L2: Clinical Validation Flashcards

1
Q

Define a biomarker and matrices they can be measured in.

A

Biological property whose in vitro measurement or identification is useful for the diagnosis, prognosis, treatment and follow up of humane disease. Biomarkers can be measured in: sera/plasma, urine, CSF, saliva, stool, synovial fluid etc.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What are a few types of biomarkers?

A

proteins, trace elements, hormones, drugs, vitamins, lipids

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are the ideal characteristics of a biomarker?

A
  1. Non-invasive, easily measured, inexpensive and produce rapid results
  2. Should be obtained from readily available sources
  3. Should have high sensitivity (allowing early detection and no overlap in disease and healthy patients)
  4. Should have high specificity (greatly up or down regulated in disease samples and unaffected by comorbid conditions)
  5. Should vary rapidly in response to treatment
  6. Should aid in risk stratification and possess prognostic value in terms of outcomes
  7. Should provide insight into underlying mechanism
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

List a few measurement methods for biomarkers.

A

CE, ELISA, MS, POCT, Spectrophotometry

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Describe the role of biomarkers in screening and their clinical utility

A

Role: identify disease in apparently healthy individuals, requires high specificity, inexpensive to assay

Utility:

  • used to identify apparently healthy people who may be at increased risk of a disease or condition (e.g. FIT)
  • usually have high sensitivity to ensure true positive results are detected
  • usually have lower specificity and produces more false positive results
  • screening results need to be followed up with diagnostic or confirmatory test
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Describe the role of biomarkers in diagnosis and their clinical utility

A

Role: high sensitivity, high tissue specificity

Utility:

  • diagnose particular disease or condition based on clinical suspicion
  • typically required to have excellent sensitivity and specificity
  • if confirmatory test follows, specificity can be somewhat lower
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Describe the role of biomarkers in confirmation and their clinical utility

A
  • verification or confirmation of test results allows clinicians to establish a diagnosis (e.g. colonoscopy for FIT)
  • design to be specific and have high PPV
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Describe the role of biomarkers in prognosis and their clinical utility

A

Role: monitor patients with chronic disease, low intra-individual variability, reflect response to treatment

Utility:

  • Help assess a patient’s future risk (HIV to AIDs)
  • difference from a diagnostic test is the need for longer follow-up and need to include specific time frame
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Describe the role of biomarkers in monitoring and their clinical utility

A
  • Periodic measurement to guide management of chronic or recurrent condition (e.g. HbA1c every three months)
  • Can be divided into five phases: 1) pretreatment, establish response, 3) monitoring treatment, 4) adjustment to establish control, 5) cessation of treatment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Define Evidence Based Laboratory Medicine (EBLM)

A
  • defined as using the best research evidence and laboratory medicine expertise to aid in disease diagnosis and stratifying risk
  • information from expert advice and textbooks often lag behind current medical and scientific literature
  • EBLM allows to acquire, appraise, and analyze the most recent data in real time
  • leads to improved care of patients, improved health outcomes, and effective use of resources
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Describe the elements of EBLM.

A

6-step method for conducting reviews. Provides framework to conduct review and understand how to assess effectiveness of QI practices.

1) Ask: frame the question
2) Acquire: identify sources and relevant studies
3) Appraise: create an evidence base by applying screening criteria related to topics, questions, and practices
4) Analyze: standardize, summarize, and rate strength of body of evidence
5) Apply: Disseminate findings for review
6) Audit/Assess: conduct audit to assess application and determine need to create new focused questions

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Framing the question: define PICOTS.

A

P (patients): what are the patients symptoms or demographics?
I (intervention): what test is being considered?
C (comparator): what is the reference or gold standard?
O (outcome): what is the endpoint of interest?
T (timing): what is the target timeframe for studies?
S (setting): what are the clinical setting for studies to have been conducted?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Define diagnostic sensitivity.

A

Proportion of those with the target disease that receive a positive result.

TP/TP+FNx100%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Define diagnostic specificity.

A

Proportion of individuals without the target disease that receive a negative result.

TN/TN+FPx100

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Scenario: In a study with a cohort of 1500 patients, 750 had the target disease. The study found 630 patients had a true positive result and 165 had a false positive result. What was the diagnostic specificity and sensitivity of the assay?

A

SPECIFICITY: 78%
SENSITIVITY: 84%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Scenario: In a study with a cohort of 1500 patients, 750 had the target disease. The study found that 6300 patients had a true positive result, 165 had a false positive result. What is the PPV and NPV of the test?

A

PPV: 79%
NPV: 83%

17
Q

Define positive predictive value (PPV).

A

Sensitivity and specificity make the underlying assumption that disease is known.
PPV: likelihood that a positive result correctly identifies disease.

PPV: TP/TP+FPx100

PPV increases with disease prevalence. More disease means more true positives.

18
Q

Define negative predictive value (NPV).

A

Sensitivity and specificity make an underlying assumption that disease in known.

NPV: likelihood that a negative result correctly identifies absence of disease.

TN/TN+FNx1000

Lower prevalence requires high diagnostic sensitivity to reduce the number of false negatives.

19
Q

Scenario: The target disease has a prevalence of 10% in population being tested. A diagnostic test was applied to a random sample of 200 individuals, yielding 15 true positives and 15 false positives. Calculate: 1) pre-test odds of disease being present, 2) likelihood ratio of the test, 3) post-test odds of disease for a patient with positive result.

A

Pretest odds: 0.11
Likelihood ratio: 9
Posttest odds: 0.99

20
Q

Define prevalence.

A

proportion of patients with the target disease in the population.

in clinical situation, prevalence is equal to pre-test odds.

TP+FN/TP+FN+FP+TN
disease/total

21
Q

Define likelihood ratio positive (LR+).

A

ratio between the probability of finding a positive test in the presence of disease and the probability of obtaining a positive result in absence of disease.

LR+ = sensitivity/(1-specificity)

22
Q

Define likelihood ratio negative (LR-).

A

ratio between the probability of finding a negative test without disease and the probability of finding a negative result in the presence of disease

LR-= 1-sensitivity/specificity

23
Q

What is a Fagan Normogram? Define post-test odds.

A

Tool to determine post-test probabilities

Visual nomogram where a line can be drawn from pre-test odds through the likelihood ratio to estimate the post-test odds.

Post-test odds: pre-test odds x LR+

24
Q

Define test efficiency.

A

Test efficiency is defined as the percentage of the times that the test gives the correct answer.

TE= TP+TN/TP+TN+FP+FN x 100

25
Q

Define Receiver Operating Characteristics (ROC) curves.

A
  • Plot true positive rate (sensitivity) on y-axis and false positive rate (1-specificity) on x-axis
  • Show pattern of sensitivities and specificities when the performance of the test is evaluated at several diagnostic thresholds
  • Overall diagnostic performance of a test is evaluated by examining the AUC (every point represents an independent cutoff)
26
Q

Describe the process of creating ROC curves.

A
  1. Extract ID, concentration, and disease status
  2. Arrange data from largest to smallest
  3. Calculate Y(sum): cumulation of individuals with disease
  4. Calculate N(sum): cumulation of individuals without disease
  5. Calculate sensitivity and 1-specificity at that concentration
  6. Graph sensitivity by false positive rate for each point
27
Q

Describe AUC in the context of ROC and how to calculate it.

A

AUC: used to determine how good the test is in a given clinical situation.

Use the trapezoid method to calculate it. Calculate the sum of all the areas between x-axis and a line connecting two adjacent points.

28
Q

Describe how to interpret a ROC curve and how to calculate Youden’s index.

A

Perfect test: curve passes through upper left corner (AUC=1, 100% sensitivity and specificity)

Not useful test: ROC is a diagonal line (AUC=0.5)

Categorize ROC curves by AUG:

  1. 9-1: VERY GOOD
  2. 8-0.9: GOOD
  3. 7-0.8: FAIR
  4. 6-0.7: POOR
  5. 5-0.6: FAIL

Youden’s index is the diagnostic cutoff that will yield the highest specificity and sensitivity = sensitivity + specificity - 100