Tempo de leitura: 6 minutos
Author: Peter Elias, MD
Too many discussions treat the diagnostic process as a discrete event, the second part of a three-part cascade where information is obtained, a diagnosis is made, and then treatment begins.
I believe it is a mistake to frame diagnosis so narrowly, reducing what is an inherently complex and iterative process to a mere event. During my 40 years in primary care, I have observed that this tendency to frame diagnosis narrowly as an event rather than a process is particularly common in discussions of diagnostic error in acute care settings like emergency departments or intensive care units. These discussions properly focus on the selection and correct interpretation of specific tests in order to determine the need for specific interventions. Does this EKG tracing justify the diagnosis and treatment of a myocardial infarction? Does this lower extremity ultrasound justify the diagnosis and treatment of pulmonary embolus? This framing is also common in discussions of surgical diagnosis: does this MRI justify shoulder surgery or spine surgery, or does this CT scan warrant abdominal surgery?
Experienced primary care clinicians inevitably learn that the best and most efficient way to make an accurate and useful diagnosis is over a period of time and across multiple visits rather than during a single encounter.
In primary care, and in specialties that deal with chronic illness, multiple illnesses, or poorly differentiated illnesses, one is rarely in the unenviable position of possessing only a limited clinical snapshot of the patient and therefore having to make a diagnosis and treatment decision based on a test result. (I use the term ‘clinical snapshot’ rather than clinical picture because it really IS a limited snapshot, obtained quickly, and in a narrow and time-limited context.)
In primary care, and in specialties that deal with chronic disease or multiple overlapping problems, the majority of diagnoses are best made using the clinical picture as it evolves over time, classically called the natural history of the illness.
Medicine doesn’t have a diagnostic ‘test’ for anxiety, depression, most causes of low back pain, most causes of headache, or most causes of fatigue, heartburn, dizziness, or cough. For these and other common primary care problems, the appropriate diagnostic process is something like this:
- Are there any things I dare not miss, right now, during this visit, because doing so would put the patient at immediate risk?
- If so, what can I do to determine the presence or absence or these diagnoses?
- If not, how do I remember this list of possible diagnoses, so I can refer back to it if the clinical picture changes?
- What are the most likely causes of what I am seeing and hearing?
- Are there any likely causes of what I am seeing and hearing that I can easily and efficiently prove or disprove?
At this point, it is essential to involve the patient as an active participant in the decision-making process.
This starts with an honest conversation with the patient about the diagnostic possibilities, the range of possible treatments and outcomes with and without treatment, and the degree of certainty or uncertainty.
The next step requires exploring with the patient what part or parts or the clinical picture the patient is most concerned with. Depending on the patient and the setting, their focus may be on obtaining a diagnosis. For other patients or in other settings, the chief concern may be about selecting treatment.
Without learning what the patient wants, we can treat the symptoms and we can treat the disease, but we cannot treat the patient.
At this point, we will have generated a collection of diagnostic possibilities of varying severity, prevalence, likelihood, and prognostic possibilities. We also know what the patient’s preferences are, because we asked and listened.
Though we are now ready to embark on trying to establish a diagnosis, we are rarely in a position to select a specific test to diagnose the current clinical problem. Instead, we have to engage the patient in selecting one of the following reasonable:
- Work on diagnosis, but hold off on treatment for now.
- Work on diagnosis, but simultaneously treat symptoms.
- Treat symptoms and observe the course. (Here natural history is ‘the test’ we are using, but it doesn’t have a discrete ‘result’ in the sense of observed Q waves, blood sugar, or potassium.)
- Not treat symptoms, monitor the course of the illness awaiting further clarifying information, and modifying our approach iteratively as further information becomes available.
- Trial and error, where we select a possible diagnosis, treat for that tentative diagnosis, and observe the response, moving on to a second, third or more empiric treatment trials as needed.
Very few patients in my practice who presented with back pain, headache, fatigue, cough, or fever have any ‘tests’ done at an initial visit. In primary care, although there are situations where tests are clearly essential at the outset of the diagnostic process, most clinical problems are best diagnosed by history, a brief exam, and tincture of time.
I doubt that more than 2 patients out of my daily schedule of 18-20 needed a ’test’ in order for the patient and I to establish a safe and effective plan.
My point here is that framing the diagnostic process narrowly around the interpretation of test results is appropriate only in a very limited piece of the diagnostic universe. This is critically important when germane: When I present to an emergency room poorly responsive and hypotensive, I want the proper tests ordered and the test results properly and quickly evaluated.
However, from the perspectives of the primary care clinician, the patient, and the caregiver, this represents only a tiny fraction of the diagnostic universe and the approach that is necessary in this setting is not easily generalized or appropriate across the broad landscape of clinical medicine.
I am much more interested in exploring ways to improve the diagnostic process in the 90% or more of circumstances where the selection and correct interpretation of a test is unlikely to be definitive.
Of course, that reflects my 40 years in the primary care front line where ambiguity is part of the air we breathe.
Peter Elias is a retired family physician. After Dartmouth College, medical school at the University of Rochester, and residency training in Madison, Wisconsin, he loved every minute of his 40-years of active primary care practice in Maine, during which he taught part-time in a Family Medicine residency.
His clinical interests include life-style modification and disease prevention, behavior change, and collaborative care with engaged and autonomous patients. He is on the Board of the Society for Participatory Medicine.
He is an avid Nordic skier, and enjoys digital photography, folk guitar, wood working, and teaching himself new computer skills. He blogs intermittently about medicine and other topics at petereliasmd.com. He is married to Cindy, his best friend and North Star, and they have two children and three thriving grandchildren.
Contact email: [email protected]