Critical incidents will happen and how much you share and help others to avoid the same mistake is up to the individual who finds the problem.
My view is you should shout loudly, especially when one considers some problems are easily missed - as in this case when the clinical software fails to take into account the human tendency to make mistakes or overlook information. To err is human but so is the desire to improve and to excel.
A critical incident occurred in our practice when a microbiology report was delegated to reception for the issue of an antibiotic prescription according to the sensitivities supplied by the laboratory.
A doctor saw the report and the sensitivities, and instructed non-clinical staff to contact the patient and to create a prescription for nitrofurantoin according to the sensitivities shown on the report, without looking into the patient's details and not expecting any allergies.
One trusts the clinical software to prevent prescribing a medication the patient is allergic to, but unfortunately it is not necessarily the case.
Sometimes one looks at documented allergies and reviews the risk and benefits with the patient and still prescribes, and what the patient suffers is a side -effect, not an allergy.
The problem arises when the information is presented in a way that is easy to overlook, mainly by non-clinical staff, responsible as they are for plenty of prescription handling. Then, it is only too easy to prescribe without the proper risk assessment.
Ensure correct coding
Information in the electronic clinical notes needs to be properly coded to serve its purpose, in this case to alert you that an allergy exists.
To our surprise, coding 'H/O penicillin allergy' in our software does not trigger an allergy alert, which staff would be used to, but a less evident warning sign.
Unfortunately overcrowded warning dialogues can be easily overlooked, as happened in this critical incident, and as a result an allergy was not taken into consideration.
The situation can be much worse when a generic code has been added (H/O drug allergy, for example) and the allergy itself has been included separately and not necessarily linked to the Read code in question.
It means the warning dialogue will fail to show the relevant information and that you need to dig deeper for the details, including where and when it was added.
Another problem is sharing information, namely through summary care records (SCR). In the previously mentioned circumstances, when the code is a generic one and the detail has been added and not linked, the SCR will show only the generic code, and in consequence it will be impossible to know or to check what the allergy is.
The bottom line is that one of the main purposes of the digital age is to provide and share additional information to make decision-making safer, but equally computers are tools, and unless used to the exact specifications, the information provided can be misleading.
In the complex clinical world, where there are many people interacting with the data at our fingertips, it is paramount we all input allergies and sensitivities in the same way, so we can all benefit and play safe.
Following the near miss, a review of the circumstances took place. A small audit in our practice found 133 patients were coded for H/O drug allergy, a recipe for mistakes.
We have discussed and corrected the information provided on our clinical system.
I invite other GPs to conduct a similar exercise on their coding of allergies, so when and if they share the information about their patients it is crystal clear what the documented allergies are and we stop playing carelessly with fire.
- Dr Millares Martin is a GP in Leeds