An observational study of drug administration errors in a Malaysian hospital (study of drug administration errors)

Full text



An observational study of drug administration errors

in a Malaysian hospital (study of drug administration


S. S. Chua*

BPharm (Hons) PhD

, M. H. Tea*

BPharm (Hons)

and M. H. A. Rahman


Departments of


Pharmacy and

Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur,



Background and objective: Drug administration errors were the second most frequent type of medication errors, after prescribing errors but the latter were often intercepted hence, administration errors were more probably to reach the patients. Therefore, this study was conducted to determine the frequency and types of drug administration errors in a Malaysian hospital ward.

Methods: This is a prospective study that involved direct, undisguised observations of drug administrations in a hospital ward. A researcher was stationed in the ward under study for 15 days to observe all drug administrations which were recorded in a data collection form and then com-pared with the drugs prescribed for the patient.

Results: A total of 1118 opportunities for errors were observed and 127 administrations had errors. This gave an error rate of 11Æ4% [95%

confidence interval (CI) 9Æ5–13Æ3]. If incorrect time errors were excluded, the error rate reduced to 8Æ7%(95%CI 7Æ1–10Æ4). The most common types of drug administration errors were incorrect time (25Æ2%), followed by incorrect technique of administration (16Æ3%) and unauthorized drug errors (14Æ1%). In terms of clinical significance, 10Æ4% of the administration errors were consid-ered as potentially life-threatening. Intravenous routes were more likely to be associated with an administration error than oral routes (21Æ3% vs. 7Æ9%,P< 0Æ001).

Conclusion: The study indicates that the fre-quency of drug administration errors in devel-oping countries such as Malaysia is similar to that in the developed countries. Incorrect time errors were also the most common type of drug admin-istration errors. A non-punitive system of report-ing medication errors should be established to encourage more information to be documented so that risk management protocol could be developed and implemented.

Keywords: drug administration error, hospital ward, medication error, observation, types of error


The National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) defines a medication error as ‘any preventable event that may cause or lead to inappropriate medication use or patient harm while the medica-tion is in the control of the health care professional, patient or consumer’ (1). Medication errors can occur during several stages of the drug delivery process, which have been categorized as prescrib-ing, transcribprescrib-ing, dispensing and administration (2).

Many studies have been conducted on medica-tion errors including reviews of terminologies and methodologies used (3). Studies on medication errors dated as far back as the 1960s by Barker and McConnell who demonstrated that medication errors happened much more frequently than could be obtained via incident reports, with about 16 errors per 100 doses (4). A more recent study found Received 5 June 2008, Accepted 2 September 2008

Correspondence: Associate Prof. Dr Siew Siang Chua, Depart-ment of Pharmacy, Faculty of Medicine, University of Malaya,

50603 Kuala Lumpur, Malaysia. Tel.: +603 7967 6688⁄+601 9336


that nearly one medication error occurred in every five doses (5). In addition, 7% of the errors were considered to be potentially harmful.

Drug administration is an activity that is prone to errors, partly attributed to the rapid develop-ment in medical technology, leading to a tre-mendous increase in types and complexity of medical devices as well as the number of medi-cations being introduced into the market (6). In addition, there are various routes of administra-tion, different dosages, dosage forms and dosing regimens which are often changed according to the patient’s clinical condition and diagnostic test results available (7).

Several studies on drug administration errors that used observational method reported error rates that varied from 3 to 44% (5, 7–12). A higher error rate (44%) has been observed in intensive care units as patients were usually on polypharmacy, parenteral drugs and were less likely to detect or intervene potential errors caused by their clinical conditions (12). Besides the variation in settings, comparison between studies is also complicated by the use of different terminologies and methods of calculating the error rates (7, 10, 12).

It has been reported that medication errors occurred most frequently at the prescribing stage, followed by administration, transcription and dis-pensing stages (13, 14). However, prescribing errors were often intercepted, unlike drug admin-istration errors (48% of prescribing errors were intercepted compared with 2%or less of adminis-tration errors) (13, 14). This means that drug administration errors are more likely to reach the patients and has higher potential for causing patient harm.

Medication errors should be viewed as sources of information concerning the safety of a healthcare system so that appropriate lessons can be learnt to improve patient safety (2). However, there is a lack of research on drug administration errors in developing countries such as Malaysia. Therefore, this study was conducted to determine the fre-quency and types of drug administration errors in a Malaysian hospital ward as health care system in Malaysia may differ from that of developed coun-tries. In addition, factors that contributed to and possible causes of drug administration errors were identified.



A drug administration error is defined as a dis-crepancy between the drug therapy received by the patient and that intended by the prescriber or according to standard hospital policies and proce-dures (6, 7, 15). In this study, it also encompassed the process of drug preparation in the ward but excluded prescribing errors. The error rate was calculated using the total opportunities for error (TOE) which is the sum of all doses ordered plus all the unordered doses given (3). The drug adminis-tration error rate was then calculated as the num-ber of doses with errors (incorrect in one or more ways) divided by the TOE and multiplied by 100 to obtain the percentage of errors, which would not exceed 100%(3, 5, 12).

Drug administration errors in this study were classified into 11 categories, similar to that used by other authors (3, 5, 11, 16): incorrect time, incorrect administration technique, unauthorized or unor-dered drug, incorrect drug preparation, incorrect dose, omission, incorrect rate, incorrect drug, deteriorated drug, extra dose and other errors which were not specified. An incorrect time error was defined as the administration of drugs an hour or more before or after the scheduled time (5, 10, 17, 18). Incorrect dose was defined as the admin-istration of doses ±10% or more of the original prescribed dose (5, 10, 12, 17).

Unauthorized or unordered drug error was the administration of a dose of medication that was not ordered by the doctor for the patient. This included the administration of a drug to the wrong patient but not the administration of a wrong medication (that is ‘incorrect drug’). Some drug administration errors would lead to subsequent errors. For exam-ple, incorrect drug preparation would result in incorrect dose but only incorrect preparation technique was considered as an error and not the incorrect dose.


This study was conducted in the haematology ward of a teaching hospital in Malaysia. The ward has 22 beds and practised the ward stock system where commonly used drugs were supplied in


bulk to the ward whereas more specific drugs were supplied as unit-of-use items on receipt of a pre-scription by the pharmacy. Orders for drugs were written by the doctors directly onto the patient’s medication file without any transcribing. The nur-ses administered the drugs with reference to the medication files. If the drugs ordered were obtained from the ward stock then the order was not checked by a pharmacist. However, prescrip-tions to be sent to the pharmacy for the supply of unit-of-use drugs were transcribed from the med-ication files by the nurses and these would be checked by a pharmacist before the drugs were supplied by the pharmacy to the ward for the specific patient.

All the oral medications and intravenous (i.v.) infusions were administered by the nurses whereas the i.v. bolus and intrathecal injections were administered by the doctors. There was no clinical pharmacist attached to the ward but any chemo-therapy required by the ward would be centrally prepared by pharmacy staff in the Sterile Complex of the hospital where it would be checked by a pharmacist. Administrations of most of the i.v. drugs in the ward were not checked by a second person except for chemotherapy and certain high risk items such as blood products.

All drug supplies to the ward were logged in the Pharmacy Information System (PIS), both for ward stocks and unit-of-use items. Any drug adminis-tration in the ward would be recorded in the patient medication file. This would be signed by the prescriber to initiate the use of the drug in a particular patient whereas subsequent administra-tions would be signed and dated by the staff who administrated the drug. Any remainder drug sup-plied as unit-of-use item to the ward would be recorded in a book and subsequently returned to the pharmacy.

The PIS kept records of all drugs purchased, received, kept and used in the hospital. It can track the distribution of drugs in the hospital such as supplies of drugs to the wards and for individual patients. It also kept records of drugs dispensed to each outpatient, the quantity and date supplied. However, it does not keep records of who used the drugs supplied as ward stock. The PIS also contained a drug information database to aid in prescription screening where any overdose or drug–drug interactions would be detected.

Data collection procedures

This is a prospective study that involved direct, undisguised observations of drug administrations by nurses or medical doctors. The methodology followed closely to that used in many other studies (5, 8, 10, 11, 19) and has been demonstrated to be the most accurate method for detecting drug administration errors (3, 19, 20).

Similar to the study by Dean and Barber (20), the ward staff was briefed about the study with partial explanation of its purpose but the word ‘error’ was avoided. One of the main concerns with direct observation method is the effect of the observer on those being observed and hence may bias the study results. However, this Hawthorne effect disap-peared after a few days of observation as the ward staff started to forget about the study and behaved as usual (21). It has also been demonstrated that such observations did not affect the error rate sig-nificantly (20).

A pilot study was conducted for 2 days in the same ward to test the practicability and feasibility of the methodology as well as to reduce the Haw-thorne effect. Amendments were made to the data collection form so that it is more user friendly and practical for the researcher to take notes while observing each drug administration process.

The study was conducted over eight consecutive weeks between November 2004 and January 2005 with a total of 15 working days (excluding one public holiday). A researcher was stationed in the ward every Wednesday and Friday during the study period from 07Æ30 to 21Æ00 hours with short lunch and dinner breaks. The researcher followed the ward staff who was involved in drug admin-istrations and recorded all aspects of drug retrieval, preparation and administration in a data collection form. After each round of observation, the researcher compared the information recorded with the doctor’s orders in the patient’s medication file to detect any discrepancies. All the data col-lected were discussed with the other members of the research team (an experienced pharmacist and a clinician). The identity of the nurses and doctors was recorded in codes and no individual was indicated in the data analysis.

For ethical reasons, the researcher would inter-vene if she noticed any drug administration errors that could result in potential patient harm but these


incidents were still recorded as errors. Other studies also allowed such interventions (11, 17). Dean and Barber (20) found that such interventions did not have any significant effect on the error rate. This study was approved by the Medical Ethics Committee of the hospital concerned.

All data collected were entered into and analy-sed using the Statistical Package for Social Sciences (SPSSSPSSInc., Chicago, IL, USA), version 12.0. Associa-tions between variables were analysed using Pearson’s chi-squared (v2) test. AnyP-value <0Æ05 was considered as statistically significant.

All the drug administration errors were classi-fied into four categories based on that used by Stubbs et al. (22), for prescribing errors: Grade (1) probably clinically insignificant, (2) minimal clini-cal significance, (3) definitely cliniclini-cally significant and could cause patient harm and (4) potentially life-threatening. The drug administration errors were initially classified independently by a clini-cian and a pharmacist. The results were tested for correlation using Kappa statistics, j. The strength of agreement was classified as: j< 0Æ20 is poor, 0Æ21–0Æ40 is fair, 0Æ41–0Æ60 is moderate, 0Æ61–0Æ80 is good and 0Æ81–1Æ00 is very good (23). A consensus on the final classification was derived with the help of another clinical pharmacist.


During the study period, 1118 TOE were observed and 127 doses had at least one error. This gives an

error rate of 11Æ4% (95% CI 9Æ5–13Æ3). When incorrect time errors were excluded 97 doses had errors and the error rate reduced to 8Æ7%(95%CI 7Æ1–10Æ4). Of the 127 doses with errors, eight had two types of errors, giving a total of 135 adminis-tration errors. Of these, 26% (35 errors) were intervened either by the researcher (19 errors or 14%), staff nurse (five errors or 4%) or patients (11 errors or 8%) and hence did not reach the patient. The types of drug administration errors observed are shown in Fig. 1. Incorrect time errors ranged from 1 to 10 h beyond the scheduled time of administration. Unauthorized drug included six cases of drugs given to the wrong patients as the nurse took the wrong medication files but these were intercepted by the researcher. The four errors under ‘Others’ included the use of the same syringe to administer the same drug to different oncology patients. Examples of the different types of drug administration errors are shown in Table 1. Possi-ble reasons for the occurrence of drug administra-tion errors were derived after interviewing the ward staff involved.

A total of 37 doses of cytotoxic drugs given had seven errors, giving an error rate of 18Æ9%(95%CI 6Æ3–31Æ5). If cytotoxic drugs were excluded, the error rate dropped to 120 of 1081 TOE that is, 11Æ1% (95% CI 9Æ2–13Æ0). The error rate associated with cytotoxic agents was higher but did not reach any statistical significance (v2= 2Æ172 andP= 0Æ141).

Classification of the drug administration errors into the four categories of clinical significance by

Fig. 1. Types of drug administra-tion errors (n= 135).


Table 1. Examples of drug administration errors and possible reasons

Types of errors Description (examples) Possible reasons Incorrect time Clauvulanate + amoxicillin (Augmentin)

at 08Æ00 hour, administered at 12Æ05 hour

Heavy workload Tranexamic acid at 8Æ00 hour, administered >1 h later Heavy workload Filgrastim (Neupogen) subcutaneous

at 16Æ00 hour, administered at 08Æ00 hour

Misread medication file Incorrect

administration techniques

Bubbles or air trapped in the i.v. infusion line. Lack of knowledge Wipe eyes with cotton wool immediately after

administration of prednisolone eye drops

Lack of knowledge Unordered or

unauthorized drug

Tranexamic acid 500 mg given but should be withheld as patient was to have CT scan

Order not passed by previous shift nurse

Diphenhydramine, chlopheniramine and

dihydrocodeine given without doctor’s authorization

Patient requested Itraconazole, allopurinol, tramadol, folic acid,

KCl, hydroxyurea, norethisterone given to patients fasting for procedures

Did not notice ‘Fasting’ sign

Tranexamic acid, folic acid, KCl and norethisterone in same file given to one wrong patient

Wrong medication file in patient’s folder Piperacillin + tazobactam (Tazocin) to wrong patient Wrong patient file Allopurinol to wrong patient Wrong patient file Incorrect


Amphotericin B not properly diluted and some powder still left in the vial

Lack of knowledge Bent the needle to syringe the drugs out of

ampoules and cause spill onto the floor

Lack of knowledge Nurse cut her finger when opening vial but

continued to prepare injection with a bleeding finger

Lack of awareness Incorrect dose 1 g tranexamic acid given as 500 mg Calculation error

2 g hydroxyurea was ordered, 1 g was administered Misread medication file 75 mg of cyclosporine given as 2 caps.

of 25 mg instead of 3 caps.

Calculation errors 4 mg i.v. metoclopramide given as 8 mg In a hurry 750 mg i.v. vancomycin given as 1 g In a hurry

Omission errors Imipenam 1 g was not given Misread medication file Chlopheniramine was ordered to be omitted on the

next day but the nurse omitted on the day ordered

Misintepreted the medication file Amlodipine, allopurinol, paracetamol not given No ward stock Famotidine not given but signed Miscommunication Incorrect rate Cytarabine to infuse over 3 h but infused for 2 h Infusion pump error

Dexamethasone should be administered slowly to avoid perineal pruritus

Lack of knowledge Vancomycin to be infused over 2 h but

infused over 30 mins

Did not follow instructions Deteriorated drug Allopurinol 200 mg expired on 27th October 2004

was given to the patient on 26th November 2004

Did not check expiry date Extra dose Itraconazole and tramadol given twice Drug given but not signed

Tacrolimus prescribed as EODabut given o.d.b Misread medication file Other errors Same syringe used to administer frusemide or heparin

saline to 4 different patients prescribed the same drug

Lack of awareness Incorrect drug Mefenamic acid 500 mg given instead of

tranexamic acid 500 mg

Misread drug name

aEOD, every other day. bo.d., every day


the two healthcare professionals were significantly correlated with j = 0Æ791 (good agreement). The final classification (Table 2) also has good and very good agreement with the separate classification by the clinician and the pharmacist withj= 0Æ749 and 0Æ833, respectively. Errors classified as potentially

life-threatening are also shown in Table 2. No death attributed to drug administration errors was recorded during the study.

Table 3 shows that the rate of drug administration errors is significantly associated with the personnel involved and the routes of administration but not with the time of administration. The time of drug administration was divided into morning (07Æ30– 14Æ00 hours) and afternoon (14Æ01–21Æ00 hours), based on the nurses’ shift duties in the ward. Other routes of administration, intrathecal (3), subcutane-ous (21) and eye drops (3) were omitted in the analysis as the numbers were very small.


The results showed that drug administration errors were common in the Malaysian hospital ward. The error rate was 11Æ4%which reduced to 8Æ7%when incorrect time errors were excluded. These results are similar to that reported in the literature (5, 11). In this study, 26% of the drug administration errors were intercepted. This implies that double checking by an independent person is essential. Some of the errors were intercepted by the patients themselves and hence, patient education is also important to prevent medication errors.

Incorrect time errors were the most common type of errors in this study but many of these errors were not likely to cause patient harm except for those drugs that require close serum concentration monitoring. However, it may be an indicator of a Table 2. Classification of drug administration errors

according to clinical significance Categories of


significance Frequency

Percentage (n= 135)

Probably clinically insignificant 48 35Æ5

Minimal clinical significance 29 21Æ5

Definitely clinically significant 44 32Æ6

Potentially life-threatening 14 10Æ4

Same syringe used for different patients


Wrong dose of cyclosporine 4

Wrong rate of cytarabine administration

3 Preparing i.v. KCl with a

bleeding finger


Extra doses of tacrolimus 1

Tazocin injection given to wrong patient

1 KCl injection given to

wrong patient

1 Tranexamic acid but gave

mefenamic acid


KCl, potassium chloride.

Table 3. Factors which may be associated with drug administration errors Factors No. of doses with errors Total No.

of doses %errors v2 P-values

OR (95%CI) Persons involved Doctors 37 89 41Æ6 87Æ664 <0Æ001* 7Æ42 (4Æ62–11Æ92) Nurses 90 1029 8Æ7 Time Morning shift 78 607 11Æ4 0Æ001 0Æ971 1Æ01 (0Æ69–1Æ47) Afternoon shift 49 384 11Æ3 Routes of administration i.v. route 60 282 21Æ3 37Æ081 <0Æ001* 3Æ15 (2Æ15–4Æ61) Oral 64 809 7Æ9 *Statistically significant atP< 0Æ01.


system failure (3, 6). Therefore, it has been recom-mended that both error rates that are inclusive and exclusive of wrong time errors should be reported (3, 6). One of the main causes of medication errors is the heavy staff workload (11, 24). In most hos-pital, the scheduled time for drug administration is the busiest time when the nurses have to make the patients’ beds, monitor patients’ physical signs and assist the doctors in their ward rounds (18). One possible solution is to increase the number of ward staff. The drug administration schedule can also be planned such that not all patients take their medi-cations at 08Æ00 hour. Medications that are pre-scribed as a daily dose could be administered during noon or evening time when the staff workload is lighter (18).

Unauthorized drug errors were associated with the administration of drugs to patients who were fasting for some procedures as the ward staff did not notice the ‘fasting’ sign. Perhaps, such sign should be more conspicuous so that it would not be missed. Drugs were almost given to the wrong patients. Misreading the medication files was a common cause of errors in this study. The ward management should rearrange patient medication files. The ward staff should always check the name of the patient before administering the drugs. Ward stock should be checked regularly to ensure that there is adequate supply and also to detect any expired or deteriorated drugs. Presence of a clinical pharmacist in the ward would serve as a resource person for drug administration and also act as a safety net in double checking the drugs used. Standard operating procedures to ensure that all drugs are checked by an independent person before being given to the patient should be established.

Cohen reported that documentation of drug administration is one of the contributory factors to administration errors in the ward (18). This was also observed in this study. Once a dose of drug has been administered, it must be signed or recorded immediately. Otherwise, it may be for-gotten and the patient may be given another dose. Similarly, recording the administration of drugs before it is given may run the risk of a dose omis-sion if the staff is called off to attend to other duties before the dose is delivered (18).

Incorrect technique of administration involved mainly the presence of air bubbles in i.v. infusion and wiping of the eyes with cotton wool

immedi-ately after administering eye drops. Incorrect drug preparation was mainly related to the incomplete reconstitution of dry powder injections. Proper methods of preparation and administration of parenteral drugs are important to prevent throm-bus formation, hypersensitivity reactions and infections (17). Training in drug administration technique and preparation as well as awareness programmes should be conducted.

Some infusion pumps did not produce the rate of infusion stated and this led to incorrect rate of administration. For example, a setting of 180 mL⁄h for 500 mL of cytarabine should take about 3 h to complete the infusion instead of around 2 h. Ward devices such as infusion pumps should be checked and calibrated regularly. This is important as some medications have to be administered over a specific time interval to achieve the optimal therapeutic effect and minimum side-effects.

Muller (25) reported that cytotoxic drugs have a higher risk of medication errors than other classes of drugs. This was found in this study but did not reach any statistical significance, perhaps because of the small number of cytotoxic drugs included as shown by the large 95%CI (6Æ3–31Æ5). Errors in the administration of cytotoxic drugs are more likely to be of clinical consequences because of its narrow margin of safety and serious adverse effects. Therefore, the preparation and administration of cytotoxic drugs should only be handled by quali-fied and experienced personnel (25).

The utilization of integrated clinical information system technology has been demonstrated to reduce medication errors (26). The teaching hospi-tal under study has a computerized Hospihospi-tal Information System, PIS, Laboratory Information System and also a computerized physician order entry (CPOE). However, these systems were not integrated and the use of CPOE was limited as inpatient prescriptions were still written manually. The 10Æ4% of errors that were considered as potentially life-threatening included using the same syringe for different oncology patients. This increased the risk of infections which could be life-threatening in these immunocompromised patients (27). These results are comparable with that of other studies (5, 10, 11).

Intravenous drug administration is more prone to errors as it involved more complex activities (24, 28). In this study, i.v. drug administration is significantly


more likely to be associated with medication errors than the oral routes (21Æ3%vs. and 7Æ9%). Bruce and Wong (17) also reported administration error rate of 17Æ0–33Æ5%for parenteral drugs. Errors associated with i.v. administrations were mainly caused by wrong administration rate and technique, similar to that reported by Wirtzet al.(29).

In this study, medical doctors were more likely to commit medication errors than nurses. One possible explanation is that medical doctors were involved more in i.v. administrations whereas oral routes were given by nurses. In addition, because of their heavier responsibilities, doctors were more likely to administer the medications later than the scheduled time, leading to incorrect time errors.

Drug administration errors in the ward were often not reported or under reported. More medi-cation error studies are recommended to increase the awareness of ward staff concerning such problems. This study was one of the first on drug administration errors to be conducted in Malaysia and it serves to open the minds of healthcare pro-fessionals and to stimulate further interest and concern in patient safety. In addition, a non-puni-tive system of reporting medication errors should be established to encourage more comprehensive data to be documented so that healthcare providers and institutions could share and learn from the mistakes of each other, and appropriate measures could be implemented to prevent any future errors. One of the limitations of the study was that there could be more than one ward staff admin-istering drugs at the same time but there was only one observer therefore, some drug administrations may have been missed. The researcher was also not able to observe some of the intrathecal administrations as these were conducted in a separate room. In addition, the data collected were limited to the days and shift the observer was stationed in the ward. Only a specific ward was studied and hence may not be representative of all wards in the hospital. The results may also be different between the various types of hospitals in Malaysia particularly between teaching and non-teaching hospitals.


The study indicates that the frequency of drug administration errors in developing countries

such as Malaysia, with an error rate of 11Æ4% is similar to that in the developed countries. Incor-rect time errors were also the most common type of drug administration errors. A system for double checking in the ward is essential as well as the utilization of integrated clinical informa-tion system technology should be considered. Training and awareness programmes should be conducted to reduce the incidence of drug administration errors. A non-punitive system of reporting medication errors should be established to encourage more information to be documented so that risk management protocol could be developed and implemented.


The authors thank Ms Che Zuraini Sulaiman, a clinical pharmacist from UMMC for helping with the classification of medication errors into the cat-egories of clinical significance. Most of all, the authors express their greatest appreciation to the ward staff for their kind assistance and cooperation throughout the observation period in the ward.


1. The National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) (2008) About Medication Errors. Available at: http:// (accessed 5 January 2008).

2. Allard J, Carthey J, Cope J, Pitt M, Woodward S (2002) Medication errors: causes, prevention and reduction. British Journal of Haematology,116,255–265.

3. Allan EL, Barker KN (1990) Fundamentals of medi-cation errors research. American Journal of Hospital Pharmacy,47,555–571.

4. Barker KN, McConnell WE (1962) The problems of

detecting medication errors in hospitals. American

Journal of Hospital Pharmacy,19,360–369.

5. Barker KN, Flynn EA, Pepper GA, Bates DW, Mikeal RL (2002) Medication errors observed in 36 health care facilities.Archives of Internal Medicine,162,1897–1903. 6. Pepper GA (1995) Understanding and preventing drug misadventures: errors in drug administration

by nurses. American Journal of Health-System


7. Greengold NL, Shane R, Schneider P et al. (2003)

The impact of dedicated medication nurses on the medication administration rate. Archives of Internal Medicine,163,2359–2367.


8. Calabrese AD, Erstad BL, Brandl K, Barletta JF, Kane SL, Sherman DS (2001) Medication adminis-tration errors in adult patients in the ICU. Intensive Care Medicine,27,1592–1598.

9. Dean BS, Allan EL, Barber ND, Barker KN (1995) Comparison of medication errors in an American and British hospital.American Journal of Health-Sys-tem Pharmacy,52,2543–2549.

10. Tissot E, Cornette C, Demoly P, Jacquet M, Barale F, Capellier G (1999) Medication errors at the admin-istration stage in an intensive care unit.Intensive Care Medicine,25,353–359.

11. Tissot E, Cornette C, Limat S et al. (2003) Observa-tional study of potential risk factors of medication administration errors.Pharmacy World & Science,25, 264–268.

12. van den Bemt PM, Fijn R, van der Voort PHJ, Gossen AA, Egberts TC, Brouwers JR (2002) Frequency and determinants of drug administration errors in the intensive care unit. Critical Care Medicine, 30, 846– 850.

13. Bates DW, Cullen DJ, Laird Net al.on behalf of the ADE Prevention Study Group (1995) Incidence of adverse drug events and potential adverse drug events: implications for prevention. Journal of the American Medical Association,274,29–34.

14. Leape LL, Bates DW, Cullen DJet al.(1995) Systems

analysis of adverse drug events.Journal of the Amer-ican Medical Association,274,35–43.

15. Dean B (1999) Errors in medication administration. Intensive Care Medicine,25,341–342.

16. American Society of Hospital Pharmacists (1993) ASHP guidelines on preventing medication errors in hospitals. American Journal of Hospital Pharmacy, 50, 305–314.

17. Bruce J, Wong I (2001) Parenteral drug administra-tion errors by nursing staff on an acute medical

admissions ward during day duty. Drug Safety, 24,


18. Cohen MR (1999) Errors associated with medication administration. In: Cohen MR, ed.Medication Errors. Washington, DC: American Pharmaceutical Associ-ation, 11.1–11.16.

19. Flynn EA, Barker KN, Pepper GA, Bates DW, Mikeal RL (2002) Comparison of methods for detecting medication errors in 36 hospitals and skilled-nursing facilities.American Journal of Health-System Pharmacy, 59,436–446.

20. Dean B, Barber N (2001) Validity and reliability of observational methods for studying medication

administration errors. American Journal of

Health-System Pharmacy,58,54–59.

21. Barker KN (1980) Data collection techniques:

obser-vation. American Journal of Hospital Pharmacy, 37,


22. Stubbs J, Haw C, Cahill C (2004) Auditing prescrib-ing errors in a psychiatric hospital. Are pharmacists’ interventions effective? Hospital Pharmacy, 11, 203– 206.

23. Altman DG (1999) Practical Statistics for Medical

Research, 1st edn. London: Chapman & Hall⁄CRC, 403–405.

24. Gladstone J (1995) Drug administration errors: a study into the factors underlying the occurrence and reporting of drug errors in a district general hospital. Journal of Advanced Nursing,22,628–637.

25. Muller T (2003) Typical medication errors in oncol-ogy: analysis and prevention strategies. Onkologie, 26,539–544.

26. Mahoney CD, Berard-Collins CM, Coleman R, Amaral JF, Cotter CM (2007) Effects of an integrated clinical information system on medication safety in a multi-hospital setting. American Journal of Health-System Pharmacy,64,1969–1977.

27. Leon TG, Pase M (2004) Essential oncology facts

for the float nurse.MEDSURG Nursing,13,165–171,


28. Cousins DH, Sabatier B, Begue D, Schmitt, Hoppe-Tichy T (2005) Medication errors in intravenous drug preparation and administration: a multicentre audit in the UK, Germany and France.Quality and Safety in Health Care,14,190–195.

29. Wirtz V, Taxis K, Barber ND (2003) An observational study of intravenous medication errors in the United

Kingdom and in Germany. Pharmacy World &