Navigating Diagnostic Uncertainty: How to Think When the Diagnosis Isn't Clear

This article reflects my clinical approach to diagnostic reasoning—not legal advice. Adapt these frameworks to your scope, setting, and clinical judgment.

One of the most common questions I get is this: how do you think when symptoms don't fit neatly; when the story shifts, when you're not sure what you're treating?

That's the clinical judgment piece most programs don't teach.

You're taught DSM criteria, medication mechanisms, treatment algorithms. What you're not taught is the cognitive process underneath: how to think when information is incomplete, conflicting, or ambiguous.

That gap, between criteria and judgment, is where most diagnostic errors live.

Here's the framework.

Start with base rates, not pattern matching

Most diagnostic errors happen because clinicians lock onto the first diagnosis that fits the symptoms, without considering how common or rare that diagnosis actually is.

Bayesian reasoning: Update your belief based on probabilities, not just symptom matching. Base rates = how common a condition actually is in the population, not just how well symptoms match.

Example: 30-year-old with low mood, low energy, anhedonia, weight gain. Also mentions "feeling disconnected from reality sometimes."

Early clinician thinks: "Dissociation plus depression…could this be dissociative identity disorder?"

The base rate problem:

• Major depression: ~7% lifetime prevalence

• Dissociative identity disorder: <1%

Even if symptoms "fit" DID, the probability is still extremely low. Major depression with dissociative features is far more likely.

The Bayesian approach:

1. Start with the most common explanation (major depression)

2. Treat that first

3. Reassess if the presentation doesn't match what you'd expect

Rare things are rare. Don't diagnose zebras when you hear hoofbeats.

This isn't about dismissing rare diagnoses—it's about earning them.

Common heuristics that mislead diagnosis

Heuristics are mental shortcuts. They speed up decision-making but fail in predictable ways.

Availability heuristic: You recently saw a patient with lithium toxicity. Next patient comes in with tremor and confusion. You think "lithium toxicity", but this patient isn't on lithium. The recent case made toxicity feel more likely than it actually was.

Anchoring: Patient's intake says "bipolar disorder." You lock onto that diagnosis. During the interview, they describe classic unipolar depression: no mania, no family history, misdiagnosed years ago. But you're anchored on "bipolar" and reluctant to revise.

Confirmation bias: You suspect borderline personality disorder. You ask about impulsivity, unstable relationships, emotional dysregulation. They say yes. You don't ask about trauma, ADHD, or bipolar…all of which could explain the same symptoms. You confirmed your hypothesis but didn't test alternatives.

Representativeness heuristic: A patient looks like "textbook borderline": fear of abandonment, intense relationships, chronic emptiness. But you didn't check: Are they in an abusive relationship? Do they have ADHD? Trauma history? The presentation fits, but you haven't ruled out other explanations.

When you feel confident about a diagnosis, ask:

• What heuristic am I using?

• What evidence would disprove this?

• What am I not asking about?

Use working diagnoses, not diagnostic certainty

You don't need to be certain before you start treatment. You need a working diagnosis—your best current hypothesis based on available data.

Working diagnosis approach:

1. What's the most likely diagnosis based on symptoms and base rates?

2. Start treatment for that diagnosis

3. Define what would make you change your mind

4. Reassess based on response

Example: Depression vs bipolar depression

28-year-old with depressive episode. No clear mania history. Reports "sometimes I get really energized and don't sleep much."

Don't wait for certainty. Treat the most probable explanation:

• Working diagnosis: Major depression (base rate ~10-15% vs bipolar ~1-2%)

• Start: SSRI

• Monitor: for mood instability, mixed features, activation

• Reassess: if bipolar features emerge

What you'd document:

"Patient presents with depressive episode. Reports periods of 'high energy' but duration, severity, and associated functional impairment are unclear; no established manic symptoms. No family history of bipolar disorder. Working diagnosis: major depressive disorder. Starting sertraline 50mg. Will monitor for mood instability and reassess if manic or mixed features emerge."

This shows:

• You considered alternatives

• You have a rationale

• You have a reassessment plan

When the story shifts

Shifting narratives aren't always "lying" or manipulation. They can mean:

• Patient is organizing their experience over time

• Shame prevented full disclosure initially

• Symptoms are actually evolving (e.g., depression → mixed features)

• New stressors are surfacing different symptoms

When the story shifts, update your working diagnosis. Changing your mind isn't failure. It's responding to new information.

Document uncertainty without sounding incompetent

Bad: "Not sure if this is depression or bipolar."

Good: "Differential includes major depression vs bipolar depression. No clear history of mania, but patient reports periods of increased energy (duration and functional impairment unclear). Treating as major depression initially with close monitoring for mood cycling. If mixed features or clear hypomania emerge, will reassess for bipolar spectrum."

You're not hiding uncertainty. You're showing you're managing it deliberately.

The goal isn't to eliminate uncertainty—it's to show you're managing it.

A quick "in the moment" tool

When you're stuck mid-visit, use this three-question reset:

1. What's the most likely explanation given the current information?

2. What's the most dangerous thing I need to rule out, or what else could explain these symptoms?

3. What new information would most change my plan?

The reasoning divide

Early clinicians ask: "What does this look like?"

Experienced clinicians ask: "What's most likely? What would make me wrong? How will I know?"

That's clinical judgment. It's not about being right every time. It's about reasoning clearly when the answer isn't obvious.