Transcript Evidence-Based Diagnosis in Physical Therapy

Evidence-Based Diagnosis
in Physical Therapy
Julie M. Fritz, PhD, PT, ATC
Department of Physical
Therapy
University of Pittsburgh
What is Diagnosis?
“The anatomic, biochemical,
physiologic, or psychologic
derangement”
DIAGNOSIS
Labeling
Pathology
What is Diagnosis?
“Diagnosis is the term which names the
primary dysfunction toward which the physical
therapist directs treatment” (Sahrmann, 1989)
DIAGNOSIS
Planning
Treatment
What is Diagnosis?
•
•
Medical Diagnosis:
•
Herniated Disc
•
CVA
Physical Therapy Diagnosis:
•
Right-sided radiculopathy centralizing
with repeated extension
•
Left-sided hemiplegia - Brunnstrom
Stage III: all movements in synergy with
marked spasticity
Three Strategies of Clinical
Diagnosis
•
Pattern recognition
•
Complete history and physical
examination
•
Hypothetico-deductive strategy
Pattern Recognition
•
Instantaneous realization that the patient
conforms to a previously learned pattern
of disease
•
Usually reflexive, not reflective
•
Usually cannot be explained to others
•
Argued to be “learned” on patients and
not “taught” in lecture halls
Complete History and Physical
(Exhaustion)
•
The pain-staking search for (but paying
no immediate attention to) all the facts
about a patient.
•
Method of a novice
•
Impractical and inefficient
Hypothetico-Deductive Method
•
The formulation, from the earliest clues of
a “short list” of potential diagnoses.
•
Subsequent tests are performed which will
most likely reduce the length of the list.
•
Requires an understanding of probability
(zebras versus horses).
Exhaustive vs. HypothesisDriven Approach
•
Exhaustion
•
•
empty the mind of all
preconceived notions
•
watch “nature in action”
•
draw conclusions after
all the facts are in
Hypothesis-Driven
•
bold hypotheses are
proposed, then
exposed to severe
criticism
•
requires
understanding of
confirmatory/disconfirmatory tests
Gathering Diagnostic Data for
a Hypothesis-Driven Approach
•
Complete versus exhaustive data
gathering
•
Must know what is good data
•
The importance of confirmatory and
disconfirmatory data
•
Rarely is one test sufficient
Appraising the Literature
Regarding Diagnostic Tests
•
The effectiveness of a hypothesisdriven approach hinges on appropriate
selection and interpretation of
diagnostic tests.
•
The clinician must be able to appraise
the literature regarding diagnostic tests.
Appraising the Literature
Regarding Diagnostic Tests
Condition Present
Condition
Absent
True Positive
False
Positive
False
Negative
True Negative
Test Positive
Test Negative
Appraising the Literature
Regarding Diagnostic Tests
•
Characteristics of Good Studies:
•
Independent Gold Standard
•
Operational Definitions
•
Representative Subjects
Condition Present
Condition
Absent
Test Positive
True
Positive A
False
Positive
B
Test Negative
False
Negative C
True Negative
D
SENSITIVITY
SPECIFICITY
A/(A+C)
D/(B+D)
Sensitivity (True Positive Rate)
•
Proportion of patients with the condition
who have a positive test result
•
Tests with high sensitivity have few false
negatives, therefore a negative result
rules out the condition. (SnNout)
Specificity (True Negative Rate)
•
Proportion of patients without the
condition who have a negative test result
•
Tests with high specificity have few false
positives, therefore a positive result
rules in the condition. (SpPin)
Appraising the Literature
Regarding Diagnostic Tests
•
Likelihood ratios combine the
information contained in sensitivity
and specificity values.
•
Permits comparisons among
competing tests.
Appraising the Literature
Regarding Diagnostic Tests
•
Positive Likelihood Ratio: Expresses
the change in odds favoring the disorder
given a positive test.
(Sensitivity/(1-Specificity))
•
Negative Likelihood Ratio: Expresses
the change in odds favoring the disorder
given a negative test.
((1-Sensitivity)
/Specificity)
Appraising the Literature
Regarding Diagnostic Tests
•
What characterizes a good test?
•
•
Large +LR (>5.0)
•
change the odds favoring the diagnosis
given a + test
•
helpful for ruling in the condition.
Small -LR (<0.30)
•
reduce the odds favoring the diagnosis
given a - test
•
. helpful for ruling out the condition.
Pre-Test
Ratio
Likelihood
Probability
X
Post-Test
Probability
=
50% (1:1)
X
5.0
=
83% (5:1)
50% (1:1)
X
0.30
=
23% (.3:1)
An Example from the Literature
•
Rubenstein et al. The accuracy of the
clinical examination of posterior cruciate
ligament injuries. Am J Sports Med.1995.
•
Performed multiple clinical tests for PCL
laxity in 39 patients (78 knees), 19 with a
torn PCL.
•
gold standard = MRI.
Test
Sens.
Posterior Drawer
90%
Posterior Sag Sign 79%
Qd. Active Drawer 54%
Reverse Pvt Shift 26%
KT-1000
86%
Spec.
99%
100%
97%
95%
94%
+ LR
90.0
~79.0
18.0
5.2
14.3
- LR_
0.10
0.21
0.47
0.78
0.15
An Example from the Literature
•
All tests had higher specificity than
sensitivity, therefore each is better as a
rule in test.
•
The posterior drawer test has a high +LR,
and small -LR, making it an excellent
diagnostic test
Your patient is a 23 year-old male s/p MVA whose knee
hit the dashboard, you think he may have injured his
PCL (25% probability). You perform a diagnostic test to
r/o the PCL injury. The result is negative.
Pre-Test
Likelihood
X
Probability
Ratio=
Probability
Post-Test
Posterior Drawer Test:
25% (.33:1) X
0.10
=
3% (.03:1)
0.78
=
20% (.26:1)
Reverse Pivot Shift Test:
25% (.33:1) X
Another Example
•
69 patients with acute, work-related
LBP
•
Waddell’s signs and symptoms
assessed prior to treatment
•
Gold standard = return to work within
four weeks
Test
Signs (2+)
Sens.
41%
Spec.
79%
+ LR
1.9
- LR
0.75
50%
81%
2.6
0.62
Signs+Symptoms (3+) 64%
62%
1.7
0.59
Symptoms (3+)
Another Example
•
None of the tests demonstrated good
LRs
•
None of the tests would function well as
a screening tool
You have a patient with acute, work-related LBP. You
know approximately 20% of such patients go on to longterm problems. You use Waddell’s tests as a screen to see
if this patient is at risk. The results are negative.
Pre-Test
Likelihood
X
Probability
Ratio=
Probability
Post-Test
Waddell’s Signs (<2):
20% (.25:1) X
0.75
= 16% (.19:1)
Waddell’s Signs+Symptoms (<3):
20% (.25:1) X
0.59
= 13% (.15:1)
Integrating Diagnostic
Information into Practice
If Data
Exists
If Data Does
Not Exist
FIND IT!!
COLLECT
IT!!
Integrating Diagnostic
Information into Practice
•
What You Need To Do:
•
Decide what you are diagnosing
•
List all possible variables
•
Decide on the “gold standard”
•
Measure Everyone !!
An Example
You are in charge of screening residents of
a long-term care facility for those who
need therapy due to increased risk of
falling.
What
are you diagnosing - Risk of falling
What
are the possible predictors?
What
will be the gold standard of fall risk?
Follow-up
everyone
THANK YOU
Review this lecture