Unscripted errors http://www.theaustralian.news.com.au/story/0,20867,19106690-23289,00.html
Computers were supposed to make prescribing safer, but they may have simply brought a new set of safety headaches, writes Kellie Bisset May 13, 2006 YOU probably haven't heard a joke about doctors' handwriting for a while. That's because electronic prescribing with its clear computer printouts has largely replaced the indecipherable scrawl that doctors used to write on prescription pads. As a result, the chances of getting the wrong drug when you go to the pharmacy have been virtually eliminated - a shift that some have described as a revolution. Today, about 90 per cent of family doctors have a computer and most use it to prescribe electronically. Hospitals are also using electronic medication management systems. Such technology has been hailed as the potential saviour in a healthcare system where preventable drug errors have been estimated to cause about 4000 patient deaths a year and cost $1.5 billion. There's an impressive weight of evidence showing that computerised prescribing systems can significantly lower medication mistakes and adverse drug events. One study in the Journal of the American Medical Informatics Association showed drug mistakes were slashed by more than 80 per cent after a system was developed in one US hospital (1999;6(4):313-321). But signs are now emerging that health technology, far from being a panacea, is causing a whole new set of problems, including potential risks to patient safety. "It is like any drug: it is going to look after people, but there might be some side effects," says professor Enrico Coiera, director of the Centre for Health Informatics at the University of NSW. While he believes the benefits of health technology enormously outweigh the costs, that doesn't stop Coiera worrying about the potential for harm, which he says is very real and something we should take seriously. "The airline industry only got safety after lots of disasters," he says. "I would like to see the e-health industry put safety at the top from the beginning, so we avoid disasters." The new breed of errors includes doctors printing out prescriptions for the wrong person, after forgetting to close a patient's file at the end of a consultation and before beginning another. And it's inevitable these sorts of mistakes will end up in court, says Therese Carroll, clinical risk manager for the Medical Defence Association of Victoria. "We have not had any claims to that effect yet, however with busy practices it is probably only a matter of time." Associate professor Helena Britt says we need to fund research on the extent of this potentially large problem. As director of the Australian General Practice Statistics and Classification Centre, Britt has already been involved in landmark research on errors made in general practice, but she says we need more information on how technology contributes to mistakes. "At the moment we are poking in the dark and with 95 million [general practice] consultations a year we have plenty of dark to poke in." Both Britt and Carroll are particularly concerned that many GPs store some patient information on paper and some electronically. If the electronic record is incomplete, software warning systems that alert doctors about serious interactions between a patient's medications can't work properly. Incomplete records are also an issue in hospitals, which often record patient medication histories "appallingly badly", according to pharmacist and patient safety researcher Ian Coombes. "If you don't have your medication history and baseline right, it does not matter how good your prescribing system is - you are probably prescribing the wrong drugs." The Royal Australian College of General Practitioners (RACGP) is trying to tackle the issue in general practice with new standards requiring doctors to store key information, such as relevant allergies and drugs, in a minimum percentage of electronic patient records. And while this might encourage more complete e-records, it doesn't address the reason many doctors aren't keen on them: paper can be quicker. The RACGP has raised concerns over whether time-poor doctors understand how to get the most out of the e-health system sitting on their desks. College president professor Michael Kidd says we need to spend more money training GPs how to use the safety features in their medical software systems. "It is critical that we continue to focus on how this new technology can be used to improve the safety and quality of what we do." Most medical software contains a basic electronic decision-support system, which gives prompts or warnings on drug interactions. Decision support can also be used to alert doctors to drug-disease interactions, drug effects in pregnant women, the elderly, and in athletes, depending on what software is used. Advanced support systems can calculate risk profiles for particular patients and link doctors to the latest clinical evidence. But evidence suggests that if it's not designed well, decision-support software can annoy doctors, leading them to ignore computer warnings (Medical Journal of Australia 2003;179:34-37), or turn them off altogether. Not only could this lead to important information being missed, it would be hard to defend in court. Coombes published a study in 2004 that showed electronic prescribing without any decision support in one Brisbane teaching hospital significantly increased the risk of patient harm compared with a handwritten system - in some cases the error rate was more than doubled (MJA 2004;180(3):140-141). But unfortunately for doctors and their patients, the answer is not as simple as making sure decision support is used correctly. Not all software is the same - the type and quality of systems used in hospitals, general practice and pharmacies varies widely and it's not subject to national standards or any kind of accreditation system. A study published by the National Prescribing Service in 2003 plugged the medication history of 20 elderly patients into four electronic prescribing packages commonly used by GPs and found they missed between eight and 16 of the 32 recommended clinically significant drug interactions. The authors said standardising the quality of decision support in prescribing software was "imperative for patient safety". Coiera doubts much has changed, given the absence of regulation. Although he was on a committee that produced draft guidelines on how software should be regulated, a national accreditation scheme remains some way off. This leaves many doctors in the invidious position of having to rely on software they have no way of knowing is any good, or what kind of evidence its information is based on. Legally, they are liable for their mistakes - even if their software has given them wrong or incomplete advice. "The progress [on decision support] is glacial," says Stephen Phillips, chairman of the National Prescribing Service (NPS). "People pay lip service to the need to move this forward and there has been an awful lot of energy, time and money spent on this in the past 10 years. Somewhere, someone needs to be charged with making a decision." Senior lecturer at La Trobe University's school of public health doctor Ken Harvey is similarly frustrated. "I have a pile of beautiful documents at least a foot high." Harvey says it's "bizarre" that best-practice guidelines are not currently linked into the decision-support software doctors use, so they don't have to wade through endless sources of information to keep up to date. He argues this would help them choose cheaper or more cost-effective drugs and help alleviate the strain on the Pharmaceutical Benefits Scheme. But Medical Software Industry Association spokesman doctor Andrew Magennis won't be holding his breath for the experts to agree on exactly what medical software should contain. Magennis, who is also medical director at HCN, the manufacturer of market-leading general practice software prescribing package Medical Director, says while standards for software content are a good idea in theory, he suspects this is utopian and unachievable. Nonetheless, there are a number of e-health projects boiling away that have the potential to improve patient safety. The NPS is working on a gold-standard, evidence-based list of the most important drug interactions, and is looking at how this could be incorporated into existing software products. And the Australian Commission on Quality and Safety in Healthcare is also looking at the creation of a national database of information that could be used to help doctors in their prescribing decisions. Then there's HealthConnect, the federal-state government joint initiative to establish a national network of electronic patient health summaries. Pilot projects are under way, though critics say privately that the federal Government has lost interest in driving the project forward and is leaving progress to state governments. There is, of course, the federal Government's recently announced $1 billion smartcard, which people can use to store patient health information, and which Coiera says could be incorporated into HealthConnect. But Phillips likens the beavering away on multiple projects to lots of people working on the trees and no one appreciating the size of the forest. He says we have a known problem with medication misadventure and patient safety in Australia, and we need to be designing good electronic systems to address it properly. "And that does not mean convening more meetings and more committees. It means making some decisions." -- Dr. Ken Harvey Adjunct Senior Research Fellow School of Public Health, La Trobe University http://www.medreach.com.au VOIP: +61 (03) 9029 0634; Mobile +61 (04) 1918 1910 _______________________________________________ Gpcg_talk mailing list [email protected] http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
