Week 4 - Diagnotic Tests Flashcards

(93 cards)

1
Q

What will a diagnostic test test the presence of?

A

A disease, condition or substance.

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

Participants in a diagnostic test study will present with symptoms already.

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

What information will a diagnostic test provide?

A

It collects information that will clarify a patient’s clinical condition, which helps to determine prognosis.

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

What types of information can be used to help clarify a patient’s clinical continence to help determine prognosis?

A

Patient characteristics.
Signs and symptoms.
Physical examinations.
Tests that have used laboratory or other technical facilities.

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

What does the clinical do in a diagnostic test?

A

They do the test.
They receive the results.
They find out whether the patient has the condition or not.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

No diagnostic test is perfect.
Might say a disease isn’t present, when it actually is.

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What does a False Positive result mean?

A

Test says you have it.
But you actually don’t.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What’s a False Negative result mean?

A

Test says you don’t have it.
But you actually do.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are practitioners in diagnostic imaging interested in?

A

Providing high quality images which can permit a medical diagnosis.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What examples of evolving technology are there?

A

Lymphangiography - this was replaced by US.

Myelography - replaced by CT.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What’s the difference between a diagnostic and a screening test?

A

Diagnostic -
Generally used when patient has some symptoms that are suggestive of the disease in question.

Screening -
Generally used on people without symptoms, who are drawn from a population with a much lower prevalence of the disease.

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

What does Screening involve?

A

Investigating apparently healthy individuals (or asymptomatic), with an aim to identify a disease early, thus enabling earlier intervention and management.

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

What may Screening involve?

A

Mass screening of the whole population.
Selected groups will be shown to have an increased risk of prevalence of a certain condition or disease.

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

What are the 2 types of screening programmes?

A

Systematic.

Or…

Opportunistic.

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

What happens in a Systematic Screening Programme?

A

People are invited.
E.g. mammography study uses female participants aged 50 to 70 years.

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

What happens in an Opportunistic Screening Programme?

A

Screening gets offered to someone who may present for a different reason.
E.g. chlamydia screening in a student presenting with depression.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What are some Advantages of Screening?

A

It can detect unrecognised disease early, meaning their prognosis of the disease can be improved.
E.g. increased survival in women with breast cancer who were diagnosed and treated early.

It can detect individuals that are at high risk of developing a certain disease where the individual or clinician can take measures to delay (or prevent) the development of the disease, done by reducing the risk.
E..g screening for high BP then offering lifestyle advice and/or drug intervention.

It can help identify people with an infectious disease, where an intervention can treat the infection, and prevent transmission of the disease to others.
E.g. chlamydia screening in sexually actively people under the age of 25.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What are some Disadvantages of Screening?

A

It might give a false sense of security (e.g. if there are some false negative results) and this delays the final diagnosis.

People tested negatively may feel they’ve avoided the disease, so still continue with their risky behaviours.
E.g. individual who eats more than recommended daily allowance of saturated fat may still eat the same amount as when they got tested and were fine, yet it will quickly get worse over time.
This may undermine primary prevention programmes, e.g. to prevent coronary artery disease by promoting healthy eating.

For true positive cases that get early treatment even though the disease is only just starting to develop could create possible side effects.

It can involve lots of medical resources and lots of money, which could be used elsewhere, especially as most people that are screened don’t need treatment.

A false positive result can cause stress and anxiety - and this may be from unnecessary investigations, especially if the procedure is invasive.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

Look at onenote for table comparison of diagnostic tests vs screening test.

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Diagnostic test results may be ordinary (on a scale) or normal/abnormal.

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is meant by diagnostic test results being Normal/Abnormal -

A

E.g. the presence or absence of a fracture on a plain radiograph.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

What is meant by a Diagnostic Test being Ordinal (on a scale)?

A

May be classed as something, ranging from a normal mammogram, to one which shows a benign disease, to one that has a suspicion of cancer or the presence of cancer.

It’s important to establish a normal, sensible range.

Boundaries have to be set in order to establish if a test is abnormal or normal.
E.g. blood pressure of 142/92 is abnormal as it’s over 140/90, but is this BP risk different from 138/88?

Treatment needs to be go/no-go decision.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

Example of an ordinal diagnostic test result -

Prostate-Specific Antigen -

A

Healthy man - has 0.5 ng/mol in their blood.
Someone with advanced prostate cancer would show will over 20 ng/mol.
A PSA of 7.4 ng/mol may be found in someone that is normal though, or someone with very early cancer.

So a PSA can’t actually prove if a man has prostate cancer.

However it may increase or decrease the odds of a diagnosis like that.

This is expressed as a likelihood ratio.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

Evaluating Health Technologies -

How would you evaluate Diagnostic Technology?

A

See if the new technology will result in improvements in the prognosis or physical health of the patient.

An imaging investigation should be performed.
And then a clinician will interpret these images.
The clinician will use this info, and other info (clinical findings, other tests), to refine the diagnosis.
The clinician should use this info to plan therapy.
And this may influence patient outcome.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Evaluating Health Technologies - A hierarchy may be used to evaluate a health technology.
Start with Technical Performance - Is the image quality sufficient to make diagnostic or therapeutic decisions? Then onto Diagnostic Performance - This is concerned with whether imaging can correctly or incorrectly assess the presence or absence of a disease. Confirmed with a ‘gold standard’ test. It depends on observer variation in interpretation of the medical images. E.g. technical performance - does MRI reliably result in good quality images which are anatomically representative? diagnostic performance - does MRI reliably the images produced allow accurate diagnoses to be made?
26
Evaluating Health Technologies - What 3 levels need to be assessed when using observation research designs?
1. Diagnostic Impact - Can the imaging technology replace other investigations? 2. Therapeutic Impact - Is concerned with whether the imaging can contribute to planning and delivery of therapy. A study can compare pre-imaging therapy plans with post-imaging therapy plans. 3. Patient Outcome - Does imaging improve patient outcomes?
27
Evaluating Health Technologies - What is the best method for evaluating the effectiveness of technologies, like MRI? Why?
An RCT. Because it can randomly assign patients to receive one diagnostic test result or an alternative.
28
Evaluating Health Technologies - Societal -
This moves beyond the clinical effects of a technology, in order to determine whether the cost of the technology is acceptable to society. Resources are limited, and policy makers have to make decisions about how they are allocated. Assess the extent to which MRI is an efficient use of resources to provide benefits to society. E.g. this could be a cost-effectiveness study that compares the cost of arthroscopy compared to an MRI of the knee. This is the area of health economics.
29
Overall Steps of Evaluating a Health Technology -
1. Technical Performance. 2. Diagnostic performance. 3. Diagnostic Impact. 4. Therapeutic Impact. 5. Patient Outcome. 6. Societal.
30
To perform a Validation Study, diagnostic tests are compared against a gold-standard test.
31
What is a gold-standard test?
One that is assumed to always give the right result, and give the patient’s true disease status.
32
The gold-standard / reference standard, is a term for what?
The most definitive diagnostic procedure, which distinguishes people with a disease from people without a disease.
33
What can be bad about a gold-standard investigation?
They are time-consuming, invasive, risky, painful or expensive.
34
This means researchers try to find simpler tests compared to the gold-standard. For example -
How valid is an exercise stress test for diagnostic coronary artery disease, compared to the gold-standard test for cardiac testing with angiography?
35
No diagnostic test is always perfect -
There can be false positives and false negatives. However you can still get true negatives and true positives from them.
36
What else is a False Positive known as?
A Type 1 error. People without the disease are incorrectly identified as having the disease.
37
What else is a False negative known as?
Type 2 error. People without the disease are incorrectly identified as being disease-free.
38
Look at onenote for a mammography example of interpreting test results.
39
What are the 4 Measures used for Diagnostic Performance?
1. Validity - 2. Prevalence. 3. Sensitivity. 4. Specificity.
40
Measures used for Diagnostic Performance - What is meant by 1. Validity?
The ability of a test to indicate which individuals have a condition and which do not. It’s the comparison of the measurements of a diagnostic test a reference standard results.
41
Measures used for Diagnostic Performance - What is meant by 2. Prevalence?
This is the proportion of all patients who have the condition.
42
Measures used for Diagnostic Performance - What is meant by 3. Sensitivity?
This measures the ability of the test to correctly identify people with the disease. It’s also known as ‘positive in disease’. Sensitivity values can range from - 0 = no ability to correctly identify patients free from the condition. 1 = ALL patients with the condition are identified.
43
Measures used for Diagnostic Performance - What is meant by 4. Specificity?
This measures the ability of a test to find those who don’t have the disease. Specificity values can range from - 0 = no ability to correctly identify patients free from the condition. 1 = if it identifies all of them.
44
What are the number of patients with the disease calculated by?
TP + FN.
45
What are the number of patients without the disease calculated by?
TN + FP.
46
What calculation identifies Sensitivity?
TP / TP + FN.
47
What is Sensitivity?
The proportion of patients with the disease who have a positive test result. I.e. how good is my diagnostic test in detecting patients with the disease.
48
What calculation identifies Specificity?
TN / TN + FP.
49
What is Specificity?
The proportion of patients without the disease who have negative test results. I.e. how good is my diagnostic test result in detecting patients without the disease?
50
What’s the mnemonic for Sensitivity?
SnNout. A highly SeNsitive test with a Negative result tends to rule OUT the diagnosis of the disease.
51
What’s the mnemonic for Specificty?
SpPin. A highly Specific test with a Positive result tends to rule IN the diagnosis of the disease.
52
What’s Trade-Off Sensitivity Specificity?
An inverse relationship. Where increasing a diagnostic test’s sensitivity (identifying true positives) typically decreases its specificity (identify true negatives), and vice versa. Lowering the decision threshold improves sensitivity but can create more false positives. But raising the decision threshold improves specificity but creates more false negatives. Compares test results to the frequency of the disease. Look at onenote for example of it in a graph.
53
What is meant by the Threshold of a test?
The cut-off that’s used in declaring that the test is positive. It can be adjusted, which will in turn alter the sensitivity and specificity of the test. Can look at onenote for example of it in a graph.
54
What happens if the threshold is shifted to the right of the graph?
Specificity increases (true negative rate). But sensitivity (true positive rate) of the diagnostic test is decreased.
55
What is meant by PPV?
Positive Predictive Value. It is the proportion of patients with a positive test result who do have the disease. I.e. looks at how well a positive test result predict the presence of a disease. PPV = TP / TP + FP (total positive).
56
What does PPV show?
The likelihood that the patient actually has the condition if the result is positive.
57
What does a value closer to 1 mean?
It is more likely that the patient really has the disease given the positive test result.
58
What is meant by NPV?
Negative Predictive Value. It is the proportion of patients with a negative test result who don’t have the disease. I.e. how well does a negative test result actually predict the absence of the disease. NPV = TN / FN + TN (total negative).
59
NPV gives the likelihood that the patient is in fact healthy if the result is negative.
60
What does a value of NPV closer to 1 mean?
It’s more likely that the patient really doesn’t have the disease, proven by the negative test result.
61
What is it called if the Disease is present in someone and the test they take shows positive?
True Positive (TP).
62
What is it called if the Disease is present in someone and the test they take shows negative?
False Negative (FN).
63
What is it called if the Disease is absent in someone and the test they take shows positive?
False positive (FP).
64
What is it called if the Disease is absent in someone and the test they take shows negative?
True Negative (TN).
65
Likelihood Ratio - What is meant by pre-test probability? What does it help with?
It is the background probability of a patient having a condition, before nay testing takes place. It helps us to move our suspicion one way or the other, to help us know what to look for, which gives us a post-test probability.
66
Likelihood Ratio - What is it?
It gives a value, that tells us how much probability changes once we know the test result.
67
Likelihood Ratio - What is meant by LR+?
It’s the multiplier for how much more likely a patient is likely to have the condition if the test result is positive. Probability that a patient with the disease has a positive result / probability that a patient without the disease has a positive result.
68
Likelihood Ratio - What is LR-?
How much the risk for having the condition has decreased if the test is negative. Probability that a patient with the disease has a negative result / probability that a patient without the disease has a negative result.
69
Look at onenote for summary table of key definitions like -
Sensitivity. Specificity. PPV. NPV. Accuracy. Likelihood Ratio.
70
Look at onenote for an example of a Prostate Cancer Screening Test.
71
Interpreting Diagnostic Performance - When is a high sensitivity preferred?
If the disease is life-threatening if it’s kept untreated. If there’s an improvement in survival rate if any treatment is initiated early on. If overdiagnosis is okay (is a screening test), because all those who screen positive will have more tests. For example - The absence of retinal vein pulsation for diagnosing raised intracranial pressure. Or Screenign programmes such as for breast cancer or HIV.
72
Interpreting Diagnostic Performance - When is a high specificity preferred?
When the disease isn’t life-threatening if left untreated. If treatment costs are high. If the pre-test probability of the condition is low. If subsequent tests/treatments are invasive (e.g. prostate biopsy) or have severe side effects (e.g. chemotherapy). For example - Screening tests for Down’s syndrome during pregnancy are highly specific, as the consequences of a wrong diagnosis are very serious.
73
Look at onenote for Exam Question Example - Interpreting Diagnostic Performance.
74
What is Interobserver Reliability?
The measure of the consistency, agreement, or correlation between 2+ observers/raters assessing the same subject, behaviour or data. It ensures results are objective, reliable and not influenced by a single observer’s biased opinions/interpretations.
75
Studies of diagnostic capability should consider if different observers that a re responsible for interpreting medical images are doing it consistently.
76
If observer performs the same diagnostic test on 2 different occasions on a participant whose characteristic haven’t changed, they will likely get different results in a proportion of cases.
77
Some tests have a high reproducibility (e.g. 99%). But others may lower, like 50%, where half the time they will get different results.
78
In diagnostic performance studies in imaging, there is variability in observer interpretation of images. What may this be down to?
Experience. Professional. Training.
79
What does the assessment of reliability involve?
Different observers interpreting the same samples of images, known as an inter-observer test. Or involves the same observers interpreting the same images on separate occasions, known as an intra-observer test.
80
What is the Kappa Statistic?
A measure of performance, used to analyse whether observers performance agree.
81
How can the Kappa Statistic be calculated?
Can be calculated when the classification of an image is - 1. Binary - presence or absence of a fracture on a plain radiograph. 2. Ordinal - normal mammogram, one which shows being disease, the suspicion of cancer or the presence of cancer.
82
Kappa has a max. value of what?
1. This is when there is a perfect agreement between observers.
83
What does a Kappa value of 0 mean?
Any agreement between observers has occurred for no better reason than chance.
84
What things can Kappa be calculated for?
A single observer observing the same image on multiple occasions. 2 different observers on the same occasion. Comparisons of multiple observers.
85
What are some key questions that would be relevant to a diagnostic test paper? (Look at onenote for more detail within them all).
1. Is the test relevant to my practice? 2. Has the test been compared with a true gold-standard? 3. Was there an appropriate range of participants selected? 4. Did everyone who got the new diagnostic test also get the gold-standard, and vice-versa? 5. Are all the results accounted for? 6. Did clinicians who interpreted the diagnostic test have knowledge of the result of the gold-standard test or vice versa? 7. Was the test shown to be reproducible both within and between observers. 8. Was there diagnostic performance of the test acceptable? 9. Were confidence intervals given for sensitivity, specificity and other diagnostic measures?
86
What does screening involve overall?
Investigating supposedly healthy individuals. With an aim to identify disease early. And this will enable earlier interventions and management.
87
New health technologies can. Be evaluated against diagnostic impact, therapeutic impact, and patient outcomes.
88
To perform a validation study, what 2 things are compared against each other?
Diagnostic tests. Gold-standard tests.
89
What is a gold-standard test (or a reference standard)?
A term for the most definitive diagnostic procedure. It distinguishes people’s with disease from people without disease.
90
Identifying new diagnostic tests may be necessary if the definitive gold-standard test is costly, risky, invasive, painful or time-consuming.
91
What 4 key measures are sued to evaluate performance of a new test?
Sensitivity. Specificity. PPV - positive predictive value. NPV - negative predictive value.
92
Diagnostic performance must be evaluated in the context of other tests.
93
Whata re the key features we need to identity in a diagnostic paper?
Range of participants. Identifying a gold-standard. Test reproducibility. Diagnostic performance. Verification, expectation, selection bias.