This article discusses a case of human error and employs the Swiss cheese model to examine the workplace conditions under which the error occurred.
If the workplace conditions are examined, better defences can be built to avert future errors or mitigate their effects. A trainee's reflective learning log is provided. The subsequent discussion illustrates how the Swiss cheese model could be used to focus the significant event analysis.
Example reflective learning log
Subject title: Delayed patient referral
A patient complained that his referral to sports medicine was delayed. He felt that our referral pathways were not robust enough and he wanted these reviewed.
When I first saw this patient, I referred him to the physiotherapist. I expected that once seen, he would get her input into what would be the best referral pathway and then he would come back to me so I could refer him on to the most appropriate service.
It is my usual practice to do the referral letter with the patient in clinic. The patient hears the dictation, corrects any errors and also declares their non-availability dates on the Choose and Book form. They hand the form to the Choose and Book referrals secretary and select their appointment immediately.
I feel the communication here broke down in a few ways:
- On the day before my annual leave started, the physiotherapist met me socially and told me to refer the patient (without tasking me) and I forgot.
- The patient didn’t book a follow-up appointment with me to trigger me writing the referral.
Possible ways to improve:
- Don’t forget verbal messages, but I am not sure that my memory is infallible. I could create a task list for myself about issues that require follow-up.
- I could get patients to routinely book follow up with me after seeing the physiotherapist.
- The physiotherapist could task me via the computer system to transfer the patient's care back to me.
- The physiotherapist, after seeing the patient, could tell him or her to see their GP to organise the referral.
What issues were raised by this significant event?
I used the Swiss Cheese model to help me understand the issues. According to this model, in an ideal medical system, there are several intact defensive layers (like steel plates) protecting against the human error, which in this case was my memory lapse.
In reality, however, the defences are less like solid steel plates and more like slices of Swiss cheese, having many holes, though unlike in the cheese, these holes are continually opening, shutting, and shifting their location. The presence of holes in any one 'slice' does not normally cause a bad outcome. Usually, an adverse outcome occurs only when the holes in many layers momentarily line up to permit a 'trajectory of accident opportunity' —all defensive slices are breached and harm occurs.
So what were the holes here?
1. Did anyone commit an unsafe act – in other words, were there 'active failures'? The physiotherapist did not task or email me with her referral advice. I did not create a task list for myself, probably because I was under time pressure with my pending holiday.
2. Are there weaknesses (or 'latent conditions') within the system? The practice does not have a leaflet for patients advising them what the referral procedure entails and who they should contact if they do not get an appointment letter. The doctor notifies the referrals secretary once the referral letter is dictated; not when the decision to refer is made. Had I tasked the referrals secretary, with a note saying the referral letter would be done after the physiotherapist had seen the patient, there would be greater awareness within the system of an outstanding task. If a referral letter is done without the patient being present, the referrals secretary could email or telephone the patient for administrative details. This would entail a change to her protocol.
What was done well?
The patient complained about the referrals system, not about individuals, which helped me to focus on examining my referral procedures to make them more robust.
The physiotherapist and I spoke about the case and I aplogised for having forgotten to act on her verbal advice. We have arranged a meeting with the patient, who feels that his delayed referral resulted in delayed treatment and he was not match ready in time for an important sports competition.
What was not done well?
When I discovered that I had not written the referral prior to going on holiday, I wrote it immediately and sent it off. I should have called the patient, apologised for my lapse and kept him informed.
What could be done differently in future?
The author of the Swiss cheese model writes: 'active failures are like mosquitoes. They can be swatted one by one, but they still keep coming. The best remedies are to create more effective defences and to drain the swamps in which they breed. The swamps, in this case, are the ever present latent conditions.'
In this case, the patient pointed out that my referral process was a 'swamp' and by not phoning him when I discovered my lapse, I acted defensively about my swamp. Perhaps it is important to recognise that when I am in the middle of a busy job, I cannot readily distinguish the 'swamp' from the fertile land. It helps to take time out to examine the land, and tackle problematic swamps.
How and when will you address these?
I will try to take a complaint as feedback and as an opportunity to examine my environment. I should ask myself: 'have things slipped and do I find myself in a swamp?'
As an educational supervisor, I would cite this reflective entry as evidence for three domains:
- Organisation - the trainee raised a significant event appropriately, discussed what happened and looked at contributing factors.
- Working with colleagues - the trainee discussed the issue with the physiotherapist and suggested reviewing the referrals protocols.
- Fitness to practise - although the care of this patient was suboptimal, the trainee engaged with quality improvement procedures and took steps to improve the quality of future care.
Dr Naidoo is a GP trainer in Oxford. She has written three books on how to pass the CSA, including CSA Practice Cases for the MRCGP.
- Reason, J. Human error: models and management. BMJ 2000: 320; 768–70. https://www.bmj.com/content/320/7237/768
- NHS Education for Scotland and The Health Foundation. Enhanced significant event analysis - a human factors system approach for primary care.