The Art of Clinical Reasoning

Clinical Reasoning

a complex cognitive process that is essential to evaluate+manage a pt’s medical problem

Dual Reasoning Process (Intuitive vs Hypothetico-Dedcutive/Rational)

not exclusive→clinicians integrate elements of both processes

intuitive

• quickly+automatically

• mental shortcuts

• rapid+intuitive

• habitual response systems

• hard to change

• used by experts

hypothetico-deductive (rational)

• explicity anaytical approach

• deliberate+controlled thought process

• logic+probability→reach conclusions

• more time

• more cognitive effort

Anchoring Bias

tendency to perceptually “lock onto” noticeable features in the patient’s initial presentation

too early in the diagnostic process

get more information later→failure to adjust

Availability Heuristic

assumption that a diagnosis is more likely or more frequently occuring if it more readily comes to mind

considering easily remembered diagnoses more likely irrespective of prevalence

Confimation Bias

seeking supportive evidence for a diagnosis at the exclusion of more persuasive information refuting it

seeking data to confirm, rather than refute the initial hypothesis

Diagnostic Momentum

prioritizing a diagnosis made by prior clinicians discounding evidence of alternative explanations

Framing Effect

interpretation of information is influenced heavily by the way in which the information is presented (framed)

Representation Error

failure to take prevalence into account when estimating the probability of a diagnosis

Visceral Bias

visceral arousal (negative and positive feelings towards pts)→leads towards poor diagnostic decisions

Representativeness

ignoring atypical features that are incosistent with the favored diagnosis

Base Rate Neglect

pursuing “zebras”

Premature Closure

stopping the diagnostic process too soon

Suggested Rules for Good Decision Making

• slow down

• be aware of base rate of disease for items on your differential diagnosis

• consider what data is truly relevant

• actively seek alternative diagnoses

• ask questions to disprove, rather than confirm, your current hypothesis

• remember you are often wrong, consider the immediate implications of this

Differential Diagnosis List

application of clinical reasoning to sort through the findings→create a list of possible causes for the pt’s issue

includes

• likely dxs

• critical dxs→can’t miss dxs

Working Diagnosis

diagnosis that best matches your pt’s issue at that time

• defined as precisely as data allows

• may be a leading diagnosis or still a hypothesis

• foundation for creating the tx plan

Process of Clinical Reasoning: Symptom to Diagnosis

1. identify the problem

2. frame the differential

3. organize the differential diagnosis list

4. limit teh list

5. explore possible diagnoses using the H+P findings

6. rank the differential list

7. test the hypotheses

8. re-rank with new data

9. test the new hypotheses

Step 1: Identify the Problems

1. chief complaint

2. acute sx+PE+DX+vitals

3. chronic active problems/dxs

4. historical problem/PMH

Step 2+3: Frame/Organize the Differential Diagnosis

develop differential diagnosis framework in a way that helps you recall

problem-specific framework of subcategories:

anatomic

• what organs/organs sytems are in this region of the CC

organ/system

• what are the cardiac/pulmonary/hepatic causes for the CC

physiologic

• what things/Es can cause the CC

pivotal points

• taking key points to narrow

• opposing descriptors to compare/contrast

-abdominal pain→pelvic vs epigastric
-leg edema→UL vs BL

Step 4: Limit the Differential Diagnosis

from large+complete differential→more focused set of dxs

gather more information→eliminate dxs that are unlikely

Step 5: Explore Diagnoses Based On History+Physical Exam (PE)

look for clinical clues that point towards most likely diagnosis

don’t be led astray by a lack of textbook/classic findings

• history details
  -current
  -previous

• exam findings
  -RFs

Step 6: Ranking the Differential Diagnosis

limitations to each approach→clinicans often integrate these approaches to rank the differentials→leads to step 7

4 approaches

• possibilistic→consider all known causes equally likely→test for all of them
  -not useful

• probabilistic→consider most likely dxs
  -higher pretest probability→dx is present before further testing is done

• prognostic→consider most serious dx

• pragmatic→consider the diagnoses most responsive to tx first

Step 6: Leading Hypothesis

the most likley dx

based on

• prevalence

• demographics

• RFs

• signs+sxs

test requirements

• high specificity

• high LR+

Step 6: Active Alternatives

life threatening dxs→can’t miss dxs

high prevalence→m/c dxs

reasonably likley dxs

based on:

• demographics

• RFs

• signs

• sxs

test requirements

• high sensitivity

• very low LR-

Step 6: Excluded Hypotheses

dxs→disproved

based on

• RFs

• signs

• sxs

• prior DX results

test requirements

• no further testing necessary

Step 7: Testing the Hypothesis (Only If Necessary)

additional data to confirm your hypotheses→dx testing

consider how much the test will change the probability the pt has the dx in question

beware of the “pan scan”
-exception→legit trauma patient

ordering DXs→be selective+intentional

why are you ordering the test:

• rule-in

• rule-out

• manage/monitor

Step 7: Considerations With DX Testing

test as little as possible

how certain that hypothesis is correct

how much more certain before starting tx

use pretest probability (probability that a dx is present before further testing is done)

Pretest Probability

probability that a dx is present before further testing is done

-is it high/low enough that no further testing is needed

• clinical decision rule (CDR)→use a doccx tool
  -often a list of criteria

• use information about prevalence of E of a sx

• use overall clinical impression (“clinical judgement”)

Step 7: Consider Harm of Testing

any risks present→consider pretest probability+what next steps should be

• harm of missing dx
  -life-threatening dx

• harm of testing
  -pain
  -radiation

• harm of tx
  -chemo
  -AE
  -C

Treatment Threshold

probability above which the dx is so likely→tx pt without more testing

Test Threshold

probability below which the dx is so likely→r/o dx without more testing

Posttest Probability

when the test is so helpful→pushes past one of 3 thresholds

1. test threshold

2. treatment threshold

3. pretest probability

Validity

does the test accurately identify whether a pt has a dx

Gold Standard/Reference Test

best measure of whether a pt has a dx

Positive Predictive Value (PPV)

probability that a person with a positive test has the dx

Negative Predictive Value (NPV)

probability that a pt with a negative DX doesn’t have the dx

Incidence vs Prevalence

incidence

• rate of new/newly occuring cases of a dx in a population during a specific period of time

• “how many new cases are happening”

• helpful for understanding dx causation+outbreaks

prevalence

• total burden of existing cases of a dx in a population during a specific period of time

• “how many people have the dx right now?”

• chronic/incurable dxs→high prevalence

• short-lived dxs→low prevalence

Testing Result Interpretations

true positive

pt with positive DX→has dx

true negative

pt with negative DX→doesn’t have dx

false positive

pt with positive DX→doesn’t have dx

false negative

pt with a negative DX→has dx

Sensitivity vs Specificity

sensitivity

• percentage of pts that have a true positive

• high sensitivity→low chance of false negatives

• negative result→true negative→r/o dx

• SnOUT

specificity

• percentage of pts who have a true negative

• high specificity→low chance of false positive

• positive result→true positive→r/i dx

• SpIN

Likelihood Ratio

likelihood that a given test result would occur in a pt with the dx compared with the likelihood that the same result would occur in a pt without the dx

how much a test result changes the dx probability vs the posttest probability

indicates how strong a specific test will raise/rule-in or lower/r/o the likelihood of the dx

Steps 8+9

step 8→continue to re-rank the DDX list

based on work-up:

• DX results

• changes in sxs/PE

• new history revealed

• new info/data

step 9→repeat process until you test the final hypothesis+diagnosis

Plan Outline

1. disposition
   -outpatient tx
   -inpatient tx
   -(ICU/step-down/telehealth/med-surg/etc.)

2. further work-up

3. consultations/referrals

4. tx
   -NP measures (therapeutic procedures+supportive care)
   -Mx/pharmacologic measures

5. follow-up

6. pt education/health promotion

Tx Adherence

the best plan→the one that the pt will understand+adhere to

collaborative approaches→foster better tx adherence+satisfaction

successful planning hingest on

• strong interpersonal skills

• understanding the pt’s life context

• ensuring their active participation in care decisions

Shared Decision Making+Patient-Centered Care

a collaborative process in which pts+healthcare providers work together to make healthcare decisions

pt-centered care

pts should be informed+involved in their care:

• communicate clincal evidence

• communicate risks vs benefits

• visual aids+decision-aiding tools

• use methods to help gauge understanding
  -teach-back method

• offer additional support/alternatives

• reduce framing+bias

Medical Decision Making (MDM)

the written record of the thought process+actions a clinician makes when dx+tx a pt

a summary of a provider’s clinical reasoning→explains your thinking all the way down to the cause of the problem

summary statement with more detail+explanation of clinical reasoning

Critically Appraising Clinical Evidence

3 basic questions (based on the evidence based working group)

1. are the results valid (believable)

2. what are the results (magnitude+precision)

3. how can you apply the results to pt care (generalizable)