Causes of intravenous medication errors—observation of nurses in a German hospital

Aim Errors in the preparation and administration of intravenous (IV) drugs are frequent events. Human error theory has recently been applied to understand the causes of IV drug errors in an ethnographic study in the United Kingdom. We used this

Please download to get full document.

View again

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 2 | Pages: 7

Extension: PDF | Download: 0

  J Public Health (2004) 12:132–138DOI 10.1007/s10389-004-0019-4 ORIGINAL ARTICLE Katja Taxis · Nick Barber Causes of intravenous medication errors—observation of nursesin a German hospital Received: 16 June 2003 / Accepted: 13 October 2003 / Published online: 6 March 2004 Springer-Verlag 2004 Abstract  Aim:  Errors in the preparation and administra-tion of intravenous (IV) drugs are frequent events. Humanerror theory has recently been applied to understand thecauses of IV drug errors in an ethnographic study in theUnited Kingdom. We used this approach to explorecauses of IV drug errors by nursing staff in a Germanhospital.  Methods:  A trained and experienced observeraccompanied nurses during IV drug rounds on two wardsin one German hospital. Information came from observa-tion and talking informally to staff. Human error theorywas used to analyse causes of IV errors.  Results:  Twenty-two nurses were observed for 13 study days. A total of 74 IV drug errors were identified while observing 161preparations and 135 administrations. Mistakes werefrequently related to the selection of the wrong fluid fordrug preparation and the co-administration of potential-ly incompatible infusions. Lack of training in IV drugpreparation and administration was the main problem.Furthermore, guidelines were ambiguous and did notcontain sufficient information. The transcription of drugorders also contributed to drug errors.  Conclusion:  Theintroduction of nurse training and guidelines provided bya multidisciplinary team, including a clinical pharmacist,may reduce the high rate of IV drug errors. Thesemeasures should be linked to a review of the legalframework to recognise IV drug administration as anursing task. Keywords  Medication errors · Human error theory ·Intravenous therapy Introduction Medication errors related to the preparation and admin-istration of drugs are common events in hospitals (Barberand Dean 1998). In observations of the administration of oral doses on four wards in two German hospitals, up to5% of errors were identified (Taxis et al. 1999). Highererror rates were associated with the preparation andadministration of intravenous drugs. We identified anerror rate of 48% observing the preparation and admin-istration of 122 doses on two wards in one German non-university hospital (Taxis and Barber 2004). This includ-ed four potentially severe errors, such as the co-admin-istration of several incompatible drugs through the sameadministration line or the fast injection of an opioidanalgesic. Two other studies carried out in Germanuniversity teaching hospitals also reported a high inci-dence of IV drug errors (Wirtz et al. 2003; Hoppe-Tichyet al. 2002). Table 1 summarises the main results of thesestudies. Similar IV drug error rates have been found inUK hospitals (Taxis and Barber 2003a; Wirtz et al. 2003;O’Hare et al. 1995).Investigating the causes of the errors is the first steptowards error prevention (Leape 1994). Human errortheory is increasingly used as a theoretical base to identifyfactors contributing to errors in medicine (Leape et al.1995; Stanhope et al. 1997; Vincent et al. 2000; Dean et al. 2002). The analysis of large-scale accidents in high-risk industries found that the design of systems, pre-existing organisational factors and the conditions, con-ventions and procedures for the use of technology cancontribute significantly to disasters. Based on suchresearch, Reason developed the model of organisationalaccident causation (Reason 2001). He distinguishes be- K. Taxis · N. BarberDepartment of Practice and Policy,The School of Pharmacy, University of London,29–39 Brunswick Square, London, WC1N 1AX, UK K. TaxisPharmazeutische Biologie, Pharmazeutisches Institut,Universitt Tbingen, Auf der Morgenstelle 8,72076 Tbingen, GermanyK. Taxis ( ) )Rijksuniversiteit Groningen,Sociale Farmacie en Farmacotherapie,Antonius Deusinglaan 1, 9713 AV Groningen,The Netherlandse-mail: katja.taxis@farm.rug.nlTel.: +31-50-3637576Fax: +31-50-3632772  tween active failures and latent failures (or latent condi-tions). Active failures are actions at the ‘sharp end’, forexample the administration of the wrong drug to thepatient. These are the final triggers of an accident. Theconsequences of an active failure are immediately obvi-ous. Latent failures are the management decisions andorganisational processes taken at a higher level whichcreate working conditions promoting the occurrence of errors and accidents at ward level. For example decisionsconcerning planning, policy making, regulating andmaintaining may create error-producing conditions suchas understaffing, high workload, poor human equipmentinterface, etc. Adverse effects due to latent failures maytake a long time to become evident. Figure 1 shows theanatomy of an organisational accident. The direction of causality is from left to right. The srcin of an accident isthe negative consequences of organisational processeswhich transmit through the system.Based on Reason’s model, Vincent et al. (2000;Vincent 2003) developed a broad framework of contrib-utory factors that can affect clinical practice. This hasbeen applied successfully in various studies (Stanhope etal. 1997; Dean et al. 2002). We have recently used thisapproach to investigate the causes of IV drug errors in twohospitals in the United Kingdom (UK) (Taxis and Barber2003b). We analysed working conditions (error andviolation-producing conditions) at the time of a drugerror (active failure) to identify organisational processesand management decisions which contributed to the error.A common type of error was the deliberate violation of guidelines when injecting bolus doses considerably fasterthan recommended. A perceived lack of risk and poor rolemodels created an atmosphere where such unsafe drug usewas commonly accepted. Lack of knowledge of uncom-mon procedures and complex design of equipment werefactors contributing to preparation errors. Latent failureswere the failure to teach practical aspects of drughandling and design failures. Although several studiesindicate a high incidence of IV drug errors in Germanhospitals (Wirtz et al. 2003; Hoppe-Tichy et al. 2002), thecauses of these errors have not been systematicallyinvestigated. Consequently we do not know whether thesolutions proposed for the UK would be effective inGermany. Therefore, we conducted an observation-basedstudy to investigate IV drug errors in a German hospital.We have reported a detailed analysis of the incidence,types and clinical importance of the IV drug errorselsewhere (Taxis and Barber 2004). Table 1 summarisesthe main results. In the present paper, we report theidentification of causes of the IV drug preparation andadministration errors, using a framework of human errortheory. Methods SettingData were collected on a surgical ward and an intensivecare unit in a German non-university, teaching hospital.On the surgical ward, doctors recorded prescriptions in Table 1  Summary of results of German studies on the incidence, types and clinical importance of intravenous drug preparation andadministration errorsTaxis/Barber 2004 Wirtz et al. 2003 Hoppe-Tichy et al. 2002Incidence of errorsOne or more errors occurred in the preparationand/or administration of 58 out of 122 intra-venous drug doses (error rate 48%, 95% CI39%–57%); 65 errors were identified in total88 errors in the preparation of 337 doses(26% preparation error rate); 93 errors in theadministration of 278 doses (34% adminis-tration error rate) (data from two German andone UK hospital)No overall error rate calculated(see below for error rates)Potential clinical importanceErrors were potentially severe in four doses(3%), potentially moderate in 38 (31%), andpotentially minor in 16 (13%)77% of errors were judged to be of moderateor severe clinical importanceNo data providedCommon types of errorsDoses involving multiple-step preparations(wrong solvent/diluent, wrong dose, omission)and the co-administration of potentially in-compatible drugs as intermittent infusionsPreparation of the wrong dose, omission of prescribed medication, wrong administrationrate, co-administration of potentially incom-patible medicationPreparation using the wrongsolvent/diluent (50%), wrongdosage form (1.5%), wrong dose(1.5%), wrong administration rate(21%; out of 425 observed IVdoses) Fig. 1  Reason’s four-stage model of human error theory. (Adaptedfrom Reason 2001)133  the inpatients’ notes and these were transcribed bynursing staff onto formatted drug charts. One or twonurses prepared and administered IV medication for allthe patients on the ward. In the intensive care unit, drugprescriptions were entered into electronic patient medi-cation records which did not require transcriptions. Onenurse cared for one to three patients. Nurses prepared andadministered IV drugs for their patients. In the intensivecare unit, a chart was available outlining common in-compatibilities and giving recommendations on the co-administration of intravenous drugs. No such guidelineswere available on the surgical ward.Data collectionParts of the data collection methods and definitions werepreviously reported in detail (Taxis and Barber 2003a,2003b). Briefly, we studied the preparation and admin-istration of IV medications that were administered topatients by nursing staff. Intravenous drug errors weredefined as deviations in this process from a doctor’sprescription, the hospital’s intravenous policy or themanufacturer’s instructions. Reason’s four-stage model of human error theory (Reason 2001) and the framework of Vincent et al. (2000) formed the basis for data collectionand analysis (Fig. 1). We adapted Vincent et al.’s pro-tocols, originally developed to analyse adverse eventsusing interview techniques (Vincent et al. 2000), toobservation-based research (Taxis and Barber 2003b).One of us (KT), a pharmacist who was trained andexperienced in observation-based medication error re-search, accompanied nurses during IV drug rounds. Withthe agreement of senior nurses in the hospital, we pre-sented the study to staff at ward level as a research projectinvestigating common problems of IV drug preparationand administration. This disguised observation-basedmethod has been shown to be a valid way to identifymedication errors (Dean and Barber 2001). The observerrecorded details of each IV drug preparation and admin-istration on a standard form. Additional information camefrom observation and talking informally to staff. Obser-vations and conversations were guided to record infor-mation on the chain of events that led to the error andactions of those involved. The researcher’s records werechecked and completed for each IV drug within 24 h of leaving the ward. Data were collected on six (intensivecare unit) or seven (surgical ward) consecutive days oneach ward, including weekends, in March 2000. Betweentwo and three drug rounds were attended each day.Notes of observation and conversation for each errorwere transcribed and translated by K. Taxis (KT) and readby both authors. Data were coded by KT and coding waschecked by N. Barber. Coding was based on existingcategories (Vincent et al. 2000), which were adapted tothe analysis of IV drug preparation and administrationerrors in a previous study (Taxis and Barber 2003b).Disagreements were discussed and resolved. Each case of IV medication error was analysed to identify the mainactive failure and factors contributing to this error. Activefailures were categorised as human errors (slips/lapsesand mistakes) and violations:– Slips or lapses were failures in the process of executinga task. The observed health care professional had anadequate plan, but the action did not proceed asintended due to recognition, attention, memory orselection failures.– Mistakes were failures at the planning or problem-solving stage of a task.– Violations were deliberate deviations from safe oper-ating practices, recommendations or guidelines, butthere was no indication that any adverse consequenceswere intended.Error and violation-producing conditions were factorsat ward level which lead to active failures. These includedproblems in handling technology, for example lack of knowledge in drug preparation; design failures; commu-nication problems; or high workload. Latent conditionsincluded any underlying organisational and managementfailures which contributed to error and violation produc-ing conditions.The study was approved by the medical, nursing andpharmacy directorate as well as the hospital’s qualityassurance committee. Each nurse was asked for permis-sion for observation. Results We observed 22 nurses during the 13-day study period.One nurse refused to participate. The following two typesof drug preparations were mainly observed:– Drugs were diluted by drawing up the drug solutionfrom the vial into a syringe and injecting it into aninfusion container (usually 50 ml or 100 ml of waterfor injection or 0.9% saline).– Drugs were reconstituted by connecting the vialcontaining the drug with an infusion container (usually50 ml or 100 ml of water for injection or 0.9% saline)via transfer needle, transferring the solvent from theinfusion container into the drug vial and waiting untilthe drug had been dissolved.Both types of drug preparations were administered asintermittent infusions over about 20 min. On the intensivecare unit, the injection of drug solutions as a bolus wasalso observed.A total of 74 IV medication errors were identifiedobserving 161 preparations and 135 administrations. Adetailed analysis of the incidence, types and clinicalimportance of these errors has been reported elsewhere(Taxis and Barber 2004; Table 1). 134  Mistakes, slips, lapses and violationsThe main active failure was identified for 71 (96%) of allIV drug errors. There were 65 (88%) mistakes, 5 (7%)slips and lapses and 1 (1%) violation. Mistakes, thefailures at the planning stage, occurred for example whena nurse selected the wrong fluid to dilute medication. Shewas unaware that the drug could precipitate if the wrongdiluent was used. She explained how she decided whichfluid to use: “I’ll ask the doctor [...] if I cannot findanything, I just give it in 100 ml of saline, that is going tobe all right” (K, 61; observer’s notes). Similar mistakeswere observed during co-administration of potentiallyincompatible drug infusions. Slips included the failure tonotice that a drug had not dissolved completely or thefailure to notice that a drug had been prescribed. Aviolation was observed once. A nurse delayed theadministration of a drug beyond the time the drug wasspecified to be stable. Initially, the patient was unwillingto have the drug infused as he wanted to have his shower.Therefore, the nurse delayed drug administration untilafter the hand-over, although she knew that the drug wasrecommended to be given within a short time afterpreparation.Error-producing conditionsTable 2 shows the type and frequency of error-producingconditions which contributed to errors. Handling technol-ogy and technology itself were the most commonlyidentified problems. Lack of knowledge of IV drugpreparation and administration procedures was the mostcommon factor in the category of handling technology.The problems of selecting the correct fluid for drugpreparations have been outlined in the preceding section.A similar lack of knowledge was observed concerningcorrect administration procedures. Nurses often adminis-tered several potentially incompatible infusions at thesame time through the same administration line. Conver-sations with nurses suggested that they were unaware thatcomplications could arise from co-administration of druginfusions. A nurse explained: “No, there is no problemwith propofol going through the same line as othermedications” (L, 24; observer’s notes). This patientreceived a continuous infusion of propofol, piritramideand midazolam as well as an intermittent infusion of mezlocillin through one lumen of a central line, althoughpropofol should always be administered through a sepa-rate line. Possible complications include complexation,precipitation or degradation of the drugs, resulting in areduced therapeutic effect; the formation of particles alsocarries the risk of embolism. A typical comment of anurse on the surgical ward was: “[...] we only give oneinfusion at a time through a peripheral line. Otherwise theflow rate is too slow, but centrally the co-administration isnot a problem” (K, 28; observer’s notes). The observerasked another nurse: “Do you sometimes have a problemwith infusing several drugs at the same time through thesame line?” The nurse answered: “No, I don’t know thatthere is a problem” (K, 26; observer’s notes).The main problems with technology involved themanufacturer’s leaflets and the guidelines on drug ad-ministration. Nurses on both wards complained about alack of guidelines and information supporting their drug-related work. They said they used the manufacturer’sleaflets, which were usually supplied with the drugs, butthese contained only basic information on drug prepara-tions, but no information on possible drug mixtures. Anurse from the surgical ward explained: “Well, we don’tget any information, we don’t get to know about anything,don’t know about new information on mixing things” (L,14, observer’s notes). There were guidelines on co-administration of IV drugs available in the intensive careunit, but parts of the guidelines were difficult to apply inpractice. For example, the guidelines contained a list of drug infusions which should be administered throughseparate lines. They also contained a list of drugs whichshould be administered through a filter. Most patients inthe intensive care unit had one catheter with two lineswithout a filter and one line with a filter. Frequently, thepatients were prescribed several drug infusions whichshould have been administered through separate lines, butthese were also recommended to be given through a filter.The nurses in general administered such drug infusionsthrough the line with the filter, which resulted in a highnumber of co-administrations of potentially incompatibledrugs. Other nurses did not know about the guidelines, ashas been already highlighted in the example of theinfusion of propofol.Errors were also observed during the transcriptionprocess (handling technology). On the surgical ward,before the drug round, one nurse checked all the drugcharts and wrote a small label for each drug that was due.The drugs were then prepared according to the informa-tion on the labels. It was observed that a nurse failed tonotice a prescription, did not write a label and thereforethe medication was omitted. Another transcription errorresulted in several administration errors. A patient was Table 2  Error-producing conditions ( n =117) relating to 71 slips,lapses and mistakesCategory Factors ( n =number of times the factorwas identified)Workload End of shift (2)Communication Information problems between:Nurses (1)Nurses and pharmacists (1)Doctor and nurses (e.g. ambiguous pre-scription) (3)Handling technology Lack of knowledge/routine/experience of:Medication preparation (27)Medication administration technique (38)Inadequate use of drug chart (4)Technology Ambiguous/complicated/unsuitable:Manufacturer information leaflet (6)Design of drug preparation/administrationtechnology (1)Administration guidelines (34)135  prescribed, amongst other drugs, pirenzepine as a pro-phylaxis against gastric ulcer while in the intensive careunit. After transfer to the surgical ward, the doctorprescribed a new list of medications in the medical notes,replacing pirenzepine with ranitidine. Subsequently, nurs-ing staff made a transcription error, both drugs werewritten on the drug chart and the patient receivedpirenzepine and ranitidine for several days. Presumably,the nurse had transcribed the drugs on the doctor’s list aswell as the drugs from the drug chart the patient hadwhilst in the intensive care unit (K, 101; observer’snotes).Communication problems were also identified. Onenurse assembled drugs on a trolley for the morning druground. She wrote the patient’s name, the room numberand dose on the labels of the infusion containers using awater-soluble pen. One of the patients was prescribed adose of 5 g mezlocillin, which was supposed to beprepared from three ampoules containing 2 g mezlocillineach. The nurse wrote 5 g on the infusion container whichwas supposed to be used for this drug preparation. Asecond nurse prepared all the drugs, but the writing wasno longer legible after touching the label. She prepared6 g of mezlocillin for the patient (K, 118; observer’snotes).Latent conditionsUnderlying problems included a lack of training in IVdrug preparation and administration, a lack of pharma-ceutical support at ward level and other organisationalissues.The hospital did not provide formal training in IV drugpreparation and administration. The nurses learnt thesetasks from each other on the ward, a process which wasneither formalised nor standardised to ensure a sufficientlevel of knowledge on common procedures. Conversa-tions with pharmacists showed that they did not knowabout the nurses’ practical preparation or administrationproblems. Therefore, they did not provide more informa-tion on common incompatibilities, set up policies on druguse, or become involved in training nurses in preparationprocedures.It was hospital policy that nursing staff on the surgicalward could administer IV drugs as intermittent infusionsor continuous infusions, but not as bolus doses orinjections. Therefore, the doctors prescribed most bolusdoses to be administered as intermittent infusions. Thesedrugs had to be diluted with a suitable infusion fluid.However, since the drugs were intended for bolus doseadministrations, the manufacturers’ leaflets did not con-tain information on possible diluents. Furthermore, theprescribing practice also meant that patients were oftengiven several intermittent infusions at the same timethrough the same infusion line, which increased thenumber of incompatibility errors. Instead, bolus dosescould have been given one after the other.The system used for documenting prescriptions on thesurgical ward involved transcribing each order twice. Thiswas identified as contributing to IV medication errors. Itwas also common practice to prescribe medication usingabbreviations for the drugs, prescribing the brand namewithout specifying the generic name and not specifyingthe dose. All of this may also have contributed to thetranscription errors. Discussion We have systematically investigated causes of IV drugerrors on two wards in a German hospital. IV drug errorswere not only caused by the immediate individual act: wefound a range of issues contributing to the errors, in-cluding training, pharmaceutical support, guidelines andthe system of documenting prescriptions.Legally, IV drug administration has to be carried outby medical staff in German hospitals (Schell 1995), but itcan be delegated to nurses (Gesetz ber die Berufe in derKrankenpflege (Krankenpflegegesetz—KrPflG) 1985). Inpractice, all IV drug administrations on the study wardswere carried out by nurses. The current legal framework indicates that in contrast to other countries, such as theUK (UKCC 2000), IV drug administration is not recog-nised as a nursing task in Germany. Therefore nurses arenot formally trained in IV drug preparation and admin-istration. There is no quantitative data on the extent towhich IV drug administration is delegated to nursing staff in other German hospitals. But we observed practicessimilar to the current hospital in three other hospitals(Hoppe-Tichy et al. 2002; Wirtz et al. 2003). The legalframework needs to be revised to clarify the roles of nurses and doctors in IV drug administration. Nationaleducational standards and local hospital policies shouldoutline nurse training courses in IV drug preparation andadministration.Clinical pharmacists have been recognised as keyhealth professionals to ensure safe drug use in the UK (Audit Commission for Local Authorities and the Na-tional Health Service in England and Wales 2001). Wefound in our study that the German pharmacists wererarely involved in ward-based activities of IV drugtherapy and knew little about the practical problems of the nurses on the wards. Furthermore, the guidelines thatwere provided by the pharmacy department were difficultto use on the ward. Clinical pharmacy activities should beextended to include ward-based activities. A multidis-ciplinary team, including clinical pharmacists, shouldidentify risks for IV drug errors and provide up-to-dateguidelines on IV drug preparation and administration. Forexample, a project in a German hospital showed that award-based clinical pharmacist was able to advise on theco-administration of drug infusions (Reinecke et al.1999). Pharmacists should also be involved in trainingcourses on IV drug preparation and administration fornurses. 136
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks