Better use of computer systems in general practice could reduce errors in GP prescribing, which are currently found in one in 20 prescriptions, a major study has recommended.
An investigation of prescribing in general practice commissioned by the General Medical Council, found most errors were classified as mild or moderate, but one in 550 was judged to be serious.
It recommends a greater role for pharmacists in supporting GPs, more effective use of IT in general practice, and extra emphasis on prescribing in GP training to address the issue.
Professor Sir Peter Rubin, chair of the GMC, said GPs were typically very busy so it was important to make sure that prescribing was given the priority it needed.
He added: “Using effective computer systems to ensure potential errors are flagged and patients are monitored correctly is a very important way to minimise errors.”
The PRACtICe study (Prevalence and causes of prescribing errors in general practice) is the largest study of its kind.
It looked at 6,048 unique prescribing items for 1,777 patients from 15 GP practices. It reviewed errors using interviews with prescribers, root cause analyses and focus groups with primary health care team members.
It found the most common errors were missing information on dosage, prescribing an incorrect dosage, and failing to make sure that patients got necessary monitoring through blood tests.
The report recommends that practices develop strategies to make the best use of the safety features that are already present on their systems.
It said the GMC should also discuss the outcome of the study with organisations with a remit for quality assurance including the Royal College of GPs and the General Pharmaceutical Council.
These discussions should cover whether improvements could be made in the training of GPs and practice staff to make the best use of prescribing features in GP computer systems.
The report also says that consideration should be given to the use of pre-specified ‘order sentences’ to encourage prescribers to provide appropriate dosage instructions. It adds that context-specific dosage guidance could be provided, taking account of patient factors such as age and renal function.
The authors say consideration should also be given to alerts for blood test-monitoring for certain drugs and for the most common and important examples of hazardous prescribing.
These alerts would be in addition to drug-drug interaction alerts, which are already present on all GP clinical computer systems in the UK.
The report outlines other strategies that could be used to improve use of IT. These include avoiding similar drug names being adjacent in pick lists, running searches on clinical records systems to identify potential prescribing errors and patients requiring blood-test monitoring, and familiarising locums with health information technologies available in practices.
Professor Tony Avery from Nottingham University’s medical school who led the report said prescribing was a skill and one that doctors should take time to develop and keep up to date.
He added: “Few prescriptions were associated with significant risks to patients but it’s important that we do everything we can to avoid all errors.
“GPs must ensure they have ongoing training in prescribing and practices should ensure they have safe and effective systems in place for repeat prescribing and monitoring.”
© 2012 EHealth Media.
Ordering drug names....use the indicationcmacfarlane 164 weeks ago
The problem of similar drug names appearing adjacent to each other in prescribing pick lists can be virtually eliminated by using the indication as a filter. The prescriber is then presented with a pick list of relevant drugs which are far less likely to have similar names.
Read the reportMary Hawking 164 weeks ago
I think you need to read the report.
The study involed 2% of the patients in 15 practices: the fact that out of 1777 patients, 1200 had had prescriptions in the study period suggests that the sample was on a combination of non-prescribing criteria.
When it comes to "monitoring errors"I was surprised to find Simvastatin as the top drug not being monitored or having a plan for monitoring recorded: there has been advice that patients should be put on 40mg and no futher testing was needed: warfarin - second on the list - is managed by the organisation responsible for the monitoring.
The study only mentions EMIS and TPP SystmOne in the glossary: GP systems differ in their presentation of information: are the conclusions applicable to all the other GP systems?
Ordering drug nameslayton 164 weeks ago
The order in which drug names are presented in pick lists is a recurring bone of contention. There is a simple and highly effective solution to the problem which will virtually eliminate this particular risk and it simply requires some intelligence in the system. Ultimately it is the doctor who decides which drug is to be prescribed so the list should present the doctor with the very limited range of drug names that he or she most commonly use with the rarer, or never, prescribed drugs right at the bottom of the list. Within that constraint alphabetic ordering can be used to speed up the selection process. The lists will then be ordered by the habits of the specific doctor prescribing at that time and the system should monitor the habits and adjust the picklist accordingly.
Extended SCR role needed to share information, to reduce risk.P Millares Martin 164 weeks ago
When patients are seeing by many doctors in different settings, the share of information (changes in prescriptions, in doses, acknowledging side effects or even allergies) need to flow more quickly than currently does.
IT is there but there is not enough use of it, too many limitations
If any prescription or allergy recorded by a clinician on any place (whether primary care centre -where today it happens only for GMS registered patients- or secondary care -A&E, outpatient, inpatient, ... where it is not allowed at present-) is put into the summary care records, and feeds into any contact from the same patient elsewhere,... there will be no doubt less chances of error, of giving the wrong medication, the wrong dose, ...
We use computers, but not their full potential.
Among the suggested strategies...HMF 164 weeks ago
"avoiding similar drug names being adjacent in pick lists".
Contra, I would argue that similar drug names SHOULD be adjacent, in the hope that the need to discriminate between them might be obvious.
If the user (here, the prescribing doctor) sees, within the visible part of the pick list -- say one screen-ful, ONLY ONE term which is "something like" what they wanted, I feel there is a danger of them saying "oh, I must have forgotten the spelling, this must be right, because it's the only one in the list".