Sharing evidence to reduce errors in medicines management

Contents.
Introduction.
Prevalence of medication errors.
How pharmacies deal with errors.
Why argue for a single check.
Other solutions. References. 

Back to Nursing Progress contents.
Back to Issue 11 contents list.

The view that single checking is safer than providing a double check is commonly held. Some observers have even proposed mathematical models to support this assertion. However, we have been unable to find reliable evidence to support the improved safety of a single check. The theory that second checking systems are less accurate is counter-intuitive and therefore it is intriguing to understand how this view has taken hold. It may have been convenient to support the assertion of a single check as second checking is time consuming and nursing time is stretched by lack of staff and extended roles. It may also have fallen on fertile ground because initially, assertions for a single check were made at a time where nursing was promoting its professionalism and accepting new responsibilities and independence. We propose that the argument for a single check is counter-intuitive, that the change to single checking should only have taken place with the support of good evidence, and that where resource constraints forced such a change that all high risk drugs should have remained on second checking. Indeed, whilst the single check concept is generally accepted throughout the Trust, second checking for chemotherapy is still the accepted procedure.

Back to contents

Prevalence of medication errors

How prevalent are medication errors? Ferner et al, 2000 suggest that up to 4 per cent of hospital inpatients will experience an adverse event resulting from a medication error related to prescribing, dispensing, or administration events, so it is imperative to develop systems and practices to reduce them. (1) The accepted administration error rate in hospitals is 5.5%, which has been shown across a range of studies. (2) A study of dispensing errors by Kayne in 1996 reported the undetected error rate in hospital is reported as being between 16 and 18 items per 100,000 dispensed items and that a hospital pharmacist makes one error that leaves a pharmacy department every nine months. (3)

Many reasons for dispensing errors have been proposed, including poor illumination. However, many of the reasons for dispensing errors remain anecdotal; noise, distractions, poor packaging and workload have all been put forward as causes of errors. Kistner and co-workers  found no correlation between the number of prescriptions dispensed per hour and the total number of errors made, although error peaks were noted during the lunch hour and in the hour before closing. (4)

Back to contents

How pharmacies deal with errors

In pharmacy, dispensing errors have historically been dealt with by focusing on the errors of individuals viewed as "free agents capable of choosing safe and unsafe modes of behaviour". (5) In addition, separating an individual person's error from any institutional responsibility is legally convenient and, by focusing on an individual, isolates the dispensing error from the dispensing processes in use. A system approach of dealing with errors places emphasis on the working conditions and builds defenses that avert errors or mitigate their effect. Many errors are blameless and effective risk management requires a reporting culture. These errors, mishaps and near misses allow recurrent error traps to be uncovered. In practice, it is best to address errors that occur as a result of both human-based and system-based errors, through a greater understanding of the processes involved in dispensing.

Research undertaken to reduce and monitor dispensing error rates in hospital, found that errors occur with all types of checking systems, including the system where a pharmacist was checked by a second pharmacist. (6) However, the study found that 57.1% of the external errors had been dispensed using a system where the pharmacists checked themselves. The authors suggested that single checking is more likely to result in errors. This is verified by a multicentre study which showed that overall, 18.1 errors were reported per 100,000 dispensed items (0.018%) and that this was significantly lower in hospitals where the dispensing of pharmacists and technicians was checked. (7) In pharmacy literature, there is not one study that suggests that a single check is safer in the dispensing process.

Back to contents

Why argue for a single check

On the surface, administration errors appear to be irrespective of knowledge errors but from our experience we would argue that this is not the case. One study was designed to compare error rates between electronic and manual systems and showed a large disparity between knowledge errors at Kings College Hospital (which had undergone extensive knowledge training) and a comparator hospital.(8) One thing is imperative; that where a second check is carried out, it should be a full second check. This is probably most important in calculations where the full calculation should be repeated. In pharmacy, the process of second checking does not allow the first dispenser to feel "safe" within the process; the second check is a full check where the original container and prescription is checked against the label and dispensed item, rather than checking the first dispenser. It can be argued that the first check picks up system errors whereas the second check picks up human errors (slips and lapses) such as errors due to knowledge (calculations etc). Therefore, in pharmacy, the second check is regarded as the final step in the process to reduce all errors.

Back to contents

Other solutions

Implementing second checking systems within limited resources is difficult, but the consequences of error probably outweigh the investment in safer systems. Such systems include electronic prescribing, bar coding both drugs and patients and the use of automated labeling. Implementation of automated check-and-sort devices reduce dispensing errors and give pharmacists more time to review patient profiles and recommend clinical interventions. One recently published study demonstrated a reduction in dispensing errors from 6.3 to 4.1 per 100,000 medication orders filled. (9) We are currently piloting electronic prescribing within the Trust and the in patient pharmacy at the Royal London Hospital has installed an automated picking system which has produced an error rate of only 4 per 100,000 items at other sites.

Back to contents

This article has highlighted some of the evidence for second checking within the dispensing process and defended the need for increased checks to reduce errors in medicines management. Of course, it is not simply about disseminating findings but working together to ensure that such sharing is effective in practice. By sharing such knowledge, we are effectively adhering to the recommendations of Organisation with a Memory (2000) to best use available information for learning within and between professions to improve the standard of healthcare delivered throughout the NHS.(10) (11) The Trust has recently funded a multidisciplinary medication risk team to help reduce errors. It will work alongside existing members of nursing and Pharmacy to help make our Trust a safer place to receive drugs.

Back to contents

1. Ferner RE, Aronson JK. Medication errors, worse than a crime. Lancet 2000;355:947-8.

Back to Nursing Progress contents

Back to Issue 11 contents list

Back to Top

Page last revised by DEB on 30/4/02

[RLH League]

[League Notices]

[Membership]

[League history]

[Education]

[Hospital History]