Brainscape's Knowledge GenomeTM


Top Flashcards (Certified)
All Flashcards

HSS3110 - Epidemiology > Week 3 > Flashcards

Week 3 Flashcards

(78 cards)

Start Studying
1
Q

What is the difference between normal and abnormal?

A

Normal is the standard of what something should be

Abnormal is when the something is off and no longer within the normal range

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

How do we determine if it’s Health or Disease?

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

How do we determine if it’s Abnormality or Normality?

A

Biological

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

What is abnormality?

A

Doctors often define abnormality as lying outside the normal range which may not be due to disease

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

T of F: The acceptance of what is a disease changes over time with some diseases disappearing and others appearing

A

True
Ex: ADHD is a new disease

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

T of F: Many factors could influence what should be treated as a disease?

A

True
Ex: sociocultural factors. Within social classes some disease are accepted as normal but some as abnormal

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

Within a clearly distinct graph of 2 curves, one represents the diseased and the other the not diseased, what is easy to find?

A

It’s easy to find the cut point.
Sensitivity and specificity are high.

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

Within a bimodal graph of 1 curves with 2 bumps, one represents the diseased and the other the not diseased, what is not easy to find?

A

The cut point
Sensitivity and specificity are low

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

Within a unimodal graph of 1 curves, what is hard to find?

A

hard to find the cut point.
Sensitivity and specificity are largely compromised

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

What does the cut off point help us find?

A

We want to indentify most of the non-disease population

For disease population, they have higher values which helps us define them

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

What is the difference between ideal circumstance and clinical circumstance?

A

Ideal is that we have all of our non disease on one side and all of our diseased on the other side of our cut off

Clinical is that we have non disease and disease on each side but most non disease are on the normal side and most disease are on the abnormal side

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

What are different types of clinical measures?

A

• Nominal data
Categorical events
e.g. ABO blood types

• Ordinal data
Ranking data
e.g. severity of disease

• Interval data
Continuous and/or discrete data
i.e., glucose level: 190.22 mg/100 mL, 190 mg/
100mL.

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

A 50-year-old man came to you and complained about his health problems. A high glucose level was found. Could you give a diagnosis of diabetes?

Next week, you asked this patient to have another blood test for his glucose level. Would you give a diagnosis?

A
  1. No you need a second look
  2. Yes, that should be enough

BUT
This patient does not trust you, as you are a junior doctor. He goes to see another doctor. This doctor gives the same diagnosis
Some patients don’t always trust the young doctors

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

What are the 2 indicators of quality of performance?

A

Validity and reliability

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

What is validity?

A

Strength of our conclusions, inferences or propositions.

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

What is reliability?

A

Consistency of our measurement, or the degree to which an instrument measures the same way each time under the same conditions with the same subjects.

Ex: when patients ask different doctors for opinions

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

A random error in the distribution affects the validity or reliability?

A

Reliability

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

A systematic error in the distribution affects the validity or reliability?

A

Validity

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

What are types of validity?

A
  • Content validity
  • Construct validity
  • Criterion validity
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is content validity?

A

• (also known as) Logical validity
a measure represents all facets of a given construct.

• Example: You are going to test our students’ mathematical skills, and you develop a questionnaire to test multiplication, then give a conclusion based on this.

Not good cuz doesn’t test everything. Multiplication doesn’t equal to skills in general

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

What is an Item-Construct-Scale?

A

Bearkdown the overall into small categories

Ex: the depression scale is broken down into different categories to figure out your level of depression

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

What is Construct validity?

A

• Agreement between a theoretical concept and a specific measuring device or procedure.
• I.e., a researcher inventing a new IQ test might spend a great deal of time attempting to “define” intelligence in order to reach an acceptable level of construct validity.
• Psychological measure

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

What is Criterion Validity?

A

• Instrumental validity
• Demonstrate the accuracy of a measure or procedure by comparing it with another measure or procedure which has been demonstrated to be valid.
i.e., diagnostic criteria (ICD-9 vs. DSM-IV) (is it the same results for both)

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

What is reliability?

A

The extent to which repeated measurements of a stable phenomenon by different people and instruments at different times and places. (If we use them multiple times, does it give us the same result)
- i.e., a clock or tape used in the Olympics Games
- i.e., lab test results. Blood and tissue samples will be kept for another potential test.

It’s important for accuracy

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
How do we test reliability?
• Test/Retest (do test on one pop and then 2nd test on same pop to see if you get same results) Expect the same results from test 1 and test 2. • Internal Consistency Estimate reliability by grouping questions in a questionnaire that measure the identical concept. - i.e., Design two questionnaires to survey the same topic with the same subjects. Estimate the correlations between the two questionnaires.
26
What are the 3 criteria’s of reliability?
• Range • Responsiveness • Interpretability
27
What is range?
Limitations of information which may influence target interest. – limited items to check about all facets of depressive symptoms Ex: temp has range but for extreme situations it won’t capture those values
28
What is responsiveness?
Mark the changes qualitatively with clinical or practical meaning.
29
What is Interpretability?
Interpret findings or results into practice. Make the findings easily used. Understandable and potential translation Ex: trying to measure someone’s mood - once we have their score, we can quantify and figure out their mood
30
What are Factors influence validity & reliability?
• Variations • Distributions
31
What are the types of variations?
Measurement Variation - Instrument: The means of making the measurement - Observer: The person making the measurement Biologic Variation - Within individuals: Changes a person with time and situation - Between individuals: Biological differences from person to person
32
What are types of distributions?
- Actual distributions - Normal distributions
33
What is Regression to the mean?
• Affects the internal validity of the experimental design. • Distort the conclusion that the significance or effect, is due to the treatment when in fact it is due to chance and by this phenomenon known as regression to the mean. Regression to the mean refers to the tendency of results that are extreme by chance on first measurement—i.e. extremely higher or lower than average—to move closer to the average when measured a second time
34
T or F: the mean of the second test will look like the mean of the first test?
True
35
T of F: An individual who scored extremely low on the first test is “likely to score higher on a second test simply because the error component is likely to be less extreme and closer to the average.
True
36
T or F: Less reliable the test is, the more likely regression to the mean will occur.
True
37
It is important to note that regression is always to the ________ mean of a group.
Population
38
Regression to the mean is a statistical phenomenon that is a________
fact of life in statistics
39
Regression to the mean essentially occurs on what?
the posttest where the measures (for example test scores) on the average regress toward the mean on average. The net effect of regression toward the mean is that the lower scores (or measurements) on the pretest tend to be higher on the posttest, and the higher scores (or measures) on the pretest tend to be lower on the posttest.
40
What are the criteria for selecting a cut-point
• Increased risk • Being sick Cost-effective, sound, reliable, and meaningful • Being treatable - If not treatable, earlier notification only causes more burden; - This abnormality is not related to what it is happening; - Once the damage occurs, it is irreversible.
41
What is a Diagnostic test?
A diagnostic test is any kind of medical test performed to aid in the diagnosis or detection of disease
42
For what do we use a diagnostic test?
- to detect or exclude diseases; (At clinical stage = for diagnosis, at pre-clinical stage for screening) - to measure the progress or recovery from disease; - to make decisions regarding treatments; - to reassure patients about diseases.
43
What are the 3 pairs of measures for accuracy?
• Sensitivity • Specificity • Positive predictive value • Negative predictive value • Positive likelihood ratio • Negative likelihood ratio
44
How do you calculate sensitivity?
Se=a/(a+c)
45
How do you calculate specificity?
Sp=d/(b+d)
46
How do you calculate Positive predictive value?
+PV=a/(a+b)
47
How do you calculate Negative predictive value
-PV=d/(c+d)
48
What is the difference between sensitivity and specificity?
• Sensitivity – the proportion of those with the condition who have a positive result (better sensitivity lower false negative rate) – high sensitivity to rule out the target condition • Specificity – the proportion of those without the condition who have a negative result (better specificity lower false positive rate) – high specificity to rule in the target condition
49
For sensitivity, how can we be nearly certain that they don’t have disease?
an important penalty for missing a disease A dangerous but treatable disease Ex: TB: ‘Rule out’: if the test is highly sensitive and the test result is negative you can be nearly certain that they don’t have disease
50
For specificity, how can we be certain that they actually have the disease?
an important penalty for misclassifying a disease – i.e., biopsy for breast cancer ‘Rule in’: if the test result for a highly specific test is positive you can be nearly certain that they actually have the disease.
51
What are Trade-offs between sensitivity and specificity?
• A graphical plot of the sensitivity, or true positive rate vs. false positive rate (1 − specificity or 1 − true negative rate), for a binary classifier system as its discrimination threshold is varied. • Represent equivalently by plotting the fraction of true positives out of the positives (TPR = true positive rate) vs. the fraction of false positives out of the negatives (FPR = false positive rate).
52
When the cut-point shifts to right, _____ decreases, ______ increases.
Sensitivity and specificity When you move the cut off you sacrifice sensitivity but specificity increases
53
What are Factors that influence sensitivity and specificity?
• Prevalence of target disease (pre-test probability - for any given task you need a pre-knowledge before implementing your test) • Population settings (community or referred patients) (for any general population, the chance of getting the given disease is lower than the studies in the hospital setting. In hospitals you get a higher %) • Characteristics of the disease itself, i.e., severity (cardio -> sudden deaths so not as many of people have it vs a mild disease) • Bias • Chance
54
What is the difference between Positive predictive value and negative predictive value?
• Positive predictive value – the probability of a patient with positive result has the target condition Ex: When the positive result of diagnostic test, what is the probability of having disease? (a/(a+b)) • Negative predictive value – the probability of a patient with negative result does not have the target condition Ex: When the negative result of diagnostic test, what is the probability of not having disease? (d/+c+d)) • Clinically relevant
55
Predictive values are strongly dependent on what?
the population pre-test probability/prevalence – For example, the prevalence of depression is likely to be higher in hospital than the general practice settings. The predictive values estimated from a primary diagnostic test should not be assumed to apply in other different settings.
56
What are the 2 levels of Pre-test probability?
• Population level – prevalence of the target condition • pre-existing literature, i.e., systematic reviews • diagnostic test 2 by 2 table • Individual level – chance of having the target condition • individual clinical history, clinical symptoms, physical examination, knowledge and experience, etc.
57
What is the correlation between Positive Predictive value and specificity?
High Positive predictive value -> High Ruling in the target condition -> Low False positive rate -> High chance of high Specificity
58
What is the correlation between Negative Predictive value and sensitivity?
High Negative predictive value -> High Ruling out the target condition -> Low False negative rate -> Chance of high Sensitivity
59
What are the Determinants of predictive values?
• Posttest probability • Sensitivity, specificity, & prevalence – More sensitive the test is, the better will be its negative predictive value (d/c+d) (c is small) – More specific the test is, the better will be its positive predictive value (a/a+b) (b is small) • If prevalence decreases, (a, c) are small, so positive predictive value is smaller; while if prevalence increases, negative predictive value is smaller.
60
What is the difference between Positive likelihood ratio and negative likelihood ratio?
Positive likelihood ratio – how much more likely a person with the target condition is to receive a positive result than a person without a condition – greater than 1 up to infinity (bigger than 10 = good positive ratio) • Negative likelihood ratio – how much more likely a person with the target condition is to receive a negative result than a person without a condition – between 0 to 1 (smaller than 0.1 = good negative ratio)
61
Which pairs of measurements are the most Clinically useful?
Positive and negative likelihood ratio
62
What does it mean when the Positive or negative likelihood ratio =1?
correctly identify 50% of patients with disease, but wrongly misclassify 50% of those without disease The further away the value is from 1 (either direction) the more useful the test is.
63
How to go from Pre-test -> post-test probability?
• Pre-test probability=probability • Pre-test odds=probability/(1-probability) • Post-test odds=pre-test odds X likelihood ratio • Post-test odds=odds • Post-test probability=Posttest odds/(1+ Posttest odds)
64
How to Increase the prevalence of disease before screening?
• Referral Process - rule out the low risk groups • Selected demographic groups-find out high risk groups • Specifics of the clinical situation-take symptoms into account
65
What are 2 types of Bias in diagnostic test?
• Patient selection – diagnostic cohort design – diagnostic case-control design • spectrum bias • Patient flow – if the number of patients who were enrolled in the study differs from the number of patients included in 2 by 2 table then there is a potential for bias • verification bias (work-up bias or referral bias) – perform gold standard testing in a random sample of study participants
66
Why is screening good?
Because if you don’t screen you won’t identify the disease The faster you detect and treat the smaller the symptoms and the longer you can live
67
What happens to your sensitivity if only a few people are measured in your study?
It will be smaller/narrow
68
What is Parallel testing and what does it do?
conducting several tests at the same time-increasing sensitivity and negative predictive value-more sensitive strategy – The degree of sensitivity in the parallel testing varies controlled by different individual tests.
69
What is Serial testing and why do we do it?
previous test results will decide whether continue the next test or not – Less risky, invasive, easier to do, and cheaper – The accuracy of serial testing depends on the assumption of independency of additional information offered by individual test. Ex: breast cancer: ultrasound, monograph, MRI
70
How do you calculate sensitivity and specificity for parallel testing?
Net sensitivity=sen1+sen2-sen1*sen2 Net specificity=spe1*spe2
71
How do you calculate sensitivity and specificity for serial testing?
Net sensitivity=sen1*sen2 Net specificity=spe1+spe2-spe1*spe2
72
What happens to the sensitivity and specificity when doing parallel testing vs serial testing?
Parallel: augmente sensitivity and diminue specificity Serial: augmente specificity and diminue sensitivity
73
What are 3 types of variation (reliability)?
• Intrasubject variation • Interobserver variation • Interobserver variation: Kappa
74
How do you calculate Kappa?
(Percent agreement observed) - (percent agreement expected by chance alone) / 100% - (percent agreement expected by chance alone)
75
What are bad, moderate and good scores for Kappa?
0.1-0.4 = bad 0.4-0.75 = moderate 0.75+ = good
76
How do you calculate the percent agreement?
(a+d/a+b+c+d) x100
77
How do you calculate the Positive percent agreement
(a/a+b+c) x 100
78
How do you calculate the Negative percent agreement?
(d/b+c+d) x100

HSS3110 - Epidemiology (6 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions