1. Collect Data
Collect information from all available sources. This includes the history and physical examination of the patient, labs, other diagnostic tests, chart review focusing on the medication list, recent primary care notes, and recent discharge summaries. Depending on the circumstances, you may consider talking to close family members and friends for collateral information provided of course that the patient consents.
2. Distill the Data into Pertinent Positive and Negative Findings.
This includes both positive and negative findings (pertinent positives and negatives.). This could be from history, physical exam, labs, other tests, or chart review.
3. Create a problem representation
Definition 1: Some experts define the problem representation as “an abstract one-sentence summary of the key features of the case synthesized in the clinician’s mind.” – Source, nih.gov
Definition 2: Dr. Eric Strong of Stanford defines a problem representation as a “1-2 sentence summary, using precise medical terminology, of the most highly relevant aspects of the patient’s history, physical exam, and diagnostic tests.” In other words, it is a brief summary of the most highly relevant key features translated into appropriate medical terminology.
You develop it after collecting clinical data and identifying key features. If a case is a referral, transfer, and some consultations, the problem may be clearly defined at the outset. In that case, data collection focuses on verifying or further refining this representation and narrowing the list of possibilities.
The problem representation should:
1) Use semantic qualifiers and
2) Synthesize related findings into clinical syndromes (whenever possible.
The problem representation is also called by the names: 1) Summary Statement, 2) Impression. Note that the problem representation is a summary of the key features. It should include the Key/forceful features from (History, physical exam, and tests, pathophysiology, illness course, etc).
A good “formula” or structure of the problem representation
Age and gender + Highly relevant PMH + Primary symptom using semantic qualifiers + Highly relevant diagnostic data using clinical syndromes when possible.
E.g. “A 60-year-old woman with a history of poorly controlled diabetes presents with chronic, progressive dyspnea, with exam and CXR findings of volume overload, and with unremarkable routine labs.”
Some clinicians like the authors of the nih.gov article cited above limit the problem representation and make it too short. The above formula is more comprehensive and best.
Use Semantic Qualifiers
Translating the story into abstractions (problem representation with semantic qualifiers) fosters retrieval of relevant “Illness scripts”
What are semantic qualifiers? Semantic qualifiers are “paired opposing descriptors that can be used systematically to compare and contrast diagnostic considerations: sharp/dull, acute/chronic, tender/non-tender, productive/nonproductive, insidious/abrupt, proximal vs. distal. “Semantic qualifiers” serve like Google search terms.”
Use OPQRST-A to come up with semantic qualifiers.
Synthesize related findings into clinical syndromes
E.g. If a patient has, “Confusion, T 39 deg, HR 120, BP 130/60, RR 24, WBC 16, cr. 2.4, positive cultures”, you will synthesize it into the clinical syndrome “severe sepsis”. Also, if a patient has, “Jaundice, ascites, confusion, asterixis, total bili 25, INR 2.0”, you will synthesize into the clinical syndrome, “Hepatic failure”.
4. Adopt a Differential Diagnosis Framework
Choose a framework to better understand the patient’s problem. It could be anatomic, physiologic, or other. Frameworks are commonly adopted from a reference source. Note that you don’t have to use one framework all the time. For example, An anatomical framework is excellent for Chest Pain, Abdominal pain, etc. However, this doesn’t mean that an anatomical framework should be used for every condition.
5. Apply the Pertinent Positive and Negative Findings to the Framework
As you apply the pertinent positive and negative findings to your framework, you generate your differential diagnosis.
Arrange it as follows:
First, list the most likely diagnoses. Second, list out the life-threatening diagnoses that you want to quickly rule out. Finally, list the rare diagnoses.
Common diagnoses – List common disorders for which this could be a typical or atypical presentation.
Life-threatening diagnoses – This is the “what can kill list”. List the “don’t miss” diagnoses that you want to rule out.
Rare diagnoses – List some rare diagnoses with some features present in your patient.
—/END/—
References
Lynn S. Bickley, Bates Guide to Physical Examination and History Taking, Eighth Edition, page 783 to 800 (Chapter 18).
Youtube Lecture series titled “How to Create a differential diagnosis” by Dr. Eric Strong, Internist from Stanford University.
Am J Med. 1985 Dec;79(6):745-9. Teaching differential diagnosis to beginning clinical students. https://www.ncbi.nlm.nih.gov/pubmed/4073109
Adv Med Educ Pract. 2016; 7: 247–248. A practical way of creating differential diagnoses through an expanded VITAMINSABCDEK mnemonic. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4853007/
J Gen Intern Med. 2011 Oct; 26(10): 1204–1208. Effusive Reasoning. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3181305/
Clinical and Diagnostic Reasoning. http://www.medicine.uiowa.edu/internalmedicine/education/MasterClinician/DiagnosticReasoningOverview/
Pediatrics, November 2012, VOLUME 130 / ISSUE 5. http://pediatrics.aappublications.org/content/130/5/795.full
Click to access Teaching_Clinical_Reasoning3of6.pdf
http://pediatrics.aappublications.org/content/130/5/795/T1.expansion.html