WASHINGTON -- A clinician's attitude toward an individual patient may be a factor in many diagnostic errors, a jarring observation that was offered during a doc-to-doc session at the meeting here.
"My mother-in-law hates me. She's hated me for twenty years," said , an internist and clinical educator at Tufts University School of Medicine in Boston.
When he sees a patient who reminds him of his mother-in-law, he has to adjust his mindset in order to very intentionally focus on offering her the same care he would any other patient.
"All of you have the equivalent of [my] mother-in-law," he told his audience. "It may be an ex-boyfriend, girlfriend, sibling or parent, and when that encounter happens, it's important to be wary of the possibility of a diagnostic error, in particular a meta-cognitive error."
This means "thinking about your thinking and thinking about your feelings and how that affects the outcome."
Trowbridge explained meta-cognitive errors, systems errors and no-fault errors, but first he shared how diagnostic reasoning occurs.
Diagnostic reasoning
"When you want to fix something that's going wrong, it's really useful to know how it actually works," he said, and doing that involves a three step process: data collection, problem representation and subconscious matching.
In the first stage, clinicians gather the patient's history, perform a clinical exam, and consider laboratory tests or imaging. Then, they look for specific details or "golden nuggets" that can direct them to a category of illness. Finally, the clinicians matches the information gleaned to an "illness script" or mental representation plucked from their "internal Rolodex of diseases."
Often physicians jump back and forth between stages and sometimes reasoning fails.
There are three types of these failures or diagnostic errors, he said.
- No fault errors: When the patient actively undermines the diagnostic process
- Systems related errors: When faulty tests or faulty data, inadequate supervision of trainees, poor communication, or an organizational failure get in the way of a correct diagnosis
- Cognitive errors: When faulty knowledge, faulty data gathering, faulty synthesis of information, or affective error impact the diagnosis.
"Rarely is their only one cause of an error," he said.
Trowbridge homed in on cognitive errors unpacking each potential circumstance.
In the case of "faulty knowledge," a clinician may have seen an illness that presented in an unfamiliar way. "If you don't have that illness script for whatever that disease is, you're not going to be able to make that diagnosis," he said.
With faulty data gathering, clinicians may gloss over important diagnostic questions, fail to conduct a comprehensive clinical exam, or disregard patient history. These type of errors are common when clinicians feel overwhelmed by time pressures.
Faulty synthesis happens when clinicians solve a puzzle too quickly, often overestimating the importance of a single finding. Also called "premature closure" Trowbridge believes it's the most common cause of diagnostic error. The reasons are simple.
"Diagnostic uncertainty really bothers us," he said, and it bothers patients. "So a lot of us push to make a diagnosis, because then we have that degree of certainty."
And lastly there's meta-cognitive error, i.e. the mother-in-law problem. Trowbridge's mother-in-law is prone to hyperbole. So he recognizes his own tendency to discount the responses of look-alike, sound-alike patients and focuses intently on giving these patients equal care. When he finds he can't overcome the "mother-in-law" bias, he refers patients to colleagues.
Preventing mistakes
Trowbridge noted that the first step in improving diagnostic reasoning is awareness of the risk for cognitive errors. Experience, seeing as many patients as possible, is also important to expanding firsthand knowledge of diseases.
In addition, to improve intuitive reasoning, Trowbridge said gaining feedback on diagnoses was critical.
When physicians guess at a diagnosis and don't know what happened later, they assume they got it right, he said.
"A lot of times we aren't right, but we assume we're right and that reinforces things we're doing incorrectly."
"We need to be talking to our institutions and our practice groups and our partners about ways to get feedback on our diagnostic decisions," he said, and that includes both cases clinicians got right and those they got wrong.
As for analytical reasoning, Trowbridge focused on four ways to avoid errors.
The first is"recognizing you're at risk for bias," he said.
For example, Trowbridge described a hypothetical case of a man who presented at an emergency department three times a week, every week -- each time the man was drunk.
This patient may have proved frustrating to hospital staff in the past, which is why it's particularly important for the emergency department staff to overcome their "familiarity" bias, "because one of these days, that man going to have a subdural hematoma."
The second strategy is a simple time-out. Trowbridge encourages clinicians to ask themselves, "Does this patient really have what we think she has, and have we looked everywhere that we need to look?"
Trowbridge said in this case it is not enough to rely upon the record -- the clinician needs to ask the patient about items included in the record. Verify every diagnosis, he said.
The third strategy is really a tool: checklists. Using presentation checklists consciously, even when a diagnosis seems obvious, is something clinicians often fail to do because the steps seem so obvious. But in many cases, clinicians are interrupted frequently during a visit, and forget to follow the most basic steps.
The fourth strategy is one of reflection, and uses the mnemonic device, SAFER.
- Serious diagnosis : Have you considered all serious diagnoses?
- Alternative diagnosis: Have you considered any alternatives?
- Feelings affect thinking: How does the way this patient make me feel affect my thinking?
- Extraneous data: Is the information you set aside really extraneous?
- Reasons: Why did this happen?
He finished by saying that clinicians need to advocate for themselves and focus on working as a team.
"We can't do this alone ... To be good diagnosticians, we need time with patients, we need time to think, we need time to discuss, we need feedback and we need help from other folks to do all these things."