One in 15 acute medical admissions is medication-related but few of these are due to GP error, says Tony Avery.
"First do no harm" is one of the most important ethical principles in medicine, and it is one that as a general practitioner I try to keep firmly in mind. It is unfortunate, therefore, that most of the medicines we prescribe can cause harm and that patients are at risk from errors in the prescribing and monitoring of drugs.
Over the past ten to 15 years, much research has been done to describe the extent of medical error and to investigate why errors occur. The vast majority of studies have been done in hospital settings, and these suggest that most errors are caused by failures in safety systems rather than gross neglect or incompetence on the part of doctors. It is generally accepted that the best way to reduce errors is to design better systems so that doctors are less likely to make mistakes.
However, relatively little research has been done in primary care, where most patient contacts take place. Our research aims to describe the extent of medication-related errors in primary care, to understand why these errors occur and to determine how they might be prevented.
One of the areas that we have focused on is medication-related hospital admissions because these provide examples of some of the most serious errors associated with GP prescribing. Studies have shown that about one in 15 acute medical admissions is medication-related and that about half of these are preventable. We have investigated more than 4,000 hospital admissions and have found that, in contrast to hospital practice, very few errors are the result of slips such as giving the wrong drug or the wrong dose. Instead, the major problems relate to balancing risks and benefits at the time that medication is prescribed, and failure to ensure that people on long-term medication are adequately reviewed so that any problems can be picked up early.
In another study, we looked at the prescription of potentially hazardous combinations of drugs in general practice. We found a low error rate, but it is clear that GPs sometimes inappropriately override hazard messages that come up on their computer systems. Although GPs say they value hazard-alert systems, they often receive messages that they believe are irrelevant. This may lead to a "cry-wolf" situation where an important alert is overridden without the GP's checking it properly. To minimise the risks of prescribing potentially hazardous combinations of drugs, attention needs to be paid to the design of hazard-alert systems so that they give GPs relevant information, but also act to prevent the most dangerous combinations of drugs being issued.
Other research includes a survey of the errors pharmacists pick up on GP prescriptions; an investigation of how GP computer systems can help identify patients at risk or those suffering from medication-related problems; and a systematic review of the literature on interventions that have been used to reduce medication-related error in primary care.
We have identified a range of approaches that might be used to reduce the number of such serious medication errors. One of these involves patients in decisions about their medicine-taking. Others include better links with community pharmacists, more involvement of pharmacists in medicine-management systems and more effective use of computers to detect errors. We will now look at how these approaches might work in practice.
Tony Avery is professor of primary healthcare and head of the division of general practice at the University of Nottingham. Researchers involved in this programme come from the universities of Nottingham, Manchester and Florida and Imperial College School of Medicine.