EMB Final Exam

In physical therapy, a diagnosis refers to

- A cluster of impairment that define why activity limitations and participation restrictions exist

- OR medical diagnosis

What questions should you ask yourself when it comes to evidence?

- Does the study provide information that is potentially relevant to my patient

- Is the information reasonably applicable to me

- Will the information affect my clinical decision

Diagnostic tests in PT are designed for

-assist in identifying key impairments

-define pathology (diagnostic label)

-indicate when to refer to MD

-can include "special tests"

Strategies for Clinical Diagnosis

-pattern recognition model

-multiple branching method

-exhaustion method

-hypothetico-deductive method

What clinical diagnosis strategy is most prone to error

pattern recognition model

Strategy for Clinical Diagnosis - Pattern Recognition Method

- The instantaneous realization that the patient's presentation conforms to a previously learned picture or pattern

- pathogenetic: characteristics/indicative of particular disease or condition

Strategy for Clinical Diagnosis - Multiple Branching Method

- The progression of the diagnostic process down but one of a large number of potential, preset paths by a method in which the response to each diagnostic inquiry automatically determines the next inquiry to be carried out

- follow yes/no algorithm

Strategy for Clinical Diagnosis - Exhaustion Method

- The painstaking, invariant search for (but paying no immediate attention to) all medical facts about the patient, followed by sifting through the data for the diagnosis

- of all possible options, only thing that's left

-typical medical model approach, time consuming, can lead to false positives

Strategy for Clinical Diagnosis - Hypothetico-deductive Method

- Formulation, from the earliest clues about the patient, of a 'short list' of potential diagnoses or actions, followed by the performance of those clinical (history and physical) maneuvers that will best reduce the length of the list

-efficient, minimize risk of false positivies

Most preferred and defensible Strategy for Clinical Diagnosis

hypothetico-deductive method

Diagnostic Evidence

- Usually in the form of a single study designed to determine the diagnostic utility of a procedure of set of procedures

The STARD statement

STAndard for the Reporting of Diagnostic accuracy studied consisting of 25 points

4 Key Points of STARD Statement

1. Independent, blind comparison with a reference standard

2. What the diagnostic test evaluated in an appropriate spectrum of patients?

3. Reference standard applied regardless of the diagnostic test results

4. Was the diagnostic test validated in a second, independent group of patients?

STARD (independent, blind comparison with a reference (gold) standard)

- Patients in the study have both tests performed (diagnostic and reference standard)

- Clinical performing tests have no knowledge of gold standard results (masked or blinded)

- Similar to concurrent validity - disease is right now, not later

STARD (What the diagnostic test evaluated in an appropriate spectrum of patients?)

- All could feasibly have the target disorder

- Complete range of severity of target disorder

- Disorders that could be confused with target disorder

- Can cause bias for examiner, need appropriate sample portion

STARD (Reference standard applied regardless of the diagnostic test results)

- May be tempting to forgo reference standard when negative diagnostic test results

- But tells only part of the story (eliminates critical information)

STARD (Was the diagnostic test validated in a second, independent group of patients?)

- Tells us about the 'stability' of the diagnostic test

- Least important

True Positive

- Subjects with disease and who test positive

False Positive

- Subjects without disease and who test positive

False Negative

- Subjects with disease who test negative

True Negative

- Subjects without disease who test negative

Formula for sensitivity

= (Patients with the condition who test positive)/(all patients with the condition)

= (True Positive)/ (true positive + false negative)

Formula for specificity

= (Patients without the condition who test negative)/(All patients without the condition)

= (True negative)/(false positive + true negative)

Sensitivity vs Specificity

- Sensitivity: Proportion of patients with the target disorder who have a positive test result (true positive rate)

- Specificity: Proportion of patients without the target disorder who have negative test results (true negative rate)

Formula for Positive Predictive Value

= (Patients with the condition who test positive)/(All patients who test positive)

= (True positive)/(True positive + false positive)

*Proportion of patients with a positive test who have the condition

Formula for Negative Predictive Value

= (Patients without condition who test negative)/(All patients who test negative)

= (True negative)/(True negative + false negative)

*Proportion of patients with negative test who do not have the condition

Two problems associated with the use of predictive values

1. They are prevalence dependent - as prevalence falls, the positive predictive value falls and the negative predictive value increases

2. They can only be calculated for two levels of the test result (+ or -)

Effect of prevalence on predictive value

- do not do tests on patients with very low probability of having a positive test

- as pre-test probability falls then positive predictive value falls, negative predictive value increases

What is often a better alternative to predictive values when you have a low pre-test pobablility

likelihood ration

To increase prevalence

increase PPV

decrease NPV

SpPin

- When specificity is high, a positive test rules in the target disorder

- A highly specific test is rarely positive in the absence of disease (few false positives)

SnNout

- When sensitivity is high, a negative test rules out the target disorder.

- A highly sensitive test is rarely negative in the presence of disease (few false negatives)

Likelihood Ratios

- Thye indicated by how much a given disagnostic test result will increase or decrease the probability of a target condition (quantify shifts in probability for individual patient decisio making)

- Describe likelihood of disorder in someone with as opposed to someone without the disorder

- Minimally affected by prevalence

Likelihood Ratio Scale:

5-10 or .1-.2

Moderate shifts in pretest - posttest probability

Likelihood Ratio Scale:

>10 or

Large and often conclusive changes from pre-test to post-test probability

Likelihood Ratio Scale:

2-5 or .2-.5

Small but sometimes important changes in probability

Likelihood Ratio Scale:

1-2 or .5-1

Small and rarely important changes in probability

Formula for Positive Likelihood ratio

= (Sensitivity)/(1-specificity)

Formula for Negative Likelihood Ratio

= (1-sensitivity)/(Specificity)

Strengths Associated with Likelihood Ratios

- They are not generally prevalence dependent (stable)

- They can be calculated for many levels of the test result

- They can be used to estimate the probability that the target disorder is present in your patient (they quantify shifts in probability for individual patient decision making)

Nomograph

graphic that guides interpretation of likelihood ratios

Nomograph is most meaningful when

pretest probability is higher (ally to individual patient)

Pre-test probability low

regardless of how good the likelihood ratio is, the post-test probability really isn't going to increase that much (and the individual still is unlikely to have the condition)

Pre-test probability too high

the post-test probability doesn't really increase that much; therefore the test isn't really informative because you are already pretty certain that the individual has the condition

Likelihood ratio = 1

no effect of likelihood of disease

Likelihood ratio > or < 1

infromative

Main study designs

-intervention

-observational

-systematic reviews and meta-analysis

Types of intervention study

-clinical trials

-experimental studies

Types of observational studies

-cohort

-case control

-cross sectional

-descriptive case reports

-case series