This site is intended for health professionals only!

Published on 1 July 2004

Share this story:

Global and national monitoring systems


Laurence A Goldberg
Consultant Pharmacist

The 2nd American College of Clinical Pharmacy (ACCP)–European Society of Clinical Pharmacy (ESCP) joint meeting, which was held in Paris (28–30 April 2004), attracted 1,700 participants from Europe and North America, with the central theme of the meeting being the optimisation of outcomes in pharmacotherapy.

WHO Collaborating Centre for International Drug Monitoring
The WHO Collaborating Centre for International Drug Monitoring (also know as the Uppsala Monitoring Centre; see Resources) now holds a database of more than three million adverse drug reaction (ADR) reports from 72 countries, Mr Sten Olsson (Head of External Affairs, Uppsala Monitoring Centre) told the audience in a session on global monitoring systems. The main contributors to the database are the USA and the UK (46% and 13% of reports, respectively). The main function of the database is to identify signals of ADRs as early as possible. The centre uses a three-step approach that involves screening of the data using a data-mining tool – the Bayesian Confidence Propagation Neural Network (BCPNN) – followed by filtering and assessment by clinical experts. The data-mining tool allows an objective assessment of all drug–ADR combinations. It generates a statistic called the information component (IC) that is a measure of the relationship between a drug and an ADR. A positive IC value indicates that a particular drug–ADR combination is reported to the database more often than expected from the rest of the reports in the database. An example of this was the link between captopril and coughing. The IC value started at zero but quickly rose above the threshold level of two standard deviations, indicating an association between the drug and the ADR.

The participating centres send reports to the database daily, but the BCPNN is run quarterly. The list generated is reviewed, and a shortlist of problems for follow-up is compiled. Items of particular interest are serious reactions, new drugs, reports of similar reactions that have been received from more than two centres, reactions that have not previously been reported, and special interest items such as blood dyscrasias. If, after review by the expert panel, the evidence is felt to represent a true signal, then a short signal report is written and sent to the national centres and to the manufacturer, if a patent medicine is involved.

Programme members can undertake web-based searches of the database. Customised searches are also available on a fee-paying basis. The centre always reminds users that it is not possible to compare frequencies of ADRs on the basis of spontaneous reports.

Asked about self-reporting of ADRs by patients, Mr Olsson said that Medwatch (see Resources) permitted reporting by patients but that other programmes were very suspicious about it. He said that 90% of healthcare professionals do not report ADRs, and he believes that consumer reports have a place. Another member of the audience asked whether the database held information about ADRs involving vaccines and traditional medicines. Mr Olsson said that there are several hundred reports involving vaccines, although some participating centres hold vaccines data separately and do not routinely submit them to the monitoring centre. In addition, there are 11–12,000 case reports in which herbal products are suspected as the cause of the ADR. There is a major problem with nomenclature: for example, 18 different plants are called ginseng.

A spontaneous report of a single case of ADR can be important in understanding the profile of a drug, according to Dr Stephen Goldman (former Medical Director, Medwatch, US Food and Drug Administration [FDA]). Postmarketing surveillance is essential to identify serious but rare adverse reactions, because most of these cannot be known at the time that a product is marketed. Moreover, there are still some drugs on the US market that were introduced before 1938 and did not go through the present-day new drug application (NDA) process. Three- thousand patients would be needed to have a 95% chance of identifying an ADR that occurs at a frequency of one in a thousand, he pointed out. Therefore, epidemiological methods are needed to identify signals of ADRs. Spontaneous reports, defined as any observations that originate outside formal studies, are also important.

Underreporting of adverse reactions always occurs, said Dr Goldman. For example, it has been estimated that in Norway, there was 90% underreporting of venous thromboembolism with oral contraceptives. Furthermore, reporting is biased because spontaneous reports arise in uncontrolled conditions and the number of patients exposed to a drug is unknown. There are, therefore, problems with both the numerator and denominator, and a percentage incidence cannot be calculated. It is critical that ADR reports are complete and that the clinical assessment of a report is made by an experienced individual.

Turning to the question of how many cases make a signal, Dr Goldman said that three or more was the rule of thumb. However, one good case can still make a difference. Regulators such as the FDA must always err on the side of public health. One example is that of felbamate, which was linked to nine reports of aplastic anaemia within one year of its launch. The background rate of aplastic anaemia is between two and five cases per million people per year. In this case, the drug was left on the market but assigned a “black box” to indicate that it should be used only when other agents are not suitable. Another example was fluoxetine. A 20-year-old woman had been taking fluoxetine for two years when it was discontinued and replaced with tranylcypromine. The patient rapidly developed what is now described as “serotonin syndrome” and, on investigation, was found to have therapeutic levels of fluoxetine five weeks after it had been discontinued. The product datasheet was changed on the basis of this single, serious report. Postmarketing surveillance is an important tool for the detection of ADR signals, Dr Goldman concluded.

National Patient Safety Agency (NPSA)
Patient safety is the freedom from accidental injury in healthcare, said Professor David Cousins (Head of Safe Medication Practice, National Patient Safety Agency, UK). The National Patient Safety Agency (NPSA; see Resources) was set up to collect and analyse information on adverse events in the National Health Service, learn the lessons and produce solutions to prevent harm. The underlying idea was to mimic the situation in aviation, where increasing numbers of reports had been associated with a decreasing number of serious events.

The NPSA believes that patients and the general public should be involved in these processes, and, to that end, it has organised a number of consultation exercises such as focus groups and face-to-face interviews. One example is oral methotrexate. Between 1993 and 2002, there were 137 patient safety incidents involving oral methotrexate. The fundamental problems were confusion between daily and weekly dosing, failure to monitor treatment closely and dispensing errors.

The NPSA set up groups that included patients and representatives of the pharmaceutical industry. A set of solutions was developed with the patients. These included reinforcing information about weekly dosing, regular blood monitoring and working with the pharmaceutical industry to improve packaging. It was frustrating to find that one manufacturer is still producing a calendar pack for methotrexate that has 28 doses, Professor Cousins commented. Although 95% of GPs in the UK have electronic prescribing, the interaction databases sometimes give clinically irrelevant information, and doctors have learned to ignore the warnings – so a new routine has been developed for methotrexate.

Clinical pharmacists can contribute to patient safety by following the guidance provided in the document Seven Steps to Patient Safety, published by the NPSA, suggested Professor Cousins. “Don’t find out about the risks after they have occurred – risk-assess the whole process in advance, for example when a new product, procedure or medical device is introduced,” he recommended.

Responding to questions from the audience, Professor Cousins said that it was critical to learn from others’ mistakes and that pharmacists could be the driving force in this case. Another aspect of safety is the purchasing of drugs – pharmacists should lead the way in refusing to purchase products that are unsafe in use, he suggested.

Dr Patrick Ball (University of Auckland, New Zealand) added that simply entering something onto a computer does not make it into a computerised system or a safe system. A survey in New Zealand has shown that 75% of computer-generated referral letters to hospitals contained erroneous information.

Institute for Safe Medication Practice (ISMP)
The mission of the Institute for Safe Medication Practice (ISMP; see Resources) is to encourage voluntary reporting of medication errors and translate errors into education, said Mr Michael Cohen (President, Institute for Safe Medication Practice, Huntingdon Valley, PA, USA).

Reports to the ISMP can be filed electronically or by telephone. In addition to actual incidents, the database also accepts near-misses and hazard reports. It works well because it is driven by altruism. ISMP always respects the confidentiality of the reporter and the location. To do otherwise would undermine the programme, said Mr Cohen.

The ISMP has been successful in persuading the pharmaceutical industry to alter labelling and change packaging to improve patient safety.

Considering the nature of medication errors, Mr Cohen said that ISMP has yet to process a report where a single thing went wrong – there are invariably multiple errors. It is usually the system that is at fault, rather than the people. Moreover, it is not just errors at the patient/nurse interface; often, the root cause is an error further back in the system that “set people up to make errors at the sharp end”.

Lack of adequate patient information accounts for a significant number of errors: for example, inadequate allergy documentation. Doctors and nurses not knowing enough (about medicines) accounts for almost half of all medication errors. High-alert drugs and activities such as concentrated potassium chloride injection, paralysing agents and free flow with infusion pumps also account for a significant number of medication errors.

WHO Collaborating Centre for International Drug Monitoring
National Patient Safety Agency
Institute for Safe Medication Practice (ISMP)

Most read

Latest Issue

Be in the know
Subscribe to Hospital Pharmacy Europe newsletter and magazine
Share this story: