In the first of this three article series on errors and mistakes in general practice, we explored the psychological basis for errors and biases in human behaviour that lead to mistakes, and then went on to think about the psychological limits that contribute to mistakes. In part two, we considered the field of system safety and how we can use these insights into human factors to build a safer, open learning culture in medicine based around the principles of Black Box Thinking.
In this final article, we'll run through a hypothetical locum session (but inspired by real events!) and the typical response to errors, finally examining how a more considered, system-based approach based on the learning from psychology and Black Box thinking could promote patient safety, practice efficiency and GP morale and retention.
Dr L finished a session at a practice that was new and unfamiliar to her. During the session, a patient complains that she had been abrupt and rude during their consultation. Two weeks later, a patient seen by Dr L phones the practice to find out when they can expect to hear about their appointment at the ENT clinic that Dr L referred them for. The practice finds no record of the referral.
Three months later, a 58 year old lady who had seen Dr L with intrusive peri-menopausal hot flushes came for review. Dr L had initiated HRT, but had incorrectly prescribed an oestrogen-only preparation in a patient with a uterus.
By this stage, the practice are very unimpressed. They discuss it at their practice meeting and decide to inform Dr L of the errors and that her performance is below what they will accept in future, so she is no longer going to be booked by them.
So that's the WYSIATI (What You See Is All There Is) summary of events, a very automatic System One (S1) way of thinking described by Daniel Kahneman in his Thinking, Fast and Slow, that we explored in the first article of this series. Also in that article, we covered the concept of cognitive load and the depletion of our slower, more effortful, discerning way of thinking, System two (S2), and reflected on the evidence of how this may affect us during our GP surgeries and leave us more prone to errors.
Taking these ideas and evidence and using the concepts of Black Box Thinking from the second article, what is revealed if we take a deeper examination of the events during this GP locum session.
08:30 - Parking
Having driven 45 mins through rush hour traffic, Dr L arrives promptly 30 minutes before her session is due to start, but the small practice car park is already full, the Doctor only spaces are taken and there is only limited on-street parking nearby. She struggles to find a parking space some streets away and is pre-occupied that the only space she can find leaves the front wheel of her car nudging a yellow line.
08.45 - Dr Who?
Dr L presents to reception staff who weren’t expecting her; no room has been allocated, no computer login activated or smartcard details registered.
09.10 - 18 patients
Two senior practice staff have to set her up to login to the IT system. Dr L sees that she has been allocated to see 18 patients, where it was her understanding that she been booked for 12 patients. Not only does she start her session later because of the delay in setting up her login, she also wonders how she is going to complete this 18 patient session in time to reach her afternoon commitment.
09.40 - Pregnancy test
She sees a patient with pelvic pain who needs a urinary pregnancy test. Having unsuccessfully searched the room, she phones reception on the only internal phone number she's been given and asks if someone can carry out the test so that she can continue seeing patients, as she's already running late. Dr L is now having to rely on effortful S2 thinking for a task that other GPs familiar with the practice would be able to do in automatic S1.
11.15 - Rude
Over two hours into the session, four hours since she last ate, Dr L’s S2 is now tired and very busy. She has just had another complex consultation with a patient who has presented with multiple problems, and so her full attention is on the computer screen as she focuses on safely generating a prescription to initiate a new medication - she is thinking of doses and interactions. The patient continues talking and asking questions, while Dr L does not register what the patient is saying. It looks to the patient as if Dr L has been rude and ignored her. The patient is also aggrieved that Dr L is running late, so decides to immediately complain.
11.55 - Referral
Dr L’s 10th patient needs a routine ENT appointment, but she has no information about how the practice manages this sort of referral, so tells the patient she will dictate the referral letter and ask the secretary to contact them if any action is needed. As a safety net, Dr L remembers to ask the patient to phone the practice in two weeks if she has not heard about an appointment.
12.50 - Prescribing error
It’s nearly six hours since Dr L ate or rested. She has drunk some water from the bottle she carries, but has not been offered any refreshment by the practice - she has not been shown where the kitchen is and does not feel comfortable taking without being offered.
Her surgery is running 40 minutes behind schedule and knows she will be late for her next assignment because of the booking error which led to her having 18 instead of 12 patients, when she meets her final patient, a 48 y old lady with disabling hot flushes.
Like many GPs, it’s been a while since Dr L initiated HRT. She looks at the new layout of the paper BNF and finds it baffling; she is having trouble focusing, her S2 is very tired and depleted: she prescribes the incorrect type of HRT.
13.15 - Dictating
The practice has a digital dictation system, but none of the admin staff available knows the login for a locum or how to show Dr L how to use it, so a helpful receptionist gives her an old fashioned manual dictaphone with a little tape; she dictates the letter and is directed to leave it on the secretary's desk, with a loose slip of paper explaining the contents; she leaves the practice to dash to her next assignment.
The secretary comes back from lunch; she has not used these tapes for years and wonders why it’s left on her desk; no-one knows quite what happened to the hurriedly scribbled note, but it was not clearly connected to the tape and so was overlooked. The secretary assumed the tape was an old, legacy tape left for her to dispose of and so she discards it, leaving no record of the ENT referral.
Three months later
Dr L is distressed when, three months later, she receives a letter from the practice saying that after three incidents of concern they prefer not to work with her again. The allegation of rudeness makes her question whether she is a good doctor and whether she needs to go on a communication skills course. The prescribing error reduces her confidence and morale. She decides to reduce the number of sessions she works to cope with the stress - a further loss of GP capacity, which harms us all.
The practice, having located the concerns to the behaviour of one individual, is satisfied it has dealt with the problem by excluding the individual. But as the months go by, they find they just cannot find any "good" locums - they all keep making mistakes. And locums don't seem to want to work with them. The practice increasingly find they have to cross cover GP absences internally. And then one of the GP partners becomes ill and has to take long term sick leave...
In the initial response, both Dr L and the practice have been duped by hindsight bias and WYSIATI and they are all the worse off for it. Once a more objective, Black Box approach to these mistakes is taken, it becomes clear that there are multiple small glitches in the system that has made it very easy for errors to weave their way through the Swiss cheese maze that should be there to protect patients and doctors.
Many small but significant marginal gains could have been identified - such as the need for an improved booking process, clearer parking arrangements, improved pre-arrival preparation by the practice and locum induction procedures - which could have led to progress, with better conditions and safety for patients, improved efficiency for the practice and improved morale and reduced stress for the GP locum. Many of these layers of error protection seem mundane and administrative but if they had been in place, the story of this locum session and the outcome for the locum and the practice could have been very different, possibly leading to a productive, mutually beneficial and supportive relationship.
Or put another way, rather than taking the populist narrative of simply not employing the locum again and singling out an individual for blame, an enquiry led by evidence and data could lead to small changes that make a big difference. And imagine the impact if these simple marginal gains are repeated over time and across many practice and locum interactions.
Cases like that of Dr Hadiza Bawa-Garba sent a shockwave through the profession, with many feeling that the use of an adversarial criminal court case to prosecute a doctor for manslaughter was not an appropriate arena to fully, justly and openly explore the system errors which led to a terrible outcome. In fact, the alleged indirect use of the doctor’s reflections on what she felt could have been done better and her perceived personal failings by the prosecution and the subsequent behaviour of the GMC has been seen as turning the clock back on safety by cementing the culture of personalising blame in healthcare. It is 19 years since the publication of Lucian Leape's report "To err is human" which identified that most errors in health care arise from faulty systems. And yet it seems the harm to patients - and doctors - of not recognising system errors continues.
There exists years of research by some of the world's top psychologists and economists, and decades of experience from safety-critical industries like aviation that have brought about staggering improvements in safety and reliability by deploying an open evolving approach to learning from failure. Learning from mistakes is the very basis of life on earth through evolution. Healthcare needs to evolve; there will be mistakes, and learning from those mistakes, is what will ultimately transform us from one of the most lethal - remember where we began this series with the statistic of 400,000 preventable deaths in the US each year - to what it should be - one of the safest.
Richard has worked as a freelance GP locum since 1995 in around 100 different practices, living and working in West Sussex and Hampshire. He founded NASGP in 1997, he is NASGP’s chairman and started the UK’s first locum chambers in 2004.
He enjoys walking, reads too many books on behavioural economics and has an unhealthy obsession with his sourdough starter.