HDC case – cognitive bias
By Dr Peter Moodie, College Clinical Advisor
15 November 2022
A recent decision by the Health and Disability Commissioner (HDC) relates to an urgent care clinic where there was an unacceptable delay in diagnosing a patient suffering from a myocardial infarction. The issues raised are equally relevant to general practice. The Commissioner specifically highlights the risk of “cognitive bias” and a “lack of critical thinking on diagnostic decision-making.”
The case took five years to come to a decision which must have been trying for both patient and clinical staff.
The case involves a middle-aged woman (Mrs B) who had type 1 diabetes with vascular complications.
She was referred with vomiting and severe pain (see below) to an urgent care clinic on a Saturday afternoon, by her regular GP. She had recently been overseas and developed gastroenteritis and she was treated with metronidazole; however soon after taking this she became acutely unwell and following a phone consultation she was diagnosed as having an adverse reaction to the medication.
The GP phoned the urgent care clinic to explain the situation, noting that she had been seen by them on three occasions during this illness.
As often seems to be the case when a complaint is generated, the narrative given by Mrs B and her husband differed from the clinical staff. Mr B claimed that he rang the clinic while they were on their way and told them that he thought Mrs B was having a heart attack and repeated this on arrival and at least five more times during their stay. On each occasion he specifically explained that Mrs B was having severe chest pain and they wanted an ECG done.
The triage nurse did record in the triage form (variously described as being electronic and then in another context handwritten) that Mrs B was feeling pain in her arms, chest, and back following “ingestion of metronidazole. However, she didn't record the chest pain in the electronic clinical notes. The nurse assigned a triage score of four to the case (this is the second lowest priority score).
In her explanation to the HDC, the nurse stated that at no time does she remember Mrs B or her husband mentioning a possible heart attack or requesting an ECG. The attending doctor also said he didn't remember any reference to heart attacks nor requests for ECGs.
Over the next five hours, Mrs B remained in extreme discomfort with vomiting, dry retching and generalised body pains which were attributed to an adverse reaction to metronidazole. Over that time, she was managed by four nurses and two doctors. She was treated with antinauseants (ondansetron), IV fluids and at least three doses of IV morphine.
After five hours, she did have an ECG and was diagnosed with a myocardial infarction and transferred to hospital where she subsequently had cardiac surgery.
The HDC was critical of the delays in making the diagnosis of myocardial infarction and found the various medical records wanting. The commissioner noted that Mrs B’s firm and sincere belief that she had made repeated requests for an ECG, however she was unable to determine whether this was the case.
Firstly, the original diagnosis of gastroenteritis was correct and the second one of an adverse reaction to metronidazole was also probably correct. Unfortunately, there was a third diagnosis of a myocardial infarction and missing this was what the Commissioner described as “cognitive bias."
It's always important to question initial diagnosis and think critically about the possibility of other differentials.
So, could this diagnostic bias have been avoided? There were two clues:
- Firstly, Mrs B was a type 1 diabetic who was plainly distressed and vomiting; to that end it could be argued she should have been given a higher priority score.
- Secondly, over the time she was in the clinic, she was not improving despite major analgesia and antinauseants.
These are important factors in what transpired. It also reminds us to continually revise provisional diagnosis and thinking of other differential possibilities.
The Commissioner always focuses on documentation in these cases. In this case, there were various lapses, however, in a busy clinic is it even practical for every encounter to be documented with the reasons for it and the conclusions for it recorded?
Plainly documentation is important, but it must not slow down the implementation of good care. In this case good communication between all those involved may well have been more useful.
Lastly, it's common for a complainant and clinical staff to have a very different interpretation of clinical events. Often, it's not easy to identify when “cognitive dissonance” is occurring but we should always be alert to a potential breakdown in communication and listen respectfully.