Transcript Slide 1

Improving Evidence Based Clinical
Decision Making
Larissa Lucas, MD, FACP
Instructor, Harvard Medical School
5th Simposio Internacional de Seguridad del Paciente
Cali, Colombia
March 13, 2014
Conflict of Interest

Senior Deputy Editor of DynaMed, an educational
product of EBSCO Information Services
Objectives




Explain the connection between cognitive
processes and errors in medical decision making
Identify the challenges to answering clinical
questions and making evidence based clinical
decisions
Describe the pros and cons of available tools
Identifies strategies for a better system of
information management
According To Traditional Stories, How Many Goats
Did Noah Take Into The Ark?
The Answer?

2?

7?

10?

14?

20?
According To Traditional Stories, How Many Goats Did
Noah Take Into The Ark?
O f e v e r y c le a n b e a s t t h o u s h a lt t a k e t o
sevens
thee by
, t h e m a le a n d h is
f e m a le : a n d o f b e a s t s t h a t a r e n o t c le a n
b y t w o , t h e m a le a n d h is f e m a le . O f f o w ls
a ls o o f t h e a ir b y s e v e n s , t h e m a le a n d
t h e f e m a le ; t o k e e p s e e d a liv e u p o n t h e
f a c e o f a ll t h e e a r t h .
G e n e sis. Ch 7 ; v 2
How Do You Approach A Problem?

Intuition – everyone knows this one

Ask an “expert” – haven’t read about this in a while

Look for the answer in reliable sources – never
heard of this type of problem before
Everyone Knew The Answer

Other Historic TRUTHS in medicine
 Post
menopausal hormone replacement therapy
 Increased
breast cancer and MI
 Anti-arrhythmic
medications for premature ventricular
complexes (PVC) post MI
 Increased
 Newborns
 Increased
 Reduce
mortality in CAST trial
should sleep prone (on their stomachs)
SIDS, now all newborns sleep on their backs
HbA1c as low as possible
 Increased
mortality in ACCORD trial
Things We Do Differently Now





Wear gloves, wash hands, universal precautions
Surgery checklists
Double check a patient’s identification before
dispensing prescriptions
Double check a patient’s blood type before
transfusion
Operating room device counts before and after
surgery
No One Is Exempt



We are all human
“To Err is Human”
We all make cognitive errors
To Err is Human…



Early 1990s: errors were either due to “ignorance”
or “implementation”
1999: To Err is Human, IOM report shifted focus
from individuals to systems of care
Now errors are:
 Systemic
 Cognitive
Croskerry P. Diagnostic Failure: A Cognitive and Affective Approach. Advances in Patient Safety: From Research to
Implementation (Volume 2: Concepts and Methodology). Rockville (MD): Agency for Healthcare Research and Quality (US); 2005
Feb.
Conditions for Cognitive Errors







Ambient conditions
Affective state
Past experience
Patient factors
Team factors
Fatigue
Violation producing factors
Croskerry P. Diagnostic Failure: A Cognitive and Affective Approach. Advances in Patient Safety: From Research to
Implementation (Volume 2: Concepts and Methodology). Rockville (MD): Agency for Healthcare Research and Quality (US); 2005
Feb.
Errors of Knowledge



We don’t know what we don’t know
Our current knowledge is wrong or outdated
We don’t know where or how to get the information
we need
Cognitive Errors in Oncology Cases



56 cases retrospectively reviewed
95% of cases had at least 1 cognitive error
Common cognitive errors
 Work
up strategy
 Interpretation of results
 Determining whether symptoms had resolved

Application of guidelines would have prevented
an error in 66%
J Gen Intern Med. 2012 Nov;27(11):1416
Why Didn’t They Follow The Guideline?




Didn’t think they had to follow the guideline
Couldn’t find the guideline
Didn’t know which guideline to read
Didn’t know how to evaluate the guideline
How to Keep Up with Guidelines?

Rely on your institution to provide algorithms


Monitor guideline-producing organizations yourself


Too much work
Government-run guideline websites


Decision support systems
Free access
Online reference resources
Alerts
 Valid trustworthy process
 Presented in relevant context

Cognitive Errors in ER Cases


79 claims reviewed
missed diagnoses were cognitive factors in 96%
 Error
of judgment in 87%
 Lack of knowledge in 58%
 Lapse in vigilance or memory in 41%
Ann Emerg Med. 2007 Feb;49(2):196
How do you maintain your knowledge base?



Consult a colleague
Attend courses
Regularly refer to an online reference
 Easy
to use and navigate
 Valid and relevant
 Comprehensive
Strategies To Reduce Cognitive Error





Provide adequate resources to make good
decisions
Step back and reflect
Decrease reliance on memory
Simplify tasks to reduce cognitive effort
Minimize time pressure
Croskerry P. Diagnostic Failure: A Cognitive and Affective Approach. Advances in Patient Safety: From Research to Implementation
(Volume 2: Concepts and Methodology). Rockville (MD): Agency for Healthcare Research and Quality (US); 2005 Feb.
Technology To Reduce Errors

Decision alerts within the EMR
 Does

Evidence based reference resources
 Do

this cause alert fatigue?
they provide too much, take too much time?
Evidence based algorithms
 How
do you know if they are current and valid?
 How do you address variances?
 Not
every patient fits a checklist
 Not every disease fits an algorithm
What Resources Do You Currently Use?

Colleagues

Online resources: DynaMed, UpToDate, EMedicine,
MDConsult, Harrisons Online

Mobile resources

PubMed

Google

Others……..
Tools To Manage Information Needs

Alerting Tools: send you important updates that
you need to read and store “Foraging”
Hunting Tools: used to search for information
_________________________________________
 Relevance: information must be put into clinical
context or you will not remember it
 Validity: information must be evaluated and
validated by a trustworthy process

Keeping
How Do Youup
Keep Up With It All?
up with
7 relevant
ublished
a family
physician
ed to
7.5 hours
he
iterature
Alper BS et al, 2004
Usefulness
Systems
Computerized decision
support
Summaries
Evidence-based databases
Synopses
Evidence-based journal
abstracts
Syntheses
Systematic reviews
Studies
Original journal articles
Tools
Alerting tools:
 Keep you current
 Use for most common
clinical scenarios
 Must be relevant

Caution: not all new
evidence should change
your practice
Shaughnessy, Slawson, Information Mastery
Hunting tools:
 Help you answer questions
 Easy to use
 Comprehensive
 Evidence labels

Caution: what is the
methodology used?
Using your instinct



Every spring a certain pond is covered in lilly-pads
The lilly-pads double in area every day
If the pond is covered
100% after 48 days,
how many days is it
covered 50%?
Case 1




59-year-old woman with >30 pack-years smoking
She is very healthy otherwise and active
She has just quit smoking but is concerned about her
risk of lung cancer
Should she be screened for lung cancer?
National Lung Screening Trial
Guidelines Followed




American College of Chest Physicians and American
Society of Clinical Oncology (ACCP/ASCO)
 JAMA 2012 Jun
American Lung Association (ALA)
 ALA 2012 Apr
American Cancer Society
 CA Cancer J Clin 2013 Mar-Apr
American Association for Thoracic Surgery
 J Thorac Cardiovasc Surg 2012 Jul
How Would You Know?



Your information resource alerted you
Your EMR suggested it using patient data
You reviewed the information in a online resource to
keep yourself current
Sometimes research changes
and we do more or do less………
Case 2


62-year-old man with well-controlled diabetes and
hypertension well controlled on HCTZ
He is having preoperative evaluation for open
colectomy for colon cancer
 Treadmill
cardiac stress test performed for chest pain 1
year ago was normal
 He has had no symptoms of chest pain in 1 year
Would you prescribe a preoperative beta-blocker?
Preoperative Beta-blockers
•2009 Update driven by new research:
•Harm shown in POISE Trial
•Benefit shown in DECREASE Trial
•2011 DECREASE Trial discredited
•2013 Heart 2013 Jul Systematic Review
•2013 Aug – joint statement released
In 2011, however, faith in the reliability
of the DECREASE trials was shattered
as a result of a scientific misconduct
scandal centering on the principal
investigator of the studies
Preoperative Beta-blockers
Sometimes research changes
and we and stop doing unnecessary
things to our patients………
Changes To Reduce Overuse
The OLD way



Preoperative Chest XRAY
“Whole body” diagnostic
CT for all trauma patients
Axillary lymph node
dissection in patients with
breast cancer
The NEW Way



Not for routine ambulatory
patients without history or
physical to suggest a
problem
Not for patients with minor
trauma or single system
trauma
Not for stage I/II if the
sentinel lymph node biopsy is
negative
Choosing Wisely – American College of Surgeons
(http://www.choosingwisely.org/doctor-patient-lists/american-college-of-surgeons/_)
MRI for Headache
Government Sources
Case 3




28-year-old healthy woman being seen for routine
annual physical
She started sexual activity at age 20 with one partner
She has never been pregnant and never has had a
sexually transmitted disease
Her last Cervical Cytology exam was 2 years ago and
normal
What test does she need: Cytology, HPV,
Nothing?
Depends On Where You Are

ACOG/ ACS/ASCCP/ASCP


For women aged 21-29
 Cervical cytology every 3 years
 NO HPV testing
World Health Organization



Start at age 30
Screen with HPV if resources available
or
Screen with cytology if colposcopy is available for follow up
How global and how relevant is your online resource?
Improving Your Approach To Decision
Making

How confident are you in what you know? How current is
your information?

Slow down

Know when to “trust your gut” and when to investigate and
confirm

Verify and update your knowledge on the most common
problems you see

Use the information you learn

Accept that you cannot know everything
Online Resources




There are many!!
Vary in breadth and depth
Some have alerting features to keep you current
Some have mobile versions
Evaluation Of Resources

Criteria for evaluation

In-line references

Transparent policy for
methodology

Date stamp on content

Indication of “new evidence”

Alerting feature

Federated search of external
content

Not included in this study
but some would consider
important

Peer review

Evidence grade labeling
Haynes, RB Journal of Clinical Epidemiology 2012
Comparing Resources
Reproduced with permission from Haynes, RB Journal of Clinical Epidemiology 2012
Data collected March 2011
Breadth of Coverage
100%
90%
80%
% topics covered
70%
60%
50%
40%
30%
20%
10%
83%
82%
70%
63%
60%
58%
48%
47%
33%
0%
Error bars: 95% Confidence Interval
Reproduced with permission from Haynes, RB Journal of Clinical Epidemiology 2012 Data collected March 2011
25%
Haynes, B. How Current Are Leading Evidence-Based Medical Textbooks? JMIR, December 2012
Mobile
Improving Your Approach To Decision
Making



Technology is not the answer for all cognitive errors
Technology can support the process of decision making
Find a tool/tools that meets your needs
Alerting to keep you updated on guidelines and practice
changing evidence
 Breadth and depth of content when you need to look
something up


Evaluate yourself when you are making a decision:
How confident are you in what you know?
 How current is your information?
 Verify and update your knowledge
 Accept that you cannot know everything

Muchas Gracias!