7.3 Differentials, Data Synthesis & Diagnosis
Key Takeaways
- Sort every vignette into must-not-miss life threats, the most-likely diagnosis, and a unifying diagnosis that explains the whole cluster.
- Two-step thinking catches the “URI” that is actually a nasal foreign body or epiglottitis; the first obvious label is often the distractor.
- Consult cardiology for a symptomatic murmur, neurology for regression, and genetics for dysmorphology — referral is part of diagnosis, not a dump.
- Treat empirically when delay threatens life or organ; test first when a false-positive treatment is the harm; document working versus confirmed diagnoses.
- PNCB asks for one best next step: in a febrile non-weight-bearing toddler, urgent exclusion of septic hip outranks labeling toxic synovitis.
Domain II.E is the payoff of the history and the tests: integrate data, decide whether a consultant is needed, prioritize the differential, and name a working or confirmed diagnosis. PNCB items are one-best-answer questions. The exam is not asking you to list every reasonable step. It is asking for the single best next step or the most accurate conclusion given the vignette.
Vignette method: must-not-miss, most-likely, unifying
Read the stem twice. Then force three columns before you look at the options:
| Column | Question you ask | Examples |
|---|---|---|
| Must-not-miss | What diagnosis, if missed today, causes death, organ loss, or permanent disability? | Septic hip, testicular torsion, epiglottitis, foreign-body airway, meningococcemia, DKA, non-accidental trauma, leukemia presenting as limp |
| Most-likely | What is common and fits the base rate plus this story? | Viral URI, transient (toxic) synovitis after a cold, constipation, GAS pharyngitis in a school-age child without viral features |
| Unifying | What one diagnosis explains the entire cluster rather than three coincidences? | Leukemia explaining limp and pallor and fever and bruising; inflammatory bowel disease explaining growth failure and abdominal pain and blood in the stool |
Prioritize life threats first. A well-appearing child with a typical cold is still allowed to have a viral URI after you have mentally cleared stridor, drooling, dehydration, and pneumonia. The order of thinking is: resuscitate and exclude catastrophe, then pick the common disease, then ask whether one unifying process is a better explanation than two unrelated ones.
A unifying diagnosis is not an invitation to choose a rare zebra over a typical story. It is the tool you use when the “obvious URI” does not explain the weight loss, night sweats, or focal finding.
Two-step thinking: the first obvious diagnosis that is wrong
PNCB stems often hand you a familiar label in the first sentence. The second look is the exam.
- “URI” that is a retained nasal foreign body: unilateral, persistent, often foul discharge after a toddler was left with beads or food. Bilateral clear rhinorrhea is a cold; one pus-soaked naris is not.
- “URI” that is epiglottitis or critical airway obstruction: toxic appearance, drooling, tripod or sniffing posture, muffled voice, stridor. Do not force a tongue-blade throat exam in primary care. Emergency airway support is the next step, not a rapid strep.
- “Gastroenteritis” that is DKA: polyuria that was called diarrhea, tachypnea that is Kussmaul, weight loss, and a POC glucose you should have obtained.
- “Growing pains” that are leukemia or SCFE: night pain that wakes the child, unilateral limp, systemic symptoms, or an older obese adolescent with hip or knee pain and obligatory external rotation.
The first label is the distractor. The second look is the answer.
Consultants: when primary care is not the right room
You do not need a consultant for every murmur or every odd-looking ear. You do need one when the data say the next safe step is outside the clinic.
| Finding | Why you involve a consultant |
|---|---|
| Murmur plus symptoms (exercise syncope, chest pain with exertion, poor feeding, failure to thrive, family sudden death, pathologic exam features) | Cardiology — not “recheck the still-asymptomatic Still’s murmur in a year” logic |
| Developmental regression, unexplained encephalopathy, or new focal neurology | Neurology (and emergency evaluation if acute) |
| Dysmorphology, multiple anomalies, or a pattern that suggests a syndrome | Genetics |
| Suspected septic joint, unstable SCFE, or fracture you cannot manage | Orthopedics / ED |
| Cytopenias, mediastinal symptoms, or limp plus a malignant-looking CBC | Hematology-oncology |
Incorporating consultants is Domain II.E language: you integrate their input into the diagnosis and the plan. Referral is not dumping; you send a question, the relevant data, and a working diagnosis.
Treat empirically versus test first
| Situation | Usual CPNP-PC move |
|---|---|
| Life-threatening bacterial disease (toxic shock, meningococcemia, septic hip once cultures are in motion in the ED) | Do not delay treatment for a perfect workup |
| GAS pharyngitis | Test, then treat — do not treat every sore throat empirically |
| Typical viral URI, bronchiolitis, or most sinus symptoms that do not meet AAP acute bacterial sinusitis criteria | Supportive care; do not shotgun antibiotics or CT |
| Febrile young infant | Test (and often treat) per age-based serious-bacterial-infection pathways — not a wait-and-see that belongs to a well toddler with a cold |
| Indicated adolescent STI screen | Test (NAAT); treat per results or per expedited protocols when indicated |
Empiric treatment is for time-critical danger. Testing first is for conditions where a false-positive treatment (unnecessary antibiotic, missed diagnosis) is the harm.
Working diagnosis versus confirmed diagnosis
Document what you actually know.
- Working diagnosis: “Suspected septic arthritis of the hip — referred to emergency orthopedics.” It guides the next step.
- Confirmed diagnosis: synovial culture, a definitive imaging finding, or a gold-standard laboratory result.
PNCB will ask you to act on a working diagnosis (send the septic hip now) without waiting for confirmation that can only be obtained in the operating room. Conversely, do not record “leukemia” as confirmed because a child limps; the CBC and the referral are the next steps, not a primary-care label carved in stone.
Worked example: the toddler with a limp
A 3-year-old stopped walking yesterday. Yesterday he had a low-grade fever; last week he had a cold. He is irritable. You will not diagnose from the doorway.
History that splits the differential. Duration and progression; fever; recent URI; trauma or unwitnessed fall; night pain or waking; weight loss, pallor, bruising, night sweats; morning stiffness (juvenile idiopathic arthritis); tick exposure in a Lyme-endemic area; sickle cell disease; immunizations and ability to bear any weight.
Exam. Toxic versus playful; temperature; whether he will bear weight; hip held flexed and externally rotated; log-roll and abduction; knee, ankle, spine, abdomen (psoas, neuroblastoma mass), skin (rash, bruises), lymph nodes. Unilateral refusal to walk is a hip-until-proven-otherwise problem in this age group, but the knee and the spine can refer.
Must-not-miss versus most-likely.
| Diagnosis | Typical pattern | Immediate implication |
|---|---|---|
| Septic arthritis of the hip | Fever, non-weight-bearing, toxic or ill, markedly elevated ESR/CRP, often elevated WBC (Kocher/Caird-type predictors) | Orthopedic emergency — aspiration, drainage, antibiotics after cultures in the right setting. Do not send home as “toxic synovitis.” |
| Transient (toxic) synovitis | Post-viral, relatively well, low-grade or no fever, still able to bear some weight, modest or normal inflammatory markers | Most-likely in a well child after URI — but it is a diagnosis of exclusion after you have thought about infection |
| Osteomyelitis | Point tenderness, fever, limp that may still allow some walking | Labs plus MRI in the appropriate setting; do not stop at a normal x-ray |
| Leukemia (or other marrow disease) | Limp with night pain, pallor, bruising, fever, abnormal CBC; x-rays may be normal | CBC is part of the workup when systemic features exist; hematology referral, not NSAIDs and a sports sticker |
| SCFE | Usually older (roughly 10–16 years), often obesity, hip or referred knee pain, obligatory external rotation | Not the typical 3-year-old — but if the stem’s age jumps, non-weight-bearing and urgent ortho; missed SCFE is a life-of-the-joint error |
| JIA, Lyme arthritis, toddler’s fracture, NAI | Morning stiffness; endemic tick plus knee swelling; spiral tibial fracture in a cruiser; injury history that does not fit | Each is in the differential once septic hip is addressed |
Labs and imaging combined. For the toxic, febrile, non-weight-bearing toddler: CBC, ESR, CRP, and urgent imaging (hip ultrasound for effusion; AP and frog-leg pelvis as indicated) on the way to orthopedics, not as a reason to delay. Kocher criteria (non-weight-bearing, fever, ESR, WBC; CRP often added) do not give a magic single cutoff PNCB will print, but multiple positive predictors mean you do not diagnose toxic synovitis. For the well, weight-bearing, post-viral child with low inflammatory markers, close follow-up of transient synovitis is reasonable after you have excluded red flags. If pallor, bruising, or cytopenias appear, the unifying diagnosis may be leukemia, and the “hip ultrasound” is no longer enough.
That is II.E: history, exam, labs, and imaging are one argument, not four separate hobbies.
PNCB tests one best answer
When options include “order labs, image, refer, and start antibiotics in the office,” ask which action is the priority. For high-probability septic hip, the best next step is urgent orthopedic/ED management, not a primary-care oral cephalexin prescription and a 10-day MRI wait. For classic viral URI without red flags, the best next step is counseling and follow-up precautions, not a CT. You may think of several steps; you choose the one that is most urgent, most specific, or most consistent with the working diagnosis.
Exam traps. (1) Treating the first obvious URI and missing foreign body or airway emergency. (2) Calling every limp toxic synovitis because last week’s lecture said it is common. (3) Ordering every test on the list instead of the best next step. (4) Documenting a confirmed rare diagnosis you have not proven. (5) Failing to call cardiology, neurology, or genetics when the data have already left primary-care range.
Integrate, prioritize must-not-miss, use consultants on purpose, and pick one best next step. That is how Domain II.E is written — and how it is tested.
A 3-year-old refuses to walk, has a temperature of 39.2°C, and holds the hip flexed and externally rotated. ESR and CRP are markedly elevated. What is the best next step?
Which approach matches the vignette method of differentials on a PNCB CPNP-PC item?
A preschooler brought in for “a cold” is toxic, drooling, and sitting in a tripod posture. Using two-step thinking and one-best-answer logic, what should the CPNP-PC do?