Epi & biostats Flashcards

1
Q

what does descriptive epidemiology look at?

A

who? - populations, persons, sex, age

what? - disease, events, injuries, beliefs

where? - geographical,

when? - time frames

how much?

NOT WHY

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

What are binary variables?

A

variables that can be divided into one vs the other.

eg:

died vs survived

IHD vs No IHD

continuous can become binary

eg:

BP - hypertension/no hypertension

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

define incidence

A

Incidence refers to the number of new cases occurring ( = arising = incident) in a specified population during a specified time period.

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

define prevalence rate or proportion

A

Prevalence rate (more correctly prevalence proportion) is the proportion of members of a population who:

  • have a disease or other attribute of interest at a particular moment (point prevalence rate or proportion)
  • or at any time during a particular period (period prevalence rate or proportion).
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

define point prevalence

A

Point prevalence refers to the number of cases of a disease or other relevant attribute (e.g. smoking habit) present in a specified population at a specified time.

The cases could have occurred (i.e. arisen, become incident, begun) at any time up to the moment at which point prevalence is assessed.

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

what does relative risk mean?

A

Relative risk or risk ratio (RR) is the ratio of the probability of an event occurring (for example, developing a disease, being injured) in an exposed group to the probability of the event occurring in a comparison, non-exposed group.

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

risk ratio = R1/R0

so the relative risk for men/relative risk for women

0.20/0.10 = 2

the risk difference is 0.10 (10%)

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

with regards to risk ratios, what would the following results tell you?

RR = 1

RR = >1

RR = <1

A
  • RR = 1 = the risk is the same for both exposed or non-exposed groups
  • RR = >1 = the risk is greater in the exposed group
  • RR = <1 = the risk is reduced for the exposed group
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

If the RR is 0.6 what does this tell us?

A

risk is 40% lower in the exposed group than the non-exposed group.

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

How do you calculate Incidence Density (ID)?

A

The number of new incidences of disease within a specified period (eg new cases of prostate cancer within 5 year follow up)

divided by

the total time at risk of contracting a disease for all studied (so add up the time from year one to disease onset, loss to follow up, death, or til the end of the study for all participants)

In the image 4/39. Round up to 100 person year risk

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

how is the relative risk calculated?

A

Relative risk (RR) is the ratio of the risk (or rate) of a disease or other outcome of interest occurring among people who were exposed to a risk factor or a treatment of interest, divided by the risk (or rate) of developing the outcome without exposure to the risk factor or treatment.

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

Describe what is meant by number needed to harm (NNH)?

A

The number of people who would need to be treated with the new medication for one bad outcome of the treatment to occur, which would not have ocurred if the other treatment had been used (current medication)

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

how do you calculate the NNH?

A

NNH = 1/absolute risk difference

absolute risk diff = risk with the treatment – risk without the treatment (R1 - R2)

so

NNH = 1/R1 - R2

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

how do you calculate the NNT?

A

NNT = 1/(risk difference)

NNT = 1/(R2 - R1)

this calculates for 10,000 subjects.

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

How do you calculate odd?

A

Odds = odds that something will happen

= probablility of event happening/probability of event not happening

eg:

total subjects 2775

female 1275 (odd being F = 1275/1500

male 1500 (odd being M = 1500/1275

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

what is the proportion?

A

the proportion of times an event happens, ie the risk of the event occuring

17
Q

how do you calculate the proportion?

A

divide the number of the event (eg men) by the total number of subjects

proportion men 1500/2775 (total subjects) = 0.54 (or 54 %)

18
Q

how can you calculate the probablility from the odds?

A

probability = odds/(1 + odds)

eg:

odds of being male = 1.18

1.18/1+1.18) = 0.54

19
Q

how do you calculate the odds from the probaility?

A

odds = probability/(1 - probaility)

20
Q

what is the attributable risk?

A

The Attributable Risk indicates the number of cases of a disease among exposed individuals that can be attributed to that exposure:

AR = Incidence (exposed) − Incidence (unexposed)

as a rate it requires a unit eg N per 1000

21
Q

what is the odds ratio?

A

An odds ratio (OR) is a measure of association between an exposure and an outcome.

The OR represents the odds that an outcome will occur given a particular exposure, compared to the odds of the outcome occurring in the absence of that exposure.

22
Q

how do you calculate the odds ratio?

A

(exposed with disease/exposed without)

divided by

(unexposed with disease/unexposed without)

23
Q

what is sensitivity?

A

The sensitivity of a diagnostic test is the likelihood that persons with the disease of interest will have positive test results.

24
Q

what is specificity?

A

The specificity of a diagnostic test is the likelihood that persons who do not have the disease of interest will have negative test results.

25
Q

how would you draw a table to assess rates of truth of a diagnostic test?

A
26
Q

how can you remember what factors to use to determine sensitivity, specificity, PPV and NPV?

A
27
Q

how do you calculate sensitivity?

A
28
Q

how do you calculate specificity?

A
29
Q

what makes a good screening test?

A
  • Reliable - reproducible and repeatable results.
  • Valid - The screening test should be sensitive and specific.
  • The screening test must be acceptable to the target population.
  • Minimal risk should be associated with the screening test.
  • Diagnostic work-up for a positive test result must have acceptable morbidity given the number of false-positive results.
30
Q

what makes a good screening program?

A
  • Morbidity or mortality of the disease must be a sufficient concern to public health.
  • A high-risk population must exist.
  • Effective early intervention must be known to reduce morbidity or mortality.
  • facilities for screening and treatment option should be available.
  • testing should be feasible and cost effective
31
Q

what is positive predictive value?

A

Positive predictive value is the probability that subjects with a positive screening test truly have the disease

32
Q

what is the negative predictive value?

A

Negative predictive value is the probability that subjects with a negative screening test truly don’t have the disease

33
Q

how do you calculate the PPV?

A

TP/TP + FP

34
Q

how do you calculate the NPV?

A

TN/TN + FN

35
Q

if the sensitivity of a test is 60% what does this tell you?

A

it correctly identifies 60% of true positives, but misses 40% of people with the disease (ie 40% are false negatives)

36
Q

if the specificity of a test is 49% what does this tell you?

A

that 49% of those tested are truely negative, but 51% test positive for a disease they do not have

37
Q

in a case control study, what does an odds ratio of 2.5 tell you?

A

that the odds of being exposed are 2.5 times higher in the disease group

38
Q

what does the confidence interval tell you?

A

A confidence interval is the range of numerical values in which we can be confident (to a computed probability, commonly 95%) that the true population value being estimated will lie.

39
Q

what other values do you need to look at before stating a CI is statistically significant?

A

Risk or odds ratio - cannot be 1

Risk or rate difference - cannot be 0