Screening, Sensitivity, Specificity/Probability and Risk

Question 1

What is the primary preventative intervention to reducing risk factors from disease?

Answer 1

-Immunisation
-Lifestyle choices (healthy eating, not smoking etc)

Question 2

What is the secondary intervention to reducing risk factors from disease?

Answer 2

SCREENING

Question 3

What are the criteria for Screening?

Answer 3

There must be significant levels of the disease in the population

Diseases with longer pre-clinical stages are more appropriate for detection by screening

There is treatment available if detected

The tests are accurate, inexpensive and acceptable to the public

Question 4

Why do we screen?

Answer 4

Early detection improves the outcome

Detecting disease at an early stage reduces both the financial and emotional cost to the patient

With diseases such as Glaucoma (with no symptoms in the early stages in open angle glaucoma), early detection means more simple treatment and preservation of vision

Question 5

What is a tests effectivity measured by?

Answer 5

-Sensitivity
-Specificity

Question 6

What does Sensitivity mean?

Answer 6

Its (tests) ability to identify the presence of disease

Question 7

What does Specificity mean?

Answer 7

Its (tests) ability to identify the absence of disease

Question 8

For a good test, does the sensitivity and specificity need to be high or low?

Answer 8

Higher sensitivity/specificity = better the test

Question 9

What would you call Px’s that have a disease and test shows a positive result?

True positives
False negatives
False positives
True negatives

Answer 9

True positives

Question 10

What would you call Px’s that do not have a disease and test shows a positive result?

True positives
False negatives
False positives
True negatives

Answer 10

False positives

Question 11

What would you call Px’s that have a disease and test shows a negative result?

True positives
False negatives
False positives
True negatives

Answer 11

False negatives

Question 12

What would you call Px’s that do not have a disease and test shows a negative result?

True positives
False negatives
False positives
True negatives

Answer 12

True negatives

Question 13

What does Bayes’ Theorem state?

Answer 13

Essentially states that it is tough to screen for a disease with a low prevalence as there are so few people with the disease and so many who are normal

Question 14

What are ways to reduce False Positives?

Answer 14

1) Repeat testing
2) Referrals to speciality optometry clinics

Question 15

Give the definition for probability

Answer 15

“Scaled-down group behaviour applied to a single individual”

Proportion ‘p’ of population exhibits a characteristic; a randomly selected individual exhibits same characteristic with probability ‘p’

Question 16

How do we write probabilities?

Answer 16

P(x) or Pr(x)

Question 17

What are the different kinds of probability?

Answer 17

-Classical probability
-Frequentist probability
-Subjective probability

Question 18

What is Classical probability?
Give some examples of Classical probability

Answer 18

Based on ideas of symmetry and equal likelihood applied to well understood objects

Examples :
-Tossing a coin – head or tail are equally likely (0.5)
-Drawing a red 6 from a standard deck of cards (only have 2 6’s that are red out of a deck of 52 cards) (6/52)
-Rolling a dice and it landing on a 6 (1/12)

Question 19

What is Frequentist probability useful for?

Answer 19

Useful for evaluating how likely a real-life events is, based on observation

Question 20

If there are two outcomes does this mean that the 2 outcomes are equally likely?

Answer 20

NO

Question 21

What is Subjective Probability?
Give some examples of Subjective probability

Answer 21

An opinion or degree of belief about how likely an event is

Examples :
-“There is a 30% chance of rain tomorrow”
-“There is a 75% chance I will attend tomorrow’s meeting”

Question 22

Give the definition for Experiment

Answer 22

Any procedure the outcome of which cannot be predicted with certainty

Question 23

Give the definition for Sample Space (S)

Answer 23

The set of all possible outcomes

Question 24

Give the definition for Event

Answer 24

Subset of interest from the sample space (S) consisting of at least one outcome from (S)

Question 25

Give the definition for Probability of an event

Answer 25

The relative frequency of this set of outcomes over an indefinitely large (or infinite) number of trials It is also the sum of probabilities of the individual outcomes it is composed of

Question 26

What can a probability value be between?

Answer 26

0 and 1 0 ≤ P(E) ≤ 1, where P(E) is the probability of event E

Question 27

What does it mean if the probability of an event, E is zero?

Answer 27

P(E) = 0 implies event E is IMPOSSIBLE

Question 28

What does it mean if the probability of an event, E is 1?

Answer 28

P(E) = 1 implies event E is CERTAIN

Question 29

What must hold true if two events are mutually exclusive?

Answer 29

Two events are mutually exclusive if they have no elements in common

Question 30

Give the formula for Mutual Exclusion (Addition Rule)

Answer 30

P(A or B) = P(A) + P(B) P(A or B or C or ...) = P(A) + P(B) + P(C) + …

Question 31

Give the formula when 2 events are not Mutually Exclusive (Addition Rule)

Answer 31

If A, B, C are not mutually exclusive events : P(A or B) = P(A) + P(B) - P(A and B)

Question 32

Why do we need to subtract the intersection (middle - P(A and B))?

Answer 32

To avoid counting the people in A and B twice

Question 33

Give the formula when 3 events are not Mutually Exclusive (Addition Rule)

Answer 33

P(A or B or C) = P(A) + P(B) + P(C) - P(A and B) - P(B and C) - P(C and A) + P(A and B and C)

Question 34

If events A and B are mutually exclusive, what can we say about P(A and B)?

Answer 34

P(A and B) = 0

Question 35

Give the formula when 2 events are independent (Product Rule)

Answer 35

P(A and B) = P(A)*P(B)

Question 36

What does it mean if events are Independent?

Answer 36

If knowing that B has occurred does not influence the probability of occurrence of event A

Question 37

Give examples of events that are not independent

Answer 37

-The probability of temperature dropping below zero, both today and tomorrow. The weather today is a good predictor of the weather tomorrow; the weather tomorrow is very dependent on the weather today -The probability of being colour-blind is not independent of gender; you are more likely to be colour-blind if you are male, than if you are female -The probability of finding someone over 6-feet tall is more likely of you are looking among professional basketball players, than if you are looking among professional jockeys -In all the examples above, the probability of an event is influenced by other events: the weather tomorrow is influenced by the weather today; colour-blindness is influenced by gender; probability of finding tall people is affected by group

Question 38

Give the formula for conditional probabilities

Answer 38

P(A given B) = P(A and B) / P(B)

Question 39

What does T | H mean?

Answer 39

Probability that you get tails the 2nd time, based on the fact (given) that you got heads the 1st time

Question 40

Define Risk

Answer 40

Chance that something will happen to you, provided it hasn’t already

Question 41

Formula to calculate risk

Answer 41

Number of incidents / Number of potential incidents

Question 42

What does risk usually imply?

Answer 42

Risk usually implies causation

Question 43

Who does risk refer to?

Answer 43

A defined or target population

Question 44

What are the two ways to compare risk?

Answer 44

-Relative risk -Attributable risk

Question 45

What is Relative risk?

Answer 45

Ratio of two risks

Question 46

What is Attributable risk?

Answer 46

The difference between two risks

Question 47

What are the better ways to present risk?

Answer 47

-Natural frequencies -Number needed to treat (NNTT)

Question 48

In general, what is the relationship between NNTT and attributable risk?

Answer 48

In general, the NNTT is one over the attributable risk

Question 49

What does Number Needed To Treat mean?

Answer 49

The number of patients you need to treat to prevent one additional bad outcome

Question 50

When does Confounding occur?

Answer 50

Confounding occurs when the real risk is : -Not recorded in the study -Associated with something that is recorded