. a.: not assessable
Knowledge deficits in the questionnaire survey In the questionnaire survey, we identified knowledge deficits in 25 % of the responses (157 wrong or missing answers in a total of 640 questions). Nurses were less familiar with IV drug administration [29 % (99/336) knowledge deficits] than with PO drug administration [20 % (58/288, p = 0.007)]. There was no correlation between the knowledge deficit of the nurses and their work experience in years (correlation coefficient R2 = 0.016, likelihood 95 %).
Emergency interventions In total, the PICU monitoring team intervened 7 times, thus preventing in one case a mix-up of ganciclovir and meropenem, in another a mix-up of cefuroxime and cefotaxime, and in 2 cases each an administration of expired caffeine solutions, a preparation of an incorrect dose of a continuous heparin infusion and a caffeine solution, and a preparation of a 10 % saline solution with was not prescribed to the patient.
Discussion
▼
Medication errors in drug handling in the PICU were rather frequent affecting 4 in 5 drug handling processes. Our intervention programme consisting of 3 intervention steps reduced the overall relative risk of medication errors in drug handling by nearly 25 %. The number of intensive care patients, however, with at least one error in drug handling did not decrease significantly, due most likely to the persistence of a considerable number of errors. Furthermore, the intervention was variably successful across dosage forms and subcategories of medication errors. As expected, less frequently handled drugs such as PO drugs were more error-prone than the more common IV drugs. In comparison to published data in paediatric wards [13, 22, 23], we observed a notably high prevalence of 83 % medication errors in the PICU. This is likely due to differences in the observation processes across studies. We decided to use monitoring to detect medication errors in daily routine. Monitoring was conducted by several monitors to avoid fatigue in objective process docu-
mentation as well as to minimise the influence of monitors’ expectation on the study data (Rosenthal effect). Furthermore, comparisons are made often with studies in adult patients that used similar observation methods and comparable error descriptions [4]. Our intervention programme, in contrast, decreased the prevalence to 63 %, albeit with variable success across dosage forms and predefined subcategories of medication errors. Although IV drug handling is considered particularly errorprone [1, 20, 27], in our PICU medication errors with IV drugs were less frequent than with PO drugs (0.64 vs. 1.33 errors/process). In our PICU, PO drugs were less frequently administered than IV drugs while nurses had less knowledge deficits in PO drugs than in IV drugs. Therefore, we assume that the staff was less familiar with the actual handling or had a reduced awareness of potential errors and potential resulting serious ADEs [7]. We also should bear in mind, that it is easier to reduce excessive error-rates such as the error-rate in PO drugs than comparatively low error-rates in the baseline assessment such as in IV drugs. This is in line with our experience that the 3-step intervention prevented mainly medication errors in PO drug handling while the rates in IV drugs were not influenced to the same extent. Particularly in young and critically ill children, additional handling steps become necessary such as the splitting or crushing of tablets. These steps provide new occasions for errors to emerge and should hence be looked at in detail. This way, nurses’ awareness of potential ADEs resulting from incorrect drug handling also in less frequently administered dosage forms such as PO drugs was likely raised by our intervention. Furthermore, in the interdisciplinary setting of the PICU a lot of different drugs including high-risk drugs were administered [12]. Strategies to improve quality frequently prioritize critical dose drugs, drugs that were administered most frequently or drugs known as error-prone [4]. Our data, however, suggest the necessity also to consider rarely administered as well as non high-risk drug groups such as electrolytes or gastrointestinal drugs, which also bear a certain risk for ADEs when handled incorrectly [28]. For instance, ranitidine may lead to nausea and vomiting when administered undiluted and should be handled with special care in children and newborns [24]. Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
Downloaded by: University of Pittsburgh. Copyrighted material.
drugs n = 1065]
n = 734]
66 Original Article
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
A programme, even addressing multiple factors of error prevention may not lead to its full success if not repeated for each new nurse. In our study, the response rate in the questionnaire survey was low with only 57 %. This indicates that for further steps in our (and other) setting incentives are required for the nurses to take part in such a programme.
Limitations
▼
(i) We focused on medication errors in drug handling without attempting to identify actual clinical ADEs that may result from these errors. (ii) A potential observation bias, possibly induced by the presence of a monitor (Hawthorne effect) may limit our results and may have caused an underestimation of the extent of the problem. (iii) In the observation, we did not include weekend, evening, and night shifts, in which the prevalence of errors may have been different.
Conclusion
▼
Drug handling errors in paediatric intensive care are highly frequent. The implementation of a comprehensive intervention programme decreased the error-prevalence. However, its success varied across dosage forms and subcategories of errors. We propose to monitor drug handling processes and to develop intervention programmes, including potentially a special course aimed at new nurses and continuous training in drug handling processes, as a means to prevent medication errors effectively.
Acknowledgements
▼
We would like to thank all participating nurses and physicians for their kind support. We thank Prof. Dr. Walter E. Haefeli for his helpful remarks in developing the manuscript and Prof. Dr. Thomas Weber for language editing of the manuscript.
Conflict of interest: The authors have no conflict of interest to disclose. References 1 Barber N, Taxis K. Incidence and severity of intravenous drug errors in a German hospital. Eur J Clin Pharmacol 2004; 59: 815–817 2 Bates DW. Preventing medication errors: A summary. Am J Health Syst Pharm 2007; 64 (14 Supplement 9): S3–S9 3 Berdot S, Sabatier B, Gillaizeau F et al. Evaluation of drug administration errors in a teaching hospital. BMC Health Serv Res 2012; 12: 60 4 Bertsche T, Niemann D, Mayer Y et al. Prioritising the prevention of medication handling errors. Pharm World Sci 2008; 30: 907–915 5 Bertsche T, Bertsche A, Krieg EM et al. Prospective pilot intervention study to prevent medication errors in drugs administered to children by mouth or gastric tube: a programme for nurses, physicians and parents. Qual Saf Health Care 2010; 19: e26 (1–5) 6 Bertsche T, Pfaff J, Schiller P et al. Prevention of adverse drug reactions in intensive care patients by personal intervention based on an electronic clinical decision support system. Intensive Care Med 2010; 36: 665–672 7 Cornish P. “Avoid the crush”: hazards of medication administration in patients with dysphagia or a feeding tube. CMAJ 2005; 172: 871–872 8 Dabliz R, Levine S. Medication Safety in Neonates. Am J Perinatol 2011; 29: 49–56 9 Davis L, Ware R, McCann D et al. Evaluation of contextual influences on the medication administration practice of paediatric nurses. J Adv Nurs 2009; 65: 1293–1299
Downloaded by: University of Pittsburgh. Copyrighted material.
In line with previous studies, we confirm in our study that antiinfective agents were administered with a high prevalence and were frequently involved in errors. Antiinfective agents are often in the focus of interventions such as an Antibiotic Stewardship Initiative that aimed to restrict and control the administration of antibiotics to prevent antimicrobial resistance as well as to minimise costs [15]. In these programmes, however, drug handling aspects itself are frequently neglected. The prevention of medication errors varied in their success rate across the predefined subcategories of errors. While the intervention decreased errors in the subcategory “Incorrect IV drug administration” from 31 to 6 %, it failed to improve medication errors of the subcategory “Not panning a 24 h-infusion”. This subcategory of medication errors, however, was mentioned by each intervention strategy but surprisingly the frequency of errors in this subcategory increased after the intervention. The missing error-reduction in this category is in line with literature stressing that nurses are not always aware of the necessity of following rules [9]. Other surprisingly frequent errors, including PO drugs with apparently self-explanatory steps such as shaking of suspensions (68 %) were responsive to the intervention. According to the literature, the health care team appears to lack awareness of the consequences [2, 18] such as under- and overdosage of the drug and changes of the concentration in the bottle, thereby putting further patients at risk for dosing deviations. Furthermore, our intervention was successful in preventing errors in subcategories such as “Incorrect dissolving/suspending of drugs for PO administration” (42 % reduced to 10 %), or “mixing drugs” (14 % reduced to 1 %) after the intervention programme. The error-prevention in these subcategories stands in contrast to previous studies, which often highlighted drugrelated errors such as wrong-time [3, 22, 23], wrong-administration-technique [23], or patient misidentification [8]. We, however, could not confirm such a high prevalence of those errors in the mentioned categories. This study underlines the different involvement of error subcategories in drug handling processes. Furthermore, it identified many ward-specific areas such as preparing 24 h infusion pumps and subcategories of medication errors, which might be helpful in future error prevention strategies. Although the number of affected patients did not decrease significantly, the overall prevalence of medication errors decreased from 83 % to 63 %. One explanation for the only moderate success of the intervention may be the work load [10, 14] and psychological stress of the nurses [19] in an intensive care unit. In comparison to general wards, in the PICU nursing care is more complex and laborious and drug handling, which can take up to 15 min [11], is only a small part of the overall nursing programme. Additionally, in our PICU fluctuations in the nursing team were frequent. Some nurses stayed only for a few months in the unit because they took part in a specialisation programme for paediatric intensive care and normally worked in another hospital. Standards in nursing care and drug handling may vary between hospitals, whereby drug handling processes may be conducted heterogeneously in the PICU. For further studies, we would, therefore, recommend to design a beginners’ programme for each new nurse, e. g. including a questionnaire survey about standard drug handling processes in the PICU, a subsequent training course by a pharmacist, and the issuance of a wardbased reference book. Continuous training in correct drug handling for all nurses may be a further solution to prevent medication errors in drug handling effectively and sustainably.
10 Endacott R. The continuing imperative to measure workload in ICU: impact on patient safety and staff well-being. Intensive Care Med 2012; 38: 1415–1417 11 Elganzouri ES, Standish CA, Androwich I. Medication Administration Time Study (MATS): nursing staff performance of medication administration. J Nurs Adm 2009; 39: 204–210 12 Franke HA, Woods DM, Holl JL. High-alert medications in the pediatric intensive care unit. Pediatr Crit Care Med 2009; 10: 85–90 13 Ghaleb MA, Barber N, Franklin BD et al. The incidence and nature of prescribing and medication administration errors in paediatric inpatients. Arch Dis Child 2010; 95: 113–118 14 Hiemisch A, Kiess W, Brähler E. Mental Job Strain in an University Children’s Hospital – A Study on Stress Experience and the Resulting Employee Health. Klin Padiatr 2011; 223: 236–241 15 Huebner J, Rack-Hoch AI, Pear A et al. Pilot Project of an Pediatric Antibiotic Stewardshio Initiative at the Hauner Children’s Hospital. Klin Padiatr 2013; 225: 223–229 16 Koren G, Haslam RH. Pediatric medication errors: predicting and preventing tenfold disasters. J Clin Pharmacol 1994; 34: 1043–1045 17 McDonnell C, Hum S, Frndova H et al. Pharmacotherapy in pediatric critical illness: a prospective observational study. Paediatr Drugs 2009; 11: 323–331 18 Merry AF, Anderson BJ. Medication errors – new approaches to prevention. Pediatr Anesth 2011; 21: 743–753 19 Pardoe P. Psychological support for nurses on paediatric intensive care units. Nurs Child Young People 2011; 23: 27–29 20 Parshuram CS, To T, Seto W et al. Systematic evaluation of errors occurring during the preparation of intravenous medication. CMAJ 2008; 178: 42–48
21 Prandstretter C, Tamesberger M, Wagner O et al. Medical prescriptions to premature and newborn infants in an Austrian Neonatal Intensive Care Unit. Klin Padiatr 2009; 221: 312–317 22 Prot S. Drug administration errors and their determinants in pediatric in-patients. Int J Qual Health Care 2005; 17: 381–389 23 Schneider MP, Cotting J, Pannatier A. Evaluation of nurses’ errors associated in the preparation and administration of medication in a pediatric intensive care unit. Pharm World Sci 1998; 20: 178–182 24 Terrin G, Passariello A, De Curtis M et al. Ranitidine is Associated With Infections, Necrotizing Enterocolitis, and Fatal Outcome in Newborns. Pediatrics 2011; 129: e40–e45 25 Tzimenatos L, Bond GR. Severe injury or death in young children from therapeutic errors: a summary of 238 cases from the American Association of Poison Control Centers. Clin Toxicol 2009; 47: 348–354 26 Uppal N, Yasseen B, Seto W et al. Drug formulations that require less than 0.1 mL of stock solution to prepare doses for infants and children. CMAJ 2011; 183: E246–E248 27 Westbrook JI, Woods A, Rob MI et al. Association of interruptions with an increased risk and severity of medication administration errors. Arch Intern Med 2010; 170: 683–690 28 Wetherton AR, Corey TS, Buchino JJ et al. Fatal intravenous injection of potassium in hospitalized patients. Am J Forensic Med Pathol 2003; 24: 128–131 29 Wetter J, Gohlke S, Stutte S et al. Pediatric Hyperglycemic Hyperosmolar Coma Diabeticum: Diagnostic Evaluation and Therapeutic Concept. Klin Padiatr 2012; 224: 26–31 30 Williams NT. Medication administration through enteral feeding tubes. Am J Health Syst Pharm 2008; 65: 2347–2357
Niemann D et al. Drug Handling in a … Klin Padiatr 2014; 226: 62–67
Downloaded by: University of Pittsburgh. Copyrighted material.
Original Article 67
