Errors committed by nursing technicians and assistants in administering antibiotics

Errors committed by nursing technicians and assistants in administering antibiotics Heloisa Helena Karnas Hoefel, RN, MSN,a and Liana Lautert, RN, PhDa Rio Grande do Sul, Brazil

Background: Success in any antibiotic treatment depends on many factors, including the associated preparation procedures and the nursing care management. Objective: The aim of this research was to analyze errors in nursing procedures in the use of cefepime. Methods: A survey was performed in the adult patients’ wing of a university hospital in Brazil, and a direct observation technique was used to process variation errors. Errors were classified by type. Results: Thirty-three nursing assistants were observed and interviewed as they prepared and administered cefepime on 99 occasions. From our sample, 20 (20%) applications were performed correctly but 126 errors were observed in 79 cases (80%). The majority (62%) of errors detected were time (scheduled and rate) related. Trained personnel executed the procedure for diluting the solution better than nontrained personnel, but no other statistically significant effect related to the knowledge level of the personnel was detected. This study concluded that errors occurred during care of patients because lack of specific procedures, such as adequate planned rate of infusion and schedule time. They were classified in an American Society of Health-System Pharmacists scheme. It is recommended that corrective practice and safety measures be introduced. (Am J Infect Control 2006;34:437-42.)

The proper administration of antibiotics—that is, the right dose, concentration, and infusion period— depends to a great extent on the skill of the respective nursing team, both in Brazil and in other countries. In fact, therapeutic success and the prevention of adverse effects, as well as the prevention of the development of resistant bacteria, is a function of how well the pharmaceutical orientations are followed.1 On the other hand, excessive concentrations and/or too-rapid infusion may cause local reactions with inflammation, infection, and the need for treatment, or even skin reactions that may cause false interpretations and lead to unnecessary changes in the prescribed antibiotic. In spite of the importance of the nursing team in the administration of antibiotics and the nurses’ responsibility for the control of hospital infections, not many publications focusing on such administration are found in the specialist literature. In a review of the last 6 years of the issues of the Journal of Infusion Nursing, only one publication2 dealt with the specific

From the School of Nursing of the Federal University of the State of Rio Grande do Sul State, Brazil and Hospital de Clinicas (HCPA) of Porto Alegre, Rio Grande do Sul, Brazil.a Reprint requests: Heloisa HK Hoefel, Rua dr. Tauphick Saadi n.260 ap. 403, Porto Alegre 90470040, RS, Brasil. E-mail: [email protected]. 0196-6553/$32.00 Copyright ª 2006 by the Association for Professionals in Infection Control and Epidemiology, Inc. doi:10.1016/j.ajic.2005.08.011

subject of the infusion of antibiotic. In Brazil, specific nursing research on antibiotics is equally rare.3 Research about safe medication is conducted, but publications in this area from a nursing perspective are scant as well,4-6 and many of the published papers are revision articles, such as the one of the American Society of Health-System Pharmacists (ASHP)7 and the paper with conclusions of the Nursing Society’s 2004 symposium in Philadelphia about research and safe medication.8 Errors in dose, concentration, infusion period, contamination risk, and other areas are related to different steps in drugs administration and may be associated to where and by whom drugs were prepared and administered. Purchasing ready-prepared intravenous drugs from pharmaceutical companies and training nurses to prepare medications are suggested by Taxis and Barber5,6 with regard to administering drugs that require multiple preparation steps. Flynn et al9 did not find a difference when drugs were prepared by a pharmacy or by nurses. Others, such as Bates et al,10 verified a 10% difference between drugs mixed by nurses and by pharmacists. Use of smart pumps to control specific flow is an alternative that is most often used when greater accuracy or higher flows are needed than what can be provided by manually adjusted gravity systems. Rosenthal11 discussed the benefits of smart pumps and bar codes. Hush et al12 concluded in their research that errors associated with smart pumps are frequent. They emphasized the necessity of interfacing smart pumps with other systems (such as bar codes and others), 437

438

Hoefel and Lautert

Vol. 34 No. 7

recognizing it as a necessary component of a safe medication system. Furthermore, during activities with students of the Nursing School at the HCPA (Hospital de Clı´nicas of Porto Alegre, Brazil), it was noted that the nursing professionals used different methods to prepare and administer the same antibiotic. In view of this fact, it is necessary to study the various techniques of antibiotic administration to ensure that the procedures are being executed in a manner most likely to lead to therapeutic success.

OBJECTIVES The objective of this study was to identify the kinds of errors committed by nursing assistants and technicians in administering antibiotics, and why they occurred.

MATERIALS AND METHODS An observational study and descriptive analysis of the data was performed in a 725-bed, privately owned, public university hospital serving a local community. Assistants and nursing technicians caring for adults from the intensive care unit, the surgery inpatient unit, and the clinical inpatient unit being given cefepime were included in the research. The administration of the antibiotic cefepime, a fourth-generation cephalosporin, was chosen because it is widely used in HCPA according to the hospital’s own pharmaceutical service. Cefepime is used in the treatment of bacterial producers of extended-spectrum beta-lactamases. To perform the sample calculation, we started with the research method developed by Hoefel et al3 with vancomycin, realizing an initial statistical analysis on a sample of 51 procedures executed by 17 professionals. An error rate of 10% was detected related to the dose, infusion rate, and concentration. Later, a new analysis was made of the first 25 professionals observed (75 procedures) to confirm the variability of the data, and the error percentage found was also 10%. Based on these results the size of the samples was calculated with an estimated error of 10% and with a 95% confidence interval. Starting with this calculation and completing the series of 3 observations per professional, the sample was constituted from 99 observations of procedures performed by 33 professionals. The researcher carried out 3 observations per participant. Five observers were trained by the researcher seeing and doing different steps of the process of administration, from the initial dilution of cefepime to infusion. Observations were registered in a structured form with the following information: doses of cefepime, solution to be diluted, volume, time beginning

and end, drops per minute, residues of solution in infusion set, whether residues were discarded or infused, and the method of discarding these residues. The subjects invited to participate in the study signed a document confirming their free and informed consent, and this document was approved by the hospital’s Ethics in Research Committee. The data were collected in two stages: (1) a semistructured interview to characterize each professional with information on his or her nursing experience both in the HCPA and elsewhere, and the degree of specific training in the administration of antibiotics; and (2) direct observation of the process of administration of the medication from the initial preparation to infusion. Both the questions asked during the interviews and the details observed during the preparation and administration of the medication were elaborated with assistance from the HCPA’s pharmaceutical service. There were 5 types of performance errors from the classification system of the ASHP13 used as the standard for classifying the kinds of errors, shown in Table 1. The statistical significance was established using the Fisher exact test. The x2 test was used to measure the tendency to follow scheduled time.

RESULTS In the period between April and November of 2003, 99 observations were made of 33 nursing professionals. After 3 observations, each professional was interviewed. The group professional experience extended from 12 months (1 year) to 253 months (21 years), with an average of 135.8 months (11 years), and the time working at the HCPA ranged from 12 to 168 months (1 to 14 years), with an average of 47.6 months (4 years). Twenty-seven professionals received specific hospital training on administering medications, and 6 did not. Five types of errors were detected in the administration of the antibiotic and were categorized in accordance with the ASHP system13 and the ASHP’s modified system14 based on a compilation of literature, as described above. Of the 99 cefepime administrations observed, 20 (20%) were correctly administered and 79 (80%) were incorrectly administered. There were observed combined errors and errors observed alone in cefepime administration (Table 2). In 43 cases (43%), only 1 error was detected; in 30 (30%), 2 types of errors were detected; and in 7 (7%), 3 types of errors were combined. Errors detected in this research on intravenous administration of cefepime can be difficult to compare with those in existing literature because other

Hoefel and Lautert

September 2006

439

Table 1. Classification of type of errors according to how they were identified

Table 2. Combined errors and errors observed alone in cefepime administration

Wrong dose error (WD): any dose that is the wrong number of preformed units or any dose above or bellow the predetermined amount. It was identified by the residue left in the equipment when administration was complete. Wrong preparation of a dose (WP): incorrect preparation of medication dose. Examples are incorrect dilution or reconstitution. In this case, the dilution volume was less than 50 mL, resulting in a final concentration greater than the 40 mg/mL that is recommended for intravenous administration cefepime hospital policy. Wrong administration technique (WAT): situations in which the drug is given via the correct route, site, and so forth, but improper technique is used. Errors of technique causing risk of contamination by incorrect manipulation (WATC) also occurred in situations in which remains of previous procedures still existed in the equipment. They were disposed of in the waste receptacles, in a paper cup, or in a nonsterile kidney dish. The correct procedure is to insert the needlepoint of the equipment into the empty serum bag to dispose of the residue without risk of contamination. Unknown infusions prepared by others and which had remained in the equipment and administered were also included in this type of error (WATU). In accordance with Brazilian nursing practices rules, medicine must be administrated by the person by whom it was diluted and prepared, except when dispensed ready for use from the pharmacy. Wrong rate error (WR): administration of a drug at the wrong rate, the correct rate being that given in the physician’s order or as established by hospital policy. These errors involved both times being too short (less than 3 minutes), causing early infusion termination (SWR), and longer (LWR) than recommended (30 minutes). Wrong scheduled time error (WT): administration of a dose of a drug greater than ‘‘plus or minus’’ X hours from its scheduled administration time, X being that set by hospital policy (WTP) (WTM). Time was determined by prescription.

Error type

n

%

WT WR WT 1 WR WATU WT 1 WATC WT 1 WD WR 1 WATU WD WT 1 WATU WT 1 WR 1 WATU WD 1 WR WP WATC WD 1 WP WATU 1 WD WT 1 WD 1 WP WR 1 WD 1 WP WT 1 WR 1 WD WT 1 WR 1 WP WT 1 WR 1 WATC Total

14 13 09 08 06 05 04 03 03 03 02 01 01 01 01 01 01 01 01 01 79

18 17 12 10 08 06 05 04 04 04 03 01 01 01 01 01 01 01 01 01 100

investigators have studied many vials and/or many drugs together. In a recent literature review carried out by Crowley et al,15 the investigators found overall preparation and administration of drugs errors ranging from 23.8% to 93.9%. Dean and Barber16 identified errors in 6067 opportunities for errors in 176 rounds: 95 unavailable medication-related and 162 other types. The 162 errors of other types represented 92% of occurrences in 176 observed administrations of various medications. However, in a study by Greengold,17 when nurses administered medications, the error rate was 14.9% and 15.7%. Flynn et al9 identified 9% of errors in intravenous administration drugs in 5 hospitals representing 5 regions in the United States. They studied manual and partly automated intravenous drug preparation by nurses and pharmacists with no difference among professionals. The number of procedures and the number of professionals who administered the cefepime were analyzed separately. The numbers of administrations shown in Table 3 may be greater than the corresponding multiples of 3 because only those professionals

Data from Hoefel H. Data collection by observation and interview. Porto Alegre, Brazil; 2004. WATC, wrong administration technique—contamination risk; WATU, wrong administration technique—unknown contents; WD, wrong dose; WP, wrong preparation; WR, wrong rate infusion; WT, wrong scheduled time.

who correctly accomplished 2 or 3 administrations of the 3 observed were considered. Training participation was not statistically significant in relation to the general successes/errors of the nursing professionals because 20 of them who had undergone training committed one or more errors in administering the antibiotic (Table 3). It is possible that the emphasis on aspects related to observed errors did not attract the attention of all of the participants in the trainings sessions. Also it should be noted that the average experience of those not trained was 11.6 years, or almost the same as the 11.2 years of the professionals who received the training. The average experience in HCPA is 5 years for the nontrained and 4 years for the trained. These similarities may help to explain why the results obtained do not show statistically significant differences between the two groups regarding the success/error rates in administering the antibiotic. In the investigation by Greengold17 with two groups of nurses, one group was made up of professionals who were responsible only for the administration of medication to 18 patients each and who had been trained in accordance with the institution’s specific methodology for the administration of medication. The other group was responsible for 6 patients each and had not been specifically trained. The conclusion is that

440

Hoefel and Lautert

Vol. 34 No. 7

Table 3. Professionals performing correctly and not at least 2 administrations of cefepime and of the correct and incorrect observations in accordance with the training received Professionals Correct

Administrations

Incorrect

Total

Correct

Incorrect

Total

Observations Training

n

%

n

%

n

%

n

%

n

%

n

%

Yes No

07 01

26 17

20 05

74 83

27 06

100 100

17 03

21 17

64 15

79 83

81 18

100 100

Fisher exact test: professionals, P 5 1; administrations, P 5 1. Data from Hoefel H. Data collection by observation and interview. Porto Alegre, Brazil; 2004.

Table 4. Observed errors in cefepime administration by type Type of error Wrong rate, long Wrong time, plus Administration technique, unknown contents Wrong dose Wrong time, minus Administration technique, contamination risk Wrong preparation Wrong rate, short Total

n

%

32 32 19 14 12 08 06 03 126

25 25 16 11 10 06 05 02 100

Data from Hoefel H. Data collection by observation and interview. Porto Alegre, Brazil; 2004.

training did not influence the observed error rate. The researcher concluded that although in some areas the use of specific personnel to administer medications may be useful, in general the errors are more related to the logistics of administration than to a lack of orientation or specific training. In the present investigation, 126 errors were observed in 79 administrations. Each type of observed error is described in Table 4. Of the wrong rate (WR) errors, 32 (25%) were longer (WRL) than 40 minutes (allowing 10 minutes of tolerance on the predetermined time of 30 minutes) and 3 (2%) were too short (WRS). The tolerance period of 10 minutes over the specified time resulted from observations of the practice of counting the drops falling initially in one quarter of 1 minute and then multiplying by four. Time established by hospital policy considered correct is 30 minutes when done by infusion set. It was considered incorrect to administrate the cefepime in less than 3 minutes because direct infusion is not specified by the pharmaceutical service of the HCPA, and also considering the research Garrelts and Wagner.18 It was also stipulated to include 10 minutes of tolerance when judging scheduled start time. There were 46 observed administrations considered as occurring at the correct scheduled time: 23 (50%) 10 minutes later and 23 (50%) 10 minutes earlier (Table 5). The 44

wrong time (WT) errors were: 32 (73%) more than 10 minutes (minus) earlier (WTM) and 12 (27%) more than 10 minutes (plus) later (WTP), shown in Table 4. Expecting that in 50% of cases the time of the beginning of administration was later than scheduled and in 50% the time was early, a tendency to begin earlier can be verified (x2: P 5 .003 and degrees of freedom 5 1). In 7 (22%) cases of the 32 WTM errors, the observed professionals were caring for many other patients who had intravenous drugs prescriptions at the same time. There were 3 cases observed beginning 1 hour before the scheduled time. On the other hand, when comparing wrong preparation (WP) errors, 4 (5%) trained professionals compared with 2 (11%) nontrained professionals (P 5 .3) diluted cefepime in a wrong way. If analyzing the performance of 33 observed persons, trained people did not show errors, ie, 27 (100%) diluted cefepime correctly. Of the 6 nontrained technicians, 2 (11%) made at least one mistake diluting cefepime and 4 (89%) performed dilution correctly. Although significant (P 5 .02), this analysis is limited because the nontrained sample is too small. Even if differences between prepared doses by trained and by nontrained professionals were not statistically significant (P 5 1), 15 incomplete doses in both groups represents 15% of all studied administrations (Table 6) and raises concern about therapy failure.

CONCLUSIONS AND RECOMMENDATIONS The types of errors most frequently identified in the administration of the antibiotic studied, in accordance with the ASHP13 classification, were related to time with 62% of the errors identified. These errors were distributed between early or late starting times (35%) and errors in the infusion of the medication (27%). The cited events are related specifically to the degree of nursing care exercised. Depending on the drug, these occurrences can lead to an alteration of the pharmacokinetic curve and the absorption of the medication, leading to an alteration of the desired effect.

Hoefel and Lautert

September 2006

Table 5. Correct and wrong administration, 10 minutes plus or minus, in accordance with late and early cefepime time set Correct

Wrong

Until 10 min

More than 10 min

Time set

n

%

n

%

Late Early Total

23 23 46

50 50 100

12 32 44

27 73 100

x2 test: P 5 0.003 e gl 5 1. Data from Hoefel H. Data collection by observation and interview. Porto Alegre, Brazil; 2004.

Preparations of medication-type errors related to the volumes of infusion with more concentrated dose were observed in 5% of the nurses’ antibiotic preparation procedures. Furthermore, interference is greater when these kinds of errors are associated with wrong dose errors, which occurred in 11% of the observations (when residues in the equipment were not administered), leading to administration of a shorter dose. Independent of the number of occasions when incomplete doses were administered, they may have been individually subinhibitory, presented below the minimum inhibitory concentration (MIC), thus increasing the selective pressure of the microorganisms.19 Therefore, they may be responsible for alterations in the sensitivity of the microorganisms to the therapy and for problems related to that alteration. Because the study evaluated only whether the recommendations had been fulfilled or not, it cannot be stated that less than the inhibitory dose was administered. However, the unadministered residues suggest that doses were smaller than necessary. The therapeutic fault, which is one of the possible consequences (level 4 according to the ASHP), in this case without a protocol or vigilance program, is difficult to associate with the respective error. This latter aspect is not easy to detect or to associate unless the individual serum levels are analyzed, which is not the normal practice for cefepime. There were many studies regarding errors that had different findings. Flynn et al9 studied 5 hospitals and found a 9% compounding error rate in different kinds of medications prepared by nurses and pharmacy. The study by Bates et al10 shows that mixing errors by the nursing staff are 20%, and by the pharmacy are 10%. Our study shows that many kinds of errors occurred even when nursing professionals attended training courses offered by the institution. Although the behavior of health professionals are not related to the degree of training offered, we expected that trained personnel would have better performance than nontrained personnel. This fact suggests that although trained persons

441

Table 6. Complete and incomplete doses administered in accordance with the training received by professionals Complete doses

Incomplete doses

Total

Administration training

n

%

n

%

n

%

Yes No

69 15

87 83

12 03

13 17

81 18

100 100

Fisher exact test: P 5 1. Data from Hoefel H. Data collection by observation and interview. Porto Alegre, Brazil; 2004.

diluted the solution better than nontrained personnel, the existing training system should be modified to obtain better results focusing on all other issues. Administering the unknown content remaining in the infusion set, or the probability of contamination, represented 22% of the errors in the wrong administration technique category. These aspects can be retrospectively revised to identify the cause of a negative consequence. However, the establishment of a protocol and vigilance system can prevent these events. Taxis and Barber6 studied causes that might explain our findings of wrong infusion rate or wrong preparation. Error-producing conditions stated by them, such as handling and design technology, may affect infusion rate, for example. Communication problems and workload also could be related to human error. This study can be expanded in a way to study these associated possibilities with a view to preventing errors. The consequences of the errors identified in this study can have repercussions, principally epidemiological, related to bacterial resistance. To identify consequences of errors as described in ASHP13 of levels other than level 0 (no error or potential error occurred), it would be necessary to extend the scope of this study by following up the cases after the administration of the cefepime. It might be necessary to alter the antibiotic therapy of patients who had been given incomplete doses that seemed like a failure of the medication but there were not associated with incomplete doses administered or other errors that may influence pharmacodynamics curves. Considering the probability that the hospital bacterial flora may also be influenced by repeated errors in the timing of administrations, or by the application of subinhibitory dosages, independently of the individual consequences, it is suggested that the ASHP13 level 0 be subdivided into two groups: (1) those errors without identified consequences for the individual patient, and (2) potential errors. In this way, those errors that might have epidemiological consequences would be classified as potential errors and studied separately, thus attributing greater importance to this parameter. In this matter, pharmacy surveillance is a way of

442

Hoefel and Lautert

Vol. 34 No. 7

detecting the errors and identifying shortcomings in administration. For the above reasons, various measures could be taken to prevent errors. Standardization of the process of administration of the drugs, specifying in what situations and in what manner any remains left in the equipment should be handled, should be a priority, whereas the technique, materials, and equipment used should be part of the responsibilities of the nursing staff. These are aspects of the process that should be clarified. It is equally important that these steps should be regularly supervised and recorded. A follow-up plan structured for predetermined days could be of great assistance in the identification of potential errors. This study calls for further action, such as the elaboration of projects to control the processes of administration of the different classes of medication, which would be useful in the development of research in nursing affairs and also would help to identify the best way to provide safe and ethical care while respecting the patient’s dignity. Limitations of this research are the small group of nontrained professionals and the possibility of the Hawthorne effect when observers were collecting data. The effect of the observer on the observed might be a possible limitation of this observational study. However, a previous observation-based study16 showed that the drug error rate was unaffected by the observer. On the other hand, a specific study about nursing supervision to minimize preventable errors in nursing care and about why errors occurred is needed. Although the results about training in this study cannot be used to generalize knowledge, we would like to emphasize the importance of education regarding drugs administration in general. Lack of education cannot only be fulfilled with information, termed by Freire20 as ‘‘bank education,’’ in which the teacher’s primary role is to ‘‘deposit’’ information in students’ minds as they would make a money deposit into a bank. Education requires internal modifications. Training in competencies may be part of the educational program. Although competencies is a contested concept in Freirean education, if one has a body of practical knowledge, one can discuss, reject, modify, or approve for inclusion some information and built new information and knowledge to follow and enrich the educational process in this matter. It is necessarily a mutual process of reflecting on and developing insights into the person’s evolving culture. Professionals are anxious to succeed when caring for patients. Anyone who understands, for example,

that a situation of physical pain can be alleviated will be moved to the respective corrective action. If this does not happen, it must be because that person has not understood the situation correctly, either for moral reasons or from a lack of knowledge. References 1. American Society of Health-System Pharmacists. Pharmacy—nursing shared vision for safe medication use in hospitals: executive session summary. Am J Health Syst Pharm 2003;60:1046-51. 2. Hammond D. Home intravenous antibiotics: the safety factor. J Infect Nurs 1998;21:81. 3. Hoefel H, Zini L, Lunardi T, Santos J, Mahmud S, Magalha˜es A. Vancomycin administration in an university hospital at general surgical units inpatients. OBJN [online]. 2004;3. Available from: www.uff.br/ nepae/objn301hoefeletal.htm. Accessed January 20, 2006. 4. Wirtz V, Taxis K, Barber ND. An observational study of intravenous medication errors in the United Kingdom and in Germany. Pharm World Sci 2003;25:104-11. 5. Taxis K, Barber N. Ethnographic study of incidence and severity of intravenous drug errors. BMJ 2003;326:684-7. 6. Taxis K, Barber N. Causes of intravenous medication errors: an ethnographic study. Qual Saf Health Care 2003;12:343-8. 7. American Society of Health-System Pharmacists. Pharmacy—nursing shared vision for safe medication use in hospitals: executive session summary. Am J Health Syst Pharm 2003;60:1046-51. 8. Burke K. The State of the Science on Safe Medication Administration symposium. J Infus Nurs 2005;28:87-92. 9. Flynn E, Pearson R, Barker K. Observational study of accuracy in compounding i.v. admixtures at five hospitals. Am J Health Syst Pharm 1997;54:904-12. 10. Bates DW, Cullen DJ, Laird N. Incidence of adverse drug events and potential adverse drug events. JAMA 1995;274:29-34. 11. Rosenthal K. Smart pumps help crack the safety code. Nurs Manag 2004;35:49-51. 12. Husch M, Sullivan C, Rooney D, Barnard C, Fotis M, Clarke J, et al. Insights from the sharp end of intravenous medication errors: implications for infusion pump technology. Qual Saf Health Care 2005;14:80-6. 13. American Society of Health-System Pharmacists. ASHP technical assistance bulletin on hospital drug distribution and control. Am J Hosp Pharm 1980;37:1097-103. 14. American Society of Health-System Pharmacists. ASHP guidelines on preventing medication errors in hospitals. Am J Hosp Pharm 1993; 50:305-14. 15. Crowley C, Scott D, Duggan C, Whittlesea C. Describing the frequency of IV medication preparation and administration errors. Hosp Pharm 2004;11:330-6. 16. Dean B, Barber N. Validity and reliability of observational methods for studying medication administration errors. Am J Health Syst Pharm 2001;58:54-9. 17. Greengold N. The impact of dedicated medication nurses on the medication administration error rate. Arch Intern Med 2003;163:2359-67. 18. Garrelts JC, Wagner DJ. The pharmacokinetics, safety and tolerance of cefepime administered as an intravenous bolus or as a rapid infusion. Ann Pharmacother 1999;33:1258-76. 19. Schrag SJ, Beall B, Dowell SF. Limiting the spread of resistant pneumococci: biological and epidemiological evidence for the effectiveness of alternative interventions. Clin Microbiol Rev 2000;13:588-601. 20. Freire P. Pedagogia do oprimido. 17 ed. Rio de Janeiro: Paz e Terra, 1987.