SciELO - Scientific Electronic Library Online

 
vol.43 issue3Survival of breast cancer women in the state of Rio de Janeiro, Southeastern BrazilReliability of food labels from products marketed in the city of São Paulo, Southeastern Brazil author indexsubject indexarticles search
Home Page  

Services on Demand

Article

Indicators

Related links

Share


Revista de Saúde Pública

Print version ISSN 0034-8910

Rev. Saúde Pública vol.43 n.3 São Paulo May./Jun. 2009 Epub Apr 17, 2009

http://dx.doi.org/10.1590/S0034-89102009005000028 

ORIGINAL ARTICLES

 

Errors in hospital prescriptions of high-alert medications

 

Errores en la prescripción hospitalaria de medicamentos potencialmente peligrosos

 

 

Mário Borges RosaI; Edson PeriniII; Tânia Azevedo AnacletoI; Hessem Miranda NeivaI; Tânia BogutchiIII

IFundação Hospitalar do Estado de Minas Gerais. Belo Horizonte, MG, Brasil
IIDepartamento de Farmácia Social. Faculdade de Farmácia. Universidade Federal de Minas Gerais. Belo Horizonte, MG, Brasil
IIIDepartamento de Estatística. Pontifícia Universidade Católica de Minas Gerais. Belo Horizonte, MG, Brasil

Correspondence

 

 


ABSTRACT

OBJECTIVE: Medication errors are currently a worldwide public health issue and it is one of the most serious prescription errors. The objective of the study was to evaluate the practice of prescribing high-alert medications and its association with the prevalence of medication errors in hospital settings.
METHODS: A retrospective cross-sectional study was conducted including 4,026 prescription order forms of high-alert medications. There were evaluated all prescriptions received at the pharmacy of a reference hospital in the state of Minas Gerais, southeastern Brazil, over a 30-day period in 2001. Prescription were checked for legibility, patient name, type of prescription, date, handwriting or writing, prescriber identification, drug prescribed, and use of abbreviations. Prescription errors were classified as writing or decision errors and how the type of prescription affected the occurrence of errors was assessed.
RESULTS: Most prescriptions were handwritten (45.7%). In 47.0% of handwritten, mixed and pre-typed prescriptions had patient name errors; the prescriber name was difficult to identify in 33.7%; 19.3% of them were hardly legible or illegible. Of a total of 7,148 high-alert drugs prescribed, 3,177 errors were found, and the most frequent one was missing information (86.5%). Errors occurred mostly in prescriptions of heparin, phentanyl, and midazolam. Intensive care and neurology units had the highest number of errors per prescription. Non-standard abbreviations were frequent and widespread. Overall it was estimated 3.3 errors per prescription order form. Pre-typed prescriptions were less likely to have errors compared to mixed or handwritten prescriptions.
CONCLUSIONS: The study results show there is a need for standardizing the prescription process and eliminating handwritten prescriptions. The use of pre-typed or edited prescriptions may reduce errors associated to high-alert medications.

Descriptors: Medication Errors. Prescriptions, Drug. Drugs with Prescription. Drugs of Special Control. Cross-Sectional Studies.


RESUMEN

OBJETIVO: Los errores de medicación son actualmente un problema mundial de salud pública, siendo los más serios los de prescripción. El objetivo del estudio fue analizar la práctica de la prescripción de medicamentos de alto riesgo y su relación con la prevalencia de errores de medicación en ambiente hospitalario.
MÉTODOS: Estudio transversal retrospectivo abarcando 4.026 prescripciones con medicamentos potencialmente peligrosos. Durante 30 días de 2001, fueron analizadas todas las prescripciones recibidas en la farmacia de un hospital de referencia del estado de Minas Gerais (Sureste de Brasil). Las prescripciones fueron analizadas con relación a: legibilidad, nombre del paciente, tipo de prescripción, fecha, caligrafía o grafía, identificación del prescriptor, análisis del medicamento y uso de abreviaturas. Los errores de prescripción fueron clasificados como de redacción o decisión, siendo evaluada la influencia del tipo de prescripción en la ocurrencia de errores.
RESULTADOS: Hubo predominio de la prescripción escrita a mano (45,7%). En 47,0% de las prescripciones escritas a mano, mixtas y pre-digitadas ocurrieron errores en el nombre del paciente, en 33,7% hubo dificultades en la identificación del prescriptor y 19,3% estaban poco legibles o ilegibles. En un total de 7.148 medicamentos de alto riesgo prescritos, fueron observados 3.177 errores, siendo más frecuente la omisión de información (86,5%). Los errores se concentraron principalmente en los medicamentos heparina, fentanil y midazolam; y los sectores de tratamiento intensivo y la neurología presentaron mayor número de errores por prescripción. Se observó el uso intensivo y sin estandarización de abreviaturas. Cuando se computaron todos los tipos de errores, se verificó 3,3 por prescripción. La prescripción pre-digitada presentó menor probabilidad de errores en comparación con las mixtas o escritas a mano.
CONCLUSIONES: Los resultados sugieren la necesidad de la estandarización en el proceso de prescripción y la eliminación de aquellas hechas a mano. El uso de prescripciones pre-digitadas o editadas podrá disminuir los errores relacionados a los medicamentos potencialmente peligrosos.

Descriptores: Errores de Medicación. Prescripción de Medicamentos. Medicamentos de Prescripción. Medicamentos de Control Especial. Estudios Transversales.


 

 

INTRODUCTION

Harvard Medical Practice studies I and II,4,12 which are milestones and pioneers in the area of patient safety, showed that adverse events are common and unexpectedly frequent in North American hospitals, causing permanent damage and deaths. From these two studies it is estimated that approximately 98,000 North Americans die each year because of errors associated with health care. Such errors are considered one of the main causes of death in the USA.10 The most frequent adverse events related to drugs were recorded in the Harvard Medical Practice Study II,12 of which a considerable number are avoidable. Over the last few years the significant increase in the number of studies into patient safety has led to greater knowledge of the subject and confirmed its importance as a worldwide problem. In line with this worrying scenario in 2004 the World Health Organization launched its World Alliance for Patient Safety. This is a permanent program that urges all member countries to guarantee the quality of care provided in health services worldwide.23

Recently, a publication on medication errors considered the level and consequences of these events to be unacceptable, since every inpatient in American hospitals is subject to one medication error per day.2 According to Barber et al3 (2003), prescription errors are the most serious among all medication use errors.

When discussing errors in health institutions there are many paradigms to be faced. Health professionals normally associate mistakes in their activities with feeling ashamed, loss of prestige and fear of punishment. Generally speaking, the environment in these health does not promote an open discussion about this issue with the aim of improving the system as a whole. This denial tendency, with the consequent under-reporting of errors, often makes the assessment of the events difficult and compromises finding out the facts.21 Another obstacle to the study and prevention of medication errors is the lack of standardization and the multiple terminology used for classifying them. This situation makes it difficult to compare studies on the subject and delays epidemiological knowledge about it.24

Although most medicines has a safe therapeutic margin, some drugs have an inherent risk of harming patients when there is a failure in the use process; these are what are here referred to as high-alert medications (HAMs). The errors that happen with these medicines are not of the most routine type, but when they occur they are serious and may lead to permanent damage or be fatal.5,8

The objective of this study was to analyze the practice of prescribing high-alert medications and their relationship with the prevalence of medication errors in hospitals.

 

METHODS

In a retrospective cross-sectional study, the data were obtained from all the second copies of prescription order forms with one or more HAMs that were prescribed between August 29 and September 27, 2001 and that were received in the pharmacy of a reference hospital in the state of Minas Gerais, Southeastern Brazil.

According to Cohen et al5 (1998) and Federico8 (2007) the following were considered to be HAMs: atracurium 25 mg ampoule; potassium chloride; 10 mL 10%; digoxin tablets 0.25mg; dobutamine 250 mg ampoules; dopamine 50 mg ampoules, epinephrine 1 mg ampoules; fentanyl 1 mL ampoules 0.5 mg and 10 mL vials with 0.5 mg; potassium acid phosphate 10 mL ampoules 2 meq/ml; calcium gluconate 10 mL ampoules 10%; heparin 0.25 mL ampoules 5000 units and 5 mL vials 5000 units/mL; NPH insulin 100 units/mL and Regular with 100 units/mL; mydazolam ampoules 50 mg, 5 mg and 15 mg; morphine 1 mL ampoules 10 mg; nalbuphine 1 mL ampoules 10 mg; norepinephrine 1 mL ampoules 1 mg; pancuronium 2 mL ampoules with 4 mg; pethidine 2 mL ampoules 100 mg; suxametonium vials with 100 mg; tramadol 10 mL vial 100 mg/mL; vecuronium 1 mL ampoule 4 mg, warfarin 5 mg tablets. To this list of 26 HAMs it was necessary to add four further definitions for record purposes on the form (unconcentrated mydazolam, heparin, insulin and fentanyl, pharmaceutical form or other parameter), because of various situations in which the HAMs appeared to have been prescribed without parameters, which made their classification on the original list impossible.

The form was prepared with eight classified research variables and was used for collecting and forming the database. It was completed by trainees, who were final year pharmacy undergraduates and reviewed by two hospital pharmacists with post-graduate degrees in hospital pharmacy. A third hospital pharmacist, with a specialization course in the area and more than 15 years experience in hospital pharmacy carried out a final review. If there was any disagreement the classification was discussed until agreement was reached. After review, discussion and consensus the forms were sent for inputting.

The following eight variables or groups of variables and two error definitions were used:

1. Legibility of the prescriptions - because of the high degree of subjectivity and dependence on the experience of the reviewer, the most homogenous assessment standard possible was established; this reduced the subjectivity involved with the final opinion. Legibility was classified as:

a) easily legible handwriting: read normally, without any difficulty in understanding what was written;

b) hardly legible or doubtful handwriting: it took longer to read and there was no certainty that all the words, numbers, symbols and abbreviations had been understood;

c) illegible handwriting: impossible to understand what was written.

2. Name of the patient - in addition to legibility criteria this variable was classified as: incomplete (omission of parts of the name, but without legibility problems); illegible/altered (name or surname of the patient altered during the observation period); not identified (patient hospitalized without documents and not identified).

3. Type of prescription - classified as: typed (printed computer-produced prescriptions generally word-processed); handwritten; mixed (prescription partly typed and partly handwritten).

4. Prescription date - in addition to its legibility it may be incomplete or omitted altogether.

5. Handwriting or spelling - applied to the body of the prescription, classified as illegible when at least 50% of the words are indecipherable.

6. Identification of the prescriber - the complete identification standard has the following parameters: prescription signed or initialed, stamped, and with name of the prescriber and number registered in the Conselho Regional de Medicina (CRM - Regional Medical Council) or Conselho Regional de Odontologia (CRO - Regional Dentistry Council) legibly written. Any other prescriber identification parameters that were not classified in accordance with the standard were considered incomplete. The anonymous prescription classification meant there was no prescriber identification.

7. Analysis of the HAM - Divided into four parts:

a) legibility of the HAM: legible or hardly illegible/doubtful.

b) pharmaceutical form: what was prescribed was compared with the reference to Clinical Pharmacology by Fuchs & Wannmacher9 (1998) and the drug medication use instructions, and classified as: correct or incorrect; incomplete (the complete pharmaceutical form was not described on the prescription, but there were no problems with legibility); hardly legible/doubtful (the pharmaceutical form prescribed left room for doubt and interpretation); omission (prescription without the pharmaceutical form of the medication).

c) concentration and route of administration: criteria the same as for the pharmaceutical form.

d) frequency and infusion rate: criteria the same as for the pharmaceutical form. The classification "not applicable" was added because in some situations frequency is not used. Example: continuous drug infusion.

All classifications are exclusive for all parameters.

8. Use of abbreviations - the type and number of abbreviations included in each prescription order form were counted and only those that were hardly legible or doubtful were recorded as errors. Legible abbreviations were not considered to be errors because there were no standardized abbreviations in use in the hospital studied. The abbreviations U or UI, meaning units, were separately counted because they were considered to be the most dangerous abbreviations.1,2,11

Prescription errors were classified in accordance with Dean et al6 (2000). Decision errors relate to prescriber knowledge, such as dosage error, the prescription of two drugs for the same purpose, medication not indicated for the patient or renal and hepatic insufficiency not considered, among others. The following were considered to be decision errors when they were wrong: pharmaceutical form, concentration, route of administration and frequency and infusion rate. Writing errors are related to the prescription writing process, such as illegibility, use of confused abbreviations (UI), omission of pharmaceutical form, concentration, route of administration, interval, frequency and infusion rate, error in the medication unit, among others. The following were classified as writing errors: patent's name (hardly legible or doubtful, incomplete, illegible, altered), date (hardly legible or doubtful, omitted, illegible, incomplete), handwriting (hardly legible and illegible), identification of the prescriber (all that were not complete), name of the HAM (hardly legible), pharmaceutical form, concentration, route of administration, interval (incomplete, hardly legible/doubtful and omitted) and abbreviations (only the hardly legible).

Data were input, worked on for consistency, reviewed and analyzed in Epi Info 6.04. Exploratory analysis of the data was carried out using descriptive statistics, with a calculation of the position measures (average, median) and variability (standard deviation and variation coefficient). Complementary analyses were carried out in SPSS 11.0, namely: univariate logistic regression, with calculation of the odds ratio (OR) for checking the relationship between the type of prescription and errors.

This study was approved by the Ethics Committee for Institutional Research on 1/30/2001.

 

RESULTS

All prescription order forms containing HAMs received in the hospital pharmacy over a 30 day period were analyzed. There were 4,026 prescriptions from 456 patients (mean= 8.8; SD= 0.4; median= 6.0 prescriptions/patient). The first quartile presented a maximum of two prescriptions and the third a minimum of 13, of which the last 5% presented more than 28 prescriptions. As for the structure of the prescription, most of them were handwritten. There were errors in the patient's name in 47.0% of all prescriptions and 19.3% problems of legibility (Table 1).

 

 

As for the identification of the professional who was the prescriber, a relationship of 1.3 prescribers per prescription was observed (5,153 records for a total of 4,026 prescriptions). In 1,734 (33.7%) of the records it was difficult or impossible to identify who prescribed it.

When the HAM prescription errors were analyzed it was seen that five types predominated, accounting for 93.5% of the records, with 86.5% referring to some type of information that had been omitted. Errors relative to the concentration of the HAM totaled 1,900 (59.8% of all those recorded) as shown in Table 2.

 

 

Therefore, 7,148 HAMs were recorded in 4,026 prescription order forms, in which some type of error was observed in 3,177 (44.5%) of them. The prevalence of errors was significantly different in the various sectors. The intensive care center (ICC) and the severe and chronic burns' sectors had the greatest relation of HAMs per prescription, while the intensive care unit (ICU) and the neurology and intermediary care sectors had the largest percentage of HAMs with errors (Figure). There was a difference of 11.2 times between the prevalence of errors per prescription in the ICC (2.8) and plastic surgery (0.2) sector. The severe/chronic burns and plastic surgery clinics used mainly typed prescriptions and the others used handwritten prescriptions.

Table 3 shows that 90% of the HAM errors are concentrated in nine drugs, among which are heparin, fentanyl and mydazolam. Heparin was the HAM that gave rise to most errors. The most frequent errors were omission of the pharmaceutical form and concentration, poor legibility and incomplete concentration. Heparin in ampoules and mydazolam were responsible for most of the errors related to concentration. There was a tendency to omit the pharmaceutical form when heparin and fentanyl were prescribed; together they gave rise to almost all the cases thus classified. With fentanyl the most frequent errors were omission of the pharmaceutical form, concentration and route of administration and hardly legible medication. For mydazolam the problems were principally related to omission of the concentration, doubtful concentration and hardly legible medication. There were a notable number of errors with the prescription of intravenous potassium chloride, such as doubtful infusion rate, hardly legible medication, doubtful concentration, lack of instruction on how to administer it, incomplete concentration, concentration omitted and route of administration it either hardly legible or doubtful.

 

 

A total of 23 decision-type prescription errors were observed, one being the pharmaceutical form, ten the concentration and 12 the route of administration. The most prevalent HAM was heparin, with 43.5% of the records. Only decision errors that could be identified from the prescription were computed.

In the 4,026 prescription order forms 70 different types abbreviation were recorded, for a total of 133,956 occurrences (an average of 33.3 per prescription). Of this total, 5,427 were classified as hardly legible. The abbreviations mg (milligram), h (hour), mL (milliliter), cp/comp/cp (c-omprimido [tablets]), EV/IV (endovenous/intravenous), gts (gotas [drops]), amp (ampoule), vo (orally), g/gr (gram), SF (soro fisiológico [saline solution]), min/X´ (minute), sc (sub-cutaneous), ABD (água bidestilada [bidistilled water]), S/N and SN (se necessário [if necessary]) and ACM (a critério médico[doctor's discretion]) represented 90% of them. The largest number of abbreviations observed in a single prescription was 211. The abbreviations UI or U were used 2,062 times in 1,971 prescriptions (48.9% of all prescription order forms).

The total number of prescription errors (writing and decision) was 13,387, with an average of 3.3 per prescription, 99.8% being writing errors. Writing errors amounted to 13,364 classified events: 3,154 (HAM errors), 1,894 (patient's name), 380 (prescription date), 775 (hardly legible or illegible prescription), 5427 (hardly legible abbreviations) and 1,734 (prescriber identification).

The univariate logistic regression analysis, shown in Table 4, reveals that the type of prescription has an influence on prescription errors (decision and writing. The type of prescription was tested: typed, handwritten and mixed. The typed prescription was considered as the reference point for comparison purposes. It was seen that the chance of errors in the handwritten prescription was approximately three times more frequent than in the typed prescription. The chance of errors was 2.5 times more frequent in the mixed prescription.

 

 

DISCUSSION

This investigation corroborates international evidence of the importance of HAM prescription errors in hospitals, whether from the point of view of their prevalence or of their potential risk for patients.2,8,14 However, some limitations must be pointed out. One has to do with the probability of mistakes when diagnosing decision errors, because an isolated analysis of prescriptions that does not consider the patient's clinical condition and without discussing the case with the prescriber does not allow some types of error to be evaluated. For this reason only the following were classified as decision errors: incorrect pharmaceutical form, concentration, route of administering the drug, frequency and infusion rate. So decision errors may not have been recorded, leading to data being underestimated.

Another relevant limitation is the subjective nature of the assessment of legibility. As a control, independent assessments were carried out by professionals with wide experience in reading prescriptions, decisions were all taken by a more experienced professional and if there was any doubt they sought consensus.

The third limitation is that only one variable was tested to explain the occurrence of prescribing errors. Other factors may have influenced the prevalence of the prescription errors detected, but they were not considered in this study, which focused on the way the medication was prescribed.

Problems associated with identifying the patient, the prescriber, and the date appeared in 47.0%, 33.7% and 9.4% of the prescriptions, respectively; these situations increase the probability of errors. Miasso & Cassiani18 (2000) identified that 33.9% of the errors in drug administration in a teaching hospital were problems with identifying the patient. As a legal document, the prescription should identify not only the patient but also the prescriber so that in situations where doubts need to be clarified it is possible to locate the person responsible for the prescription. In a hospital situation, where hundreds or even thousands of doses of medication are dispensed every day by the hospital pharmacy, omitting the date from the document may also lead to errors.

The difference in the occurrence of errors in the prescription of HAMs between sectors (Figure) is inversely proportional to the use of the typed prescription. This relationship reinforces the role of the type of prescription when it comes to determining errors and may be partly explained by the type of prescription used in the various sectors where patients are hospitalized.

Over the years, heparin has been associated with high error rates and is one of the ten drugs most frequently listed in error notifications in the United States where patients suffer harm. Between 1999 and 2002 heparin was very high on the register of serious errors (4.5% to 5.5%) and in 2002 it was responsible for 9.5% of the errors that caused harm in patients.19

In fentanyl prescriptions, the second drug most associated with errors (21.1%), the omission of the pharmaceutical form was the most frequent error. This is a mistake that may cause the exchange of one presentation for another. Fentanyl is a highly powerful opioid derivative and if used incorrectly it may cause the patient serious problems. The prescription of mydazolam where the concentration is omitted or the concentration leaves room for doubt is also relevant because it is frequently used in intensive therapy and it is essential that the correct dose is administered.

Errors in prescribing intravenous potassium chloride highlight the need for preventing serious accidents. A wrong dose or dilution of this drug, or the administration of the concentrated product, may have fatal consequences.2,8,13 In a study carried out in 24 intensive care units in the United Kingdom, in which 21,589 prescriptions were analyzed, the five drugs most associated with incorrect prescribing were potassium chloride (10.2%), heparin (5.3%), magnesium sulphate (5.2%), paracetamol (3.2%) and propophol (3.1%).20

Lack of standardization and the frequent use of abbreviations (33.3/prescription) show latent failures that may contribute to the occurrence of medication errors. The widespread use of the abbreviations UI or I may lead to serious errors, because of the possibility of them being confused with the number zero, leading to the administration of a concentration ten or 100 times higher than prescribed.1,2 The use of abbreviations in medical prescriptions is among the most mentioned causes of medication error because of their potential for confusion and communication breakdowns; the idea of banning their use is an old one. The Institute for Safe Medication Practices (ISMP) has been concerned with this problem for several years. The Joint Commission on Accreditation of Healthcare Organizations, an institution that certifies hospitals worldwide, prohibited the use of a list of abbreviations, among which are U and UI, in those hospitals that are candidates for accreditation.1 According to Koczmara et al11 (2005), some names should never be abbreviated due to the frequent mistakes this practice causes; among them are units, micrograms, sub-cutaneous and cubic centimeter.

A classic example of error in the use of the abbreviation UI was recorded in Canada, where a wrong dose of insulin was administered by mistake. The patient suffered permanent damage because he/she received 70 units of insulin instead of the seven prescribed, because the abbreviation U was confused with a 0.11

The high prevalence of writing errors in hospitals was also observed by Lisby et al15 (2005). These authors found that 75% of the mistakes found in prescriptions may be classified as writing errors. The most prevalent problems involved the pharmaceutical form, omission of the dose and route of administering the medication. In another study carried out in medical clinics in 60 health care units in the United States it was seen that non-clinical or writing errors were responsible for 79.1% of those recorded.7

The writing errors found in this research and related to HAMs could be minimized with standardization and the preparation of medication prescribing norms.2,17,22 The absence of a defined standard and the constant use of abbreviations in the prescriptions studied, such as UI meaning units, emphasize the need for standardization in the use of abbreviations in prescriptions. It is suggested that some abbreviations, like UI, be eliminated from prescriptions due to the potential for error associated with them.1

As for legibility problems, there would be a significant decrease if the type of prescription used in hospitals were standardized. Prescription errors can be reduced by using text editing programs or typed prescriptions, thereby avoiding the use of handwritten prescriptions as far as possible.22

A different approach must be used for reducing decision-type prescribing errors since measures for improving the prescriber's knowledge and intercepting these type of error are different from those suggested for writing errors.2 The importance of decision problems in prescriptions was seen in a study on antibiotics carried out in a Brazilian university hospital, where 91 incidents were recorded, of which 7.7% were considered to be prescribing and decision-type errors.16

According to Aspden et al2 (2007) the recommendations with the greatest scientific evidence that will lead to preventing medication errors in hospitals are the use of electronic prescriptions with appropriate clinical support, the inclusion of pharmacists on clinical rounds, making it possible for pharmacists to be contacted 24 hours a day for solving doubts with regard to drugs and the utilization of special procedures and written protocols for HAM use. This last measure can prevent a large number of the errors that occur with high-alert medications.2,8,13 However, research has to be done in Brazilian hospitals to check the true impact of the introduction of protocols in the use of HAMs and the prevention of events related to these drugs within the Brazilian context.

The introduction of the electronic prescription may have a major impact on prescription errors and attempts should be made to institute such a measure. Since its cost may be prohibitive for a proportion of Brazilian hospitals it is recommended that typed or edited prescriptions be adopted to avoid as far as is possible handwritten prescriptions.2 However, pre-typed prescriptions or the use of text editing programs are measures that need to be carefully prepared in order to avoid new types of error or the simple transposition of old problems to a new way of prescribing drugs.

Finally, when comparing the findings from this investigation with international literature we can see that medication errors involving HAMs tend to follow defined patterns, an important fact when it comes to taking decisions aimed at controlling them. However, all knowledge of this nature, when produced in diverse environments, implies the need for adapting it to the cultural reality of the location where intervention is planned and to the profile of the problems detected in each such location.

 

REFERENCES

1. Abushaiqa ME, Zaran FK, Bach DS, Smolarek RT, Farber MS. Educational interventions to reduce use of unsafe abbreviations. Am J Health Syst Pharm. 2007;64(11):1170-3. DOI: 10.2146/ajhp060173        [ Links ]

2. Aspden P, Wolcott J, Bootman JL, Cronenwett LR, Committee on Identifying and Preventing Medication Errors. Preventing medication errors. Quality Chasm Series (Hardcover). Washington: National Academies Press; 2007.         [ Links ]

3. Barber N, Rawlins M, Dean Franklin B. Reducing prescribing error: competence, control, and culture. Qual Saf Health Care. 2003;12(Suppl 1):i29-32. DOI: 10.1136/qhc.12.suppl_1.i29        [ Links ]

4. Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients: results of the Harvard Medical Practice Study I. N Engl J Med. 1991;324(6):370-6.         [ Links ]

5. Cohen MR, Proulx SM, Crawford, SY. Survey of hospital systems and common serious medication errors. J Healthc Risk Manag. 1998;18(1):16-27.         [ Links ]

6. Dean B, Barber N, Schachter V. What is prescribing error? Qual Health Care. 2000;9(4):232-7. DOI: 10.1136/qhc.9.4.232        [ Links ]

7. Devine EB, Wilson-Norton JL, Lawless NM, Hansen RN, Hazlet TK, Kelly K, et al. Characterization of prescribing errors in an internal medicine clinic. Am J Health Syst Pharm. 2007;64(10):1062-70. DOI: 10.2146/ajhp060125        [ Links ]

8. Federico F. Preventing harm from high-alert medications. Jt Comm J Qual Patient Saf. 2007;33(9):537-42.         [ Links ]

9. Fuchs FD, Wannmacher L. Farmacologia clínica: fundamentos da terapêutica racional. Rio de Janeiro: Guanabara Koogan; 1998.         [ Links ]

10. Kohn LT, Corrigan JM, Donaldson MS. To err is human: building a safer system. Washington: National Academy Press; 1999.         [ Links ]

11. Koczmara C, Jelincic V, Dueck C. Dangerous abbreviations: "U" can make a difference! Dynamics. 2005;16(3):11-5        [ Links ]

12. Leape LL, Brennan TA, Laird N, Lawthers AG, Localio AR, Barnes BA, et al. The nature of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med. 1991;324(6):377-84.         [ Links ]

13. Leape LL, Berwick DM, Bates DW. What practices will most improve safety? JAMA. 2002;288(4):501-7. DOI: 10.1001/jama.288.4.501        [ Links ]

14. Leape LL, Berwick DM. Five years to err is human. What have we learned? JAMA. 2005;293(19):2384-90. DOI: 10.1001/jama.293.19.2384        [ Links ]

15. Lisby M, Nielsen LP, Mainz J. Errors in the medication process: frequency, type, and potential. Int J Qual Health Care. 2005;17(1):15-22. DOI: 10.1093/intqhc/mzi015        [ Links ]

16. Louro E, Romano-Lieber NS, Ribeiro E. Eventos adversos a antibióticos em pacientes internados em um hospital universitário. Rev Saude Publica. 2007; 41(6):1042-8.         [ Links ]

17. Meyer TA. Improving the quality of the order-writing process for inpatient orders and outpatient patients. Am J Health Syst Pharm. 2000;57(Suppl 4):S18-S22.         [ Links ]

18. Miasso AI, Cassiani SHB Erros na administração de medicamentos: divulgação de conhecimentos e identificação do paciente como aspectos relevantes. Rev Esc Enferm USP. 2000;34(1):16-25.         [ Links ]

19. Niccolai CS, Hicks RW, Oertel L, Francis JL. Heparin Consensus Group. Unfractionated heparin: focus on a high-alert drug. Pharmacotherapy. 2004;24(8 Pt 2):146S-155S. DOI: 10.1592/phco.24.12.146S.36107        [ Links ]

20. Ridley SA, Booth SA, Thompson CM, Intensive Care Society's Working Group on Adverse Incidents. Prescription errors in UK critical care units. Anaesthesia. 2004;59(12):1193-200. DOI: 10.1111/j.1365-2044.2004.03969.x        [ Links ]

21. Rosa MB, Perini E. Erros de medicação: quem foi? Rev Assoc Med Bras. 2003;49(3):335-41. DOI: 10.1590/S0104-42302003000300041        [ Links ]

22. Tran M. The impact of introducing pré-printed chemotherapy medication charts to a day chemotherapy unit. J Oncol Pharm Pract. 2000;6(2):64-9. DOI: 10.1177/107815520000600206        [ Links ]

23. World Health Organization. World alliance for patient safety: forward programme 2006-2007. Geneva; 2006 [citado 2007 dez 29]. Disponível em: http://www.who.int/patientsafety/en        [ Links ]

24. Yu KH, Nation RL, Dooley MJ. Multiplicity of medication safety terms, definitions and functional meanings: when is enough enough? Qual Saf Health Care. 2005;14(5):358-63. DOI: 10.1136/qshc.2005.014159        [ Links ]

 

 

Correspondence:
Faculdade de Farmácia - CEMED
Universidade Federal de Minas Gerais
Av. Antônio Carlos 6627, sala 3111
31270-901 Belo Horizonte, MG, Brasil
E-mail: mariobr_ca@yahoo.com

Received: 2/7/2008
Revised: 7/9/2008
Approved: 11/3/2008
Research partially funded by Fundação Hospitalar do Estado de Minas Gerais.

 

 

Article based on the masters dissertation by Rosa MB, presented to the Universidade Federal de Minas Gerais, Escola de Veterinária, in 2002.