MEDICATION DISPENSING ERRORS AT A PUBLIC PEDIATRIC HOSPITAL

Avaliar a segurança na dispensação de medicamentos através da determinação da taxa de erros de dispensação constituiu o objetivo deste trabalho. O método utilizado foi o estudo transversal que avaliou 2 620 doses de medicamentos dispensados entre agosto e setembro de 2006, em um serviço de farmácia de um hospital pediátrico do Espírito Santo. Os erros de dispensação foram definidos como qualquer desvio entre o dispensado e o prescrito na receita médica. Os erros foram categorizados em conteúdo, rotulagem e documentação. A taxa de erro de dispensação foi calculada dividindo o número de erros total/total de doses dispensadas. Os resultados mostraram que, dos 300 erros identificados, 262 (87,3%) foram de conteúdo. Nas categorias erros de rotulagem a taxa foi de 33 (11%) e 5 (1,7%) na de erros de documentação. Concluiu-se que a taxa total de erros de dispensação foi elevada quando comparada à descrita em estudos internacionais. A categoria de erro mais freqüente foi a de erro de conteúdo.


INTRODUCTION
Medication error is a difficult topic to address.
Discussions about it are generally directed at seeking the culprit and do not provide system improvement opportunities with a view to preventing failures (1) .Some authors suggest that the occurrence of errors in different social and professional systems might originate in system failures (systemic errors).
Medication errors are considered human errors and, thus, can be caused by these failures (2)(3)(4) .They represent a severe social and health problem with important economic repercussions (3) and are classified as medication prescription, dispensing and administration errors.Dispensing errors occur during the medication dispensing process (4)(5)(6) .
Literature suggests that the incidence of medication errors in pediatrics is twice or three times as high as in adults, and also that pediatric patients are at higher risk of death when compared to adults (7- 11) .These errors also represent an important economic cost and measures to reduce and prevent them are necessary (2,10) .Discovering their frequency and defining ways to prevent them is an important strategy to reduce risks, especially in special populations like children.The Brazilian Health Surveillance Agency (ANVISA) recently included medication errors in its pharmacosurveillance program as a strategic area for patient safety in the process of medication use (4) .
The pharmacy service is responsible for the safe and efficient use of medication in hospitals and plays an essential role in integrating the prescription, dispensing and administration processes and should have policies and procedures to prevent errors (4-5,8)   .
The rate of errors is considered one of the best indicators of quality of medication distribution systems and is still used to evaluate the safety of these systems (4,6,9) .Some studies, published in the United States and England, showed an incidence of dispensing errors of about 10%, even in hospitals with advanced medication distribution systems, such as unit doses (7,11) .
A few studies were carried out in Brazil, specifically on medication dispensing processes, and presented high rates of errors, above 10% (5,8)   .In the pediatric area, especially in Brazil, no study was found on dispensing errors and, to date, their cause and epidemiology are unknown.
Therefore, this study aims to evaluate the dispensing process of the pharmacy service at a public pediatric hospital.Important indicators were used to measure the dispensing process and the rate of total dispensing errors was the global indicator of its quality.

Variables Operationalization
Errors were classified according to the criteria used in a previous study (10) in:

Content errors
Incorrect medication -Drug dispensed differs from the one prescribed.Excluding therapeutic substitution of medication due to hospital standards or procedures; Incorrect concentration -dispensing drug with correct quantity of medication (Mg or mL) but with incorrect adjustment of dosing instructions; Incorrect dosage form -Dispensing correct medication but in a dosage form different from that prescribed.
It includes providing a modified release formulation when a standard formulation was prescribed; Dose added -Dispensing a larger quantity of medication (in number, units, or times a day) than that prescribed.
Missing doses -Dispensing a quantity smaller (in number, units, or times a day) than that prescribed.Additional warning -Omission or incorrect use of warnings according to the bibliographic references.

Pharmacy address -Fail to include the correct pharmacy address on the label;
Other labeling errors -Any labeling error not included in the previous categories; for instance, illegible name or number.

Absent or incorrect controlled medication documentation -Absent or incorrect documentation of controlled drug registration according to law;
Other documentation errors -any documentation errors not included in the category above.

Statistical analysis
Statistical Package for the Social Sciences (SPSS Chicago -IL, version 9.0, 1998) was used to build the database and make statistical calculations.Categorical variables were expressed as proportions (relative frequency).

Ethical aspects
This study was approved by the hospital direction.The pharmacists and pharmacy technicians filled out the informed consent term, allowing the observation of dispensing activities, and were informed about the study aims to evaluate the dispensing system.They were not aware of the objectives and method used; these aspects were kept blind to the pharmacy team to avoid known biases (6- 7) .All errors that occurred during the study were codified and kept confidential.They were corrected only after the last control performed by the pharmacist-researcher (7) .

RESULTS
During the study, 239 prescriptions were evaluated and included 655 prescribed medications, totaling 2620 dispensed doses.
The total rate of errors including the three criteria (content, labeling and documentation errors) was 11.5% (300 errors/2620 doses).Table 1 shows the frequency of dispensing errors in each of the categories.The content category presented the highest rate of error, followed by labeling and documentation errors."content errors".The most frequent errors in this category were "too high doses", "missing doses" and omission errors; incorrect dosage and deteriorate medication errors were not registered.On the other hand, the most common errors in "labeling errors" were the subcategory "other labeling errors" and "incorrect dosage", as shown in Table 3.There was no occurrence for the items: medication name, quantity, pharmaceutical presentation, date, instructions, warning, and incorrect pharmacy address.The total error rate of the category "labeling error" was 1.7%.The most frequent errors in this category were: "other labeling errors" with 75% and "incorrect dosage" with 21.2%.(Table 3).

DISCUSSION
The total dispensing error rate, according to the adopted classification system (content, labeling, and documentation errors), was 11.5% for the total of dispensed doses in the study period.This rate represents one error for approximately nine dispensed doses.Data analysis showed a much higher frequency in the criterion "content errors" 87.3% (262/2620).This higher predominance is due to the fact that errors occur more often during the dispensing process itself than during the labeling and documentation of this activity (3) .
When the criterion "content errors" is separately analyzed, the most frequent errors were "too high doses" and "missing doses" with 49.6% and 28.6%, respectively.These data are in accordance with other studies that appoint that dosage errors are the most frequent in pediatrics.Considering only the contribution of this category (content errors) in the total error rate by dispensed doses, an error rate of 10% is found, that is, almost the total error rate found in the sample.
The categories 'labeling errors' and 'documentation errors' represented 11% and 1.7% of the errors, respectively.In the category 'documentation errors', the absence of documentation was the most common error.There are different methods to evaluate dispensing errors, which makes it difficult to compare results between different studies (14) .In the study sample, the total error rate was considered high when compared with other studies using a similar method (4)(5)10) . The auses of errors presented in this study can be of several origins, which require a deeper evaluation with qualitative studies, though some critical points can be raised as possible causes.An important point observed is that, most of the times, the pharmacist did not check doses prepared by pharmacy technicians.Despite the adequate rate of pharmacists per number of beds (30:1), it was verified that the pharmacist's actions are focused on the administrative aspect of the dispensing process, and not on its care.There is strong evidence that the distribution system of unit doses can reduce the error rate and increase safety in the medication use process (4)(5)(6)(7)(8) .
Another interesting aspect is that part of the doses is dispensed in 'unit doses', that is, they are ready for use, with no need for manipulation before administering medication, which represents an improvement.However, the adopted system does not allow for a reduction in dispensing errors because there are no rigid controls.The majority of errors can be avoided if a distribution system concentrating the process of dosage preparation in the pharmacy service is in place and the pharmacist checks the prescription before it is dispensed (4)(5) .On the other hand, effective interaction between the nursing and the pharmacy services is essential because many errors that occur during the dispensing process can be avoided at the moment the medication is administered by the nurse.A multicenter study involving four hospitals in different regions of the country identified high error rates during medication preparation and administration.Authors suggest that, to improve safety of medication distribution systems, changes need to be adopted in the institutional culture with a view to solid improvements (9) .
Results of two Brazilian studies on dispensing errors in adult hospitals showed very high error rates (13.8% and 17%), although pharmacists inspected the doses prepared by auxiliaries in both studies (5,8) .
However, one has to be careful in making comparisons between these two studies because there is an important methodological difference in error classification (7,10) .
This study presents some limitations, the main of which is that one cannot generalize its results to other hospitals of the same size and specialty, because there are other important variables that can influence the dispensing error rate.Another issue is that a representative sample of Brazilian public hospitals was not used, considering number and type of clinical units as well as their complexity level (4) .

CONCLUSION
The total dispensing error rate in the study sample is high when compared to international studies (12)(13)15) . The ost frequent error category was the "content error", while "missing doses" and "wrong dose" were the most prevalent in this category.The categories "labeling error" and "documentation error" represented a small influence on the total error rate.
Further research is necessary to evaluate this issue, not only on medication dispensing but also on administration and prescription.

Table 1 -
Distribution of errors by category at a

Table 2
presents the results for the category

Table 2 -
Distribution of errors according to the category "content errors" at a pediatric hospital in

Table 3 -
Distribution of errors according to the