The Pharmaceutical Journal
Vol 270 No 7233 p130-131
25 January 2003

This article
Reprint
Photocopy

American Society of Health-system Pharmacists summary

A strong focus on safe systems for medicines management

Certain types of medication error are more likely to cause harm or death to patients, according to a report from the United States Pharmacopeia's Centre for the Advancement of Patient Safety (CAPS). Presenting the report, Diane Cousins, vice-president of CAPS, said that incorrect administration technique, such as the improper dilution of IV products was four times more likely to cause harm in hospital patients.

The report, entitled "Summary of information submitted to Medmarx in the year 2001: a human factors approach to medication errors", is the third annual report produced by Medmarx. Medmarx, founded in 1998, is the anonymous medication error reporting programme and quality improvement programme run by the USP. The scheme provides comprehensive reports, trends and national alerts. With more than 550 participating hospitals it is able to provide benchmarking reports and help hospitals reach the required standards for JCAHO (Joint Commission for Accreditation of Hospital Organisations) audits.

Ms Cousins emphasised that the work of CAPS was a continuous cycle of collecting data, analysing, collaborating and educating. Much effort is devoted to teaching people to analyse their data, in contrast to other error reporting systems that tend to focus on errors detected and actions taken, she said.

The 2001 report covered 368 hospitals nationwide and a total of 105,603 medication error records. There were 2,539 harmful errors (2.4 per cent), of which 353 required initial or prolonged admission to hospital, 70 required interventions to sustain life and 14 resulted in death. The remainder were non-harmful actual errors (86.9 per cent) or potential errors (10.7 per cent). More fatal errors were counted than in the previous year but this was almost certainly due to improved reporting, said Ms Cousins.

Turning to the phases of the medication use process at which errors occurred, Ms Cousins said that 37 per cent of reported errors occurred at the administration phase. Documentation accounted for 26 per cent of errors, dispensing 21 per cent, prescribing 15 per cent and monitoring 1 per cent. Reporting of administration and dispensing errors is well developed, noted Ms Cousins, and reporting of prescribing errors has improved since the 2001 survey was completed.

The types of error that occurred most frequently were errors of omission (29 per cent), improper dose or quantity (21 per cent), prescribing errors (14 per cent) and unauthorised drug supply (13 per cent). Together these accounted for 77 per cent of errors, although they were not necessarily most likely to cause harm. When the "harmful" category of errors was further analysed, four types error accounted for a disproportionate number of harmful outcomes: they were wrong drug preparation (2.7 per cent), wrong route (2.8 per cent), improper dose or quantity (3.3 per cent) and wrong administration technique (8.0 per cent). Errors of omission (the most common type of error) were least likely to cause harm.

The top two causes of errors were performance deficits (38 per cent) and failure to follow the agreed protocol (20 per cent). A performance deficit described the situation where an individual had the required knowledge or skills but simply failed to perform. Computer entry errors, either computerised prescribing or transcription of a handwritten order, accounted for 11 per cent of errors.

Contributory factors to errors included distractions (47 per cent) and increasing workload (24 per cent). "A very good individual performing in a stressed situation can still make mistakes," noted Ms Cousins. Staffing issues — mainly inexperienced staff and insufficient staff — were implicated in 43 per cent of errors, and this was an increase compared with previous reports, she added.

The drugs most commonly involved in harmful medication errors were insulins, morphine, heparin, warfarin, potassium chloride, furosemide, vancomycin, hydromorphone, meperidine (pethidine), and diltiazem. The use of these products often involved complex protocols, she said.

Emergency department and paediatric medicine departments deserved special attention. Emergency departments are organised to deliver prompt, life-sustaining care in dynamic situations and therefore they function differently from other patient care areas. In emergency departments more errors reached patients than elsewhere — in this survey 7.6 per cent of the errors in this area resulted in patient harm. The errors arose predominantly from prescribing or administration mistakes. One of the reasons for this was that the usual pharmacy oversight of prescriptions was absent and so the opportunity to intercept errors was missing, she suggested.

In paediatric medicine improper dosing, omission errors and incorrect timing accounted for the majority of errors reported. The lack of manufacturer-prepared doses was a contributory factor but organisational issues, such as the frequent absence of a procedural safety net and the use of verbal medication orders, also played significant roles.

Ms Cousins concluded that although many of these findings were not new, the size of the database from which they were derived gave them a weight that could not be ignored.

She hoped that health care practitioners would see the report as a call to action. Sharing of data and experience is essential, she said, as is federal protection of reported information, otherwise professionals will be reluctant to report incidents.

Only 500 hospitals out of a possible 5,200 are currently using the Medmarx system. Many pharmacists would like to use it but their hospital risk managers have other systems in use. In order to make the process easier the latest release of Medmarx software has made it possible to upload data from any other system.

In addition to implementing the Medmarx reporting programme, the USP suggests that hospitals should also adopt a number of other measures to help prevent medication errors.

First, it recommends the implementation of standardised procedures and strict protocols for the dispensing and administration of high alert medications

The second recommendation is to purchase pre-mixed intravenous solutions and unit dose medication. Thirdly it suggests automation of the medication process through the use of advanced technology, such as medication bar-coding systems, wristband scanning devices, patient photo-identification on wristbands, computerised medication administration records, or a combination of any of these.