Transcript Teaching Trainees to Think Geriatrically

D.W. Reynolds Foundation Annual Meeting
October 24, 2011
Amit Shah, MD University of Texas Southwestern Dallas
Cynthia Brown, MD, MSPH University of Alabama at Birmingham
Houman Javedan, MD Harvard University
What is a geriatrician?
“….as a geriatrician, I am by definition an
expert in subtlety and complexity”
Hazzard, W. “I Am a Geriatrician” JAGS 52:161, 2004
Teaching with aphorisms and pearls…..
 “Always remember the atypical presentation of disease.
The older patient often won’t have read the textbook.”
 “If you’ve seen one 80 year old, you’ve seen one 80 year
old.”
 “Occam’s Razor is dangerous”
 “Multifactorial etiologies demand multifactorial
solutions”
We can do better…
Teaching trainees to think
Outline for today
 Background about clinical reasoning
 Teaching of clinical reasoning in novice
trainees (medical students)
 Teaching and assessment of clinical
reasoning in advanced trainees
(fellows)
 Clinical Reasoning Case Practice (in a small group)
What is clinical reasoning?
AKA:
 problem-solving
 decision-making
 “judgment”
 diagnostic reasoning
 Major domain of clinical competence
 Thinking and decision-making processes that are used
in clinical practice
Two Major Types of Clinical Reasoning
 Analytic (hypothetico-deductive)
 What is generally emphasized in
teaching
 Bayes' theorem
 EBM
 Generation of rules
 Non-analytic
 Illness Scripts
 Pattern recognition
 “Experience”
 Subconscious/automatic
Figures from: Eva KW. What every teacher needs to know about clinical reasoning. Medical Education 2004; 39: 98–106
Which is better?
 Non-analytic plays a much larger role in clinical care
than we teach
 Non-analytic has not been shown to be inferior
 Excessive reliance on pattern recognition can cause
diagnostic errors
 Want to teach students to use both
 Example: EKG reading teaching study
Norman GR, Brooks LR, Colle CL, Hatala RM. The benefit of diagnostic hypotheses in clinical reasoning:
experimental study of an instructional intervention for forward and backward reasoning. Cognit Instruct
2000;17:433–48
Non-analytic Teaching
 Can teach illness scripts and pattern recognition
 One Model: Many cases, rapid fire method
 KBIT at TCOM (Dr. Frank Papa)
 Pattern matching and pattern discrimination
 Diagnostic competence is
 problem specific
 disease-specific
 Explains why PBL/CBLs have limitations
 Rapid Fire, multiple cases have a role (eg Prognosis app)
Papa FJ, Oglesby MW, Aldrich DG, Schaller F, Cipher DJ. Improving diagnostic capabilities of medical students
via application of cognitive sciences-derived learning principles. Medical Education, 41:419-425, 2007
Areas of Clinical Reasoning
Wong et al’s analytical model
 Hypothesis Cues
 Hypothesis Refinement
 Testing Threshold
 Treating Threshold
Geriatrically Thinking about Hypothesis Cues
 Atypical Presentation of Disease
 Importance of Corollary Informants
 Picking up subtle clues from the environment
 Home visits
Geriatrically Thinking about Hypothesis Refinement
 Multifactorial etiologies
 Syndromal Presentation
 Incidence specific to age groups
Geriatrically Thinking about Testing Thresholds
 Increased risk of harm from testing
 Increased likelihood of false positive tests
 Prognostication
 Goals of care
 Cost/benefit
Geriatrically Thinking about Treating Thresholds
 Polypharmacy/non-pharmacologic management
 Paucity of therapeutic evidence in advanced age
groups
 Individualized threshold/fluctuating threshold
 Importance of interdisciplinary care
 Impacts of goals of care
The Old Way of Becoming An Expert
Mismatch Between Teaching and Practice
What is Taught
What they observe
 Step-by-step approaches
 Quick, snap judgments
 Book knowledge
 “Wisdom” “Experience”
 EBM and Bayesian Analysis
 “Fast and Frugal” or “Flesh
 Hypothesis testing
and Blood” (real world)
 Pattern recognition and
seemingly automatic
retrieval from the
subconscious
 Shortcuts
 Thorough/ Luxury of Time
Clinical Guidelines and Cookbook Medicine
 Great for simple, straightforward, typical patient
 NOT SO GREAT FOR GERIATRICS!
 Help reduce variability in clinical practice
 But we teach the “heterogeneity of aging” :
 Speed rate new knowledge applied
 Can be good, but previous disasters with rapid adoption
of untested practice/medications in the geriatric patient
(eg, Vioxx, RALES trial, etc)
Why trainees love them:
 Clear answer of “what to do next”
 Clear targets / goals of treatment
 Simplify things / “take away the uncertainty”
Problems of Novice Trainees
 Analytic strategies more rigid and simplistic
 Over-reliance on algorithms
 Inability to account for uncertainty
What can we do?
 Teach which hypothesis cues are important in a given
context
 Refining hypotheses when additional data available.
How will we teach them?
 New National Curriculum: POGOe web-GEM
 Disclaimer: Both myself and Amit Shah are involved
in this initiative
 Teach trainees explicitly about thinking and
cognitive errors
 Our niche as geriatricians, given our patients
disproportionately suffer consequences of cognitive
errors!
POGOe web-GEM Curriculum
 Standardized Peer-Reviewed Curriculum
 34 cases in development
 Authors from numerous institutions
 Linked to AAMC Competencies
 At the 3rd/4th year medical student level of detail
 Have a “clerkship home” to allow integration for
schools without mandatory geriatrics rotation
 Emphasize core topics important to third year medical
students in the case’s “clerkship home”
POGOe web-GEMs and Thinking Skills
 Using the CASUS platform
 Same platform used by medU (SIMPLE, CLIPP)
 1.5 million cases completed to date
Clinical Reasoning Features:
 Hypothesis Generation
 Diagnostic Networks
 Hypothesis Refinement
POGOe web-GEM Curriculum
 Explicit focus on teaching novice trainees to “think
like a geriatrician”
 Syndromal Presentations
 Atypical Presentations
 Interprofessional approaches to evaluation and
treatment
Teaching about Cognitive Errors
 Well developed literature (from cognitive
psychology) about diagnostic errors
 Popularized by Dr. Jerome Groopman
 Explicit teaching to novice trainees may:
 help to develop good thinking habits
 demystify process of coming to diagnosis
Croskerry P. The importance of cognitive errors in diagnosis and strategies to minimize
them. Acad Med. 2003;78:775-780.
Croskerry P. Cognitive forcing strategies in clinical decision making. Ann Emerg Med.
2003;41:110-120.
Redelmeier DA. Improving patient care. The cognitive psychology of missed diagnoses.
Ann Intern Med. 2005;142:115-120.
How do cognitive errors occur?
 Typically not in isolation
 Cascade of sequential cognitive mistakes
 Much more common that novice trainees think
 Up to 10% of autopsies reveal a clinically relevant
diagnosis that was missed
 Misdiagnosis occurs 15-20% of the time; about 80% are
due to cognitive errors
 Technology not a solution
 Increased technology can increase misdiagnosis
Types of Cognitive Errors
 Over 30-40 types of cognitive errors described in
literature
 Common types of error easily taught during care of
geriatric patients:
 Premature Closure
 Framing or Diagnosis Momentum
 Availability
 Representativeness Errors
 Attribution
 Commission Bias and Omission Bias
Premature Closure or “Freezing”
 Once have something that “fits” you stop thinking
 Also has been called “satisficing error”
 Satisfy + suffice
 Examples in Geriatrics:
 Hospital patient with pneumonia who has a MI
 “The fracture most commonly missed is the second”
 Delirious patient with multifactorial etiology
(stopping work-up with positive U/A and miss MI or
meningitis)
Framing or Diagnosis Momentum
Once several doctors/specialists agree on a
diagnosis it is easier to perpetuate it rather
than take the time to question its accuracy
 In Geriatrics:
 “Chart Lore” for PMH
 “ED admits this delirious patient with a UTI
from the NH”
Availability Errors
 Choosing most likely or most memorable
diagnosis
 Over-estimation of frequency of vivid or easily
recalled events
 Under-estimation of frequency of ordinary or hard
to recall events
Representativeness
 Removing a disease from differential diagnosis list
because it does not match standard or usual
presentation
 A major problem in diagnosis of patients in whom
“atypical presentations” are common, like
geriatrics
Commission Bias
(inability to follow geriatric mantra: “Don’t just do something,
stand there”)
 Urge to act rather than do nothing even when nothing is
preferable.
 Rooted in belief that beneficence involves active
intervention
Omission Bias
(“They are just old”)
 The tendency towards inaction
 Events attributed to natural events of disease better than
those related to a physician’s intervention
 Rooted in non-malfescence
Attribution Errors
 Stereotyping
 Judgmental
 Gender Bias
 Racial Bias
 Age Bias
Teaching avoidance of cognitive errors
 Make thinking explicit – Think out loud!
 Feedback
 Reflective Practice
 Become comfortable with uncertainty
 Acknowledgement we get it wrong at least 10% and
up to 20% of time
 Metacognition: “Cognitive Pills for Cognitive Ills”
What is different about the
advanced learner?
 1- Larger non-analytic reasoning data base
 2- Already adopted a form of analytical model
 3- Due to variability of experience and graduate
medical education, variability of baseline knowledge
and skills
 1+2+3 = Adult Learner1
1. Kolb, David (1984). Experiential learning: Experience as the source of learning and development.
Englewood Cliffs, NJ: Prentice-Hall.
What is different about the new
Geriatric learner?
 Less likely to be exposed to geriatric clinical reasoning
domains described at beginning of presentation
 Well trained in clinical guidelines not based on the
elderly (worked hard and invested heavily in achieving
recognition for it)
 Less likely to have experience in a nursing home or
rehab setting
Where to Begin?
 Clinical Reasoning studies
 Think out loud2,3
 Ability to communicate reasons
2 Kassirer, J.P., Wong, J.B, Kopelman R.I. (2009). Learning Clinical Reasoning. New York: Lippincott Williams
& Wilkins
3 Alberdi E, Taylor P, Lee R. Elicitation and representation of expert knowledge for computer aided
diagnosis in mammography. Methods Inf Med. 2004;43(3):239-46.
Can You Teach Advanced Learners?
 Adult Learning Curricula2
 Experiential learning (why, how, what, if)3
 Case based learning
 Real life application (a.k.a. relevant)
 Learning Objectives
 Reflection4
2 Armstrong, Liz and Kegan, Robert. Harvard Macy Institute Curriculum for Educators.
3 Armstrong E, Parsa-Parsi R. How can physicians' learning styles drive educational planning? Acad Med. 2005 Jul;80(7):680-4 .
4 Stark P, Roberts C, Newble D, Bax N. Discovering professionalism through guided reflection. Med Teach. 2006 Feb;28(1):e25-31
What cognitive aspect makes it
specifically geriatric?
 If you had to develop a Geriatric Clinical Reasoning
“Mini-Cog” to screen for geriatric clinical executive
function what would be the clock draw?
 UNCERTAINTY
Clinical Reasoning Curriculum
 Task:
 A validated adult learning method that incorporates
recognized clinical reasoning methods targeting
geriatric reasoning by using uncertainty as a core theme.
Clinical Reasoning Curriculum
 Design:
 A case based learning format
 Part of weekly didactics
 Introductory lecture describing basics of analytical and
non-analytical clinical reasoning
 Reflective session half way through
 Provide copy of learning objectives with relevant list of
geriatric knowledge and skills topics
 Provide template of Wong model presentation
PowerPoint for case presentation
Clinical Reasoning Curriculum
 Design:
 Designated fellow will present a real clinical case
encountered during rotation or NH overnight call
 After presenting the case the presenting fellow will
choose question from one of the Wong model domains
 The group will then divide into groups of 2 to 3 fellows
and discuss the question for 5-10 minutes
Clinical Reasoning Curriculum
 Design:
 A representative from each group will present the
reasoning behind their clinical decision to the whole
group
 The presenting fellow will summarize each groups key
reasons to confirm they are understood correctly
 The presenting fellow will share the outcomes of the
case if known
 The presenting fellow will present the results of a
literature search if any relevant evidence exists
Learning Objectives
Template
 Hypothesis Cues
 What signs and symptoms are significant and why?
 Hypothesis Refinement
 What are your likely diagnoses and why?
 Testing Threshold
 What tests would you request and why?
 Treatment Threshold
 What treatment would you implement and why?
Clinical Reasoning Curriculum
 How is this different from morning report?:
 1- Case must not have a single correct answer (must
include uncertainty)
 2- The level of uncertainty should increase over the
course of the year
 3- The focus is on why a decision should be made not
the list of options?
Practice Case
Mrs. M
 PC: Fatigue
 HPC: Mrs. M, an 80 year old woman presents with her daughter,
Sarah, to outpatient clinic complaining of fatigue. She has been
feeling more tired over the last 8 months but things have gotten
to a point that Sarah is concerned. “Mom is just not herself” she
says. Sarah says Mary is tired “all the time”. She would usually go
out and walk around the block but now she only makes it to the
driveway mailbox and back. She has had no weight loss, appetite
is fair, no chest pain, no palpitations, no cough, no fever, no
chills, no dysuria, no increased urinary frequency, no change in
bowel habit.
 PMHx:
 Diabetes, CAD s/p MI 2001 stent to circumflex, Moderate mitral
regurgitation, Hypertension, Osteoarthritis
 Social Hx
 Lives with her daughter and son in-law. Daughter is with her 24
hours a day.
Functional history: Independent for most ADLs until 1 month ago,
now needs help with dressing and showering. Needs help with
instrumental activities of daily living for the past 3 years. Uses cane
when walking out of the house.
 Medications:
 Lisinopril 10mg
 Metoprolol ER 75mg
 Glipizide 5mg BID
 Hydrochlorothiazide 12.5mg
 Aspirin 81mg, Tylenol prn
 Allergies: NKDA
 Physical Exam
 Gen: Pleasant elderly female sitting comfortably in the chair, looks quietly
at the ground
Vitals: P 60 regular, BP 130/70, T97.7, Sat 96% RA
HEENT: no cervical lymphadenopathy, dry oral mucosa
CVS: No JVP, HS: S1 + S2 + 2/6 pansystolic murmur at apex radiating to
axilla
Resp: Sparse inspiratory crackles at both bases
Abdomen: Soft, non-tender, bowel sounds present, no hepatosplenomegaly
Ext: 1+ edema, no cyanosis, no clubbing
Neuro: CN intact, Power 4+/5 all four limbs, Reflexes difficult to elicit
Cognitive: Montreal Cognitive Assessment 25/30- missed last trail, missed
2 serial sevens, 2 delayed recall missed but able to recall with categorical
cue, missed date
Get up and go: failed, unable to get out of chair
 Labs
 1 month ago: WCC 11, Hct 32, MCV 85, Plt 250, Na 132, K 4.0, BUN 24, Cr 1.4
 Hypothesis Cues
 What signs and symptoms are significant and why?
 Hypothesis Refinement
 What are your likely diagnoses and why?
 Testing Threshold
 What tests would you request and why?
 Treatment Threshold
 What treatment would you implement and why?
 Hypothesis Refinement:
 What are your likely diagnoses and why?
Assessment
 No simple validated MCQ- but is this appropriate?
 How are we assessing Clinical Reasoning currently?
 Rotation evaluations bring in high variability of
confounding variables if trying to hone in on clinical
reasoning itself
 Similar to cognitive testing => we get an overall picture
without stressing necessarily
Assessment
 Reflection important and specifically placed at mid-
way point to address concerns sooner than later
 Faculty observation and participation during session
 Guided observation based on learning objectives
Faculty Observer Examples
 How well learner incorporates geriatric clinical
reasoning domains described at beginning of
presentation? (knowledge + skills)
 How large is a learner’s geriatric non-analytical data
base and does it grow over time? (skills)
 Anxiety with regards to uncertainty (world of “rule out
MI”) (attitudes)
What CR adds to assessment
 Able to observe a trainees knowledge base,
communication, and cognitive style (analytical vs nonanalytical).
 The Good Student
 Be able to assess which geriatric domains are missing
 The Bad Student
 Be able to identify where the deficiency is in a more
controlled environment
References
 Elstein AS, Schwarz A. Clinical problem solving and diagnostic decision
making: selective review of the cognitive literature. BMJ. 2002;324:729-732.
 Gladwell M. Blink: The Power of Thinking Without Thinking. New York, NY:
Little, Brown and Company; 2005.
 Groopman, Jerome E. How Doctors Think. Houghton Mifflin, 2007
 Graber ML, Franklin N, Gordon RR. Diagnostic error in internal medicine.
Arch Intern Med. 2005;165:1493-1499.
 Croskerry P. The importance of cognitive errors in diagnosis and strategies to
minimize them. Acad Med. 2003;78:775-780.
 Croskerry P. Cognitive forcing strategies in clinical decision making. Ann
Emerg Med. 2003;41:110-120.
 Redelmeier DA. Improving patient care. The cognitive psychology of missed
diagnoses. Ann Intern Med. 2005;142:115-120.
 Eva KW. What every teacher needs to know about clinical reasoning. Medical
Education 2004; 39: 98–106
 Kassirer, J.P., Wong, J.B, Kopelman R.I. (2009). Learning Clinical Reasoning.
New York: Lippincott Williams & Wilkins