Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

STEP3 > Biostats > Flashcards

Biostats Flashcards

1
Q

What does it mean for a data set to be skewed to the right or left? How do mean and median relate in these populations?

A
  • to the right is because there outliers above the data set, causing the mean to be above the median
  • to the left is because there are outliers below the data set, causing the mean to be below the median
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is nominal data?

A

qualitative data for which there is no particular order

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

What is ordinal data?

A

that which has a particular order (i.e. numerical)

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

How is incidence defined?

A

the number of new cases over a particular time

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

How is prevalence defined?

A

the number of existing cases at any particular time

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

What are the differences between accuracy and precision in statists?

A
  • accuracy is equivalent to valid, the combination of sensitivity and specificity, how true it is
  • precision is how immune it is from randomness and scattering
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How much data is included in 1, 2, and 3 standard deviations?

A
  • 1 SD includes 68% of the data
  • 2 SD includes 95% of the data
  • 3 SD includes 99.7% of the data
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

How is standard error of the mean calculated? What does it mean?

A
  • it is calculated as standard deviation divided by the square root of n
  • it means that as n increases, the data becomes clustered more tightly around the mean and the data is more precise
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How should confidence interval be interpreted?

A
  • a narrower range suggests more precise data

- when CI includes 1, this means the results are not significant

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

How can you narrow a confidence interval?

A

since a 95% confidence interval is 2 SEM, and SEM is calculated as SD/sqrt(n), taking four times as many measurements will narrow the CI by half

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

How do t-tests, ANOVA tests, and chi-squared tests relate?

A
  • t- and ANOVA tests answer “are the means between these groups different”
  • chi-squared tests answer “are these two groups related” and utilize when data comes in discrete categories
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Describe cohort and case-control studies.

A
  • cohort: an observational, prospective study that takes two groups (one with an exposure and the other without) and then compares disease incidence, uses relative risk
  • case control: a retrospective observational study that takes two groups (one with disease and the other without) and then compares prior exposures, uses odds ratio
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Describe a cohort study. What statistic is generated from such studies?

A
  • an observational, prospective study in which two groups (one with an exposure and one without) are analyzed and evaluate for disease incidence
  • produces a relative risk
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Describe a case control study. What statistic is generated from these?

A
  • an observational, retrospective study that takes two groups (with disease and without) and then analyzes prior exposures/risk factors
  • produces an odds ratio
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is the difference between risk ratio and odds ratio?

A
  • risk ratio comes from prospective cohort studies

- odds ratio comes from retrospective case control studies

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

Calculate relative risk.

A

RR = ratio of exposed to get disease/ratio of unexposed get disease

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

Calculate odds ratio.

A

OR = (diseased exposed/diseased not exposed) / (controls exposed/controls not exposed)

18
Q

What are the following biases:

  • Berkson
  • Hawthorne
  • Lead-Time
A
  • Berkson: when hospitalized patients are used in a trial instead of the general population
  • Hawthorne: those being studied know they are being watched and this changes their behavior
  • Lead-time: early detection is confused with increased survival
19
Q

What is lead-time bias?

A

when early detection is confused with increased survival

20
Q

What is the Hawthorne effect? How is it prevented?

A
  • when those being studied know they are being watched

- fixed with placebo-control and double blinding

21
Q

What is Berkson bias?

A

when hospitalized patients are used in a trial instead of the general population

22
Q

What does a p value of 0.5 mean?

A

that if a study were repeated, there is a 95% chance it would produce findings consistent with the trial’s findings

23
Q

What are type I and type II errors?

A
  • type I is the alpha error, is a false positive, and is rejecting the null hypothesis when it is really true
  • type II is beta error, is a false negative, and is accepting the null when it is false
24
Q

Define sensitivity. How is it calculated?

A
  • it is the likelihood that a test will detect all the people with disease
  • sensitivity = TP/(TP + FN)
25
Q

Define specificity. How is it calculated?

A
  • it is the likelihood that people without a disease are correctly identified as disease-free
  • specificity = TN/(TN + FP)
26
Q

What is the biggest between sensitivity and specificity versus positive and negative predictive value?

A

sensitivity and specificity are features of the test and don’t change with population statistics like disease prevalence whereas PPV and NPV do

27
Q

How is positive predictive value calculated?

A
  • TP / TP + FP

- the odds that if you have a positive test that you have disease

28
Q

How is negative predictive value calculated?

A
  • TN / TN + FN

- the odds that if you have a negative test that you don’t have disease

29
Q

Define:

  • sensitivity
  • specificity
  • PPV
  • NPV
A
  • sensitivity: the likelihood that that a test will identify everyone with disease; TP / (TP + FN)
  • specificity: the likelihood that people without disease are correctly identified; TN / (TN + FP)
  • PPV: the likelihood that you have the disease given a positive test; TP / (TP + FP)
  • NPV: the likelihood that you don’t have the disease given a negative test; TN / (TN + FN)
30
Q

How do PPV and NPV change with disease prevalence?

A
  • PPV increases with increasing disease prevalence

- NPV decreases with lower disease prevalence

31
Q

How is absolute risk reduction calculated? relative risk reduction?

A
  • ARR is calculated as the difference in percentage the two groups
  • RRR puts the ARR over the risk in the control population (usually increasing it)
32
Q

What is the coefficient of determination and how is it calculated?

A
  • it is the amount of variation in one variable explained by the other
  • calculated as r^2
33
Q

How is the accuracy of a test calculated?

A

accuracy = TP + TN / all tests

34
Q

What is receiver operating curve? How is accuracy determined?

A
  • it is a graph of sensivitiy versus 1 - specificity

- accuracy equals the area under the curve

35
Q

What is a positive likelihood ratio? How is it calculated?

A
  • calculated as sensitivity / (1 - specificity)

- it represents the frequency with which tests are positive in diseased individuals compared to those without disease

36
Q

What kind of bias forms when there is a difference in follow up rates between the treatment and control groups?

A

selection bias

37
Q

What is non-response bias?

A

bias in a study that originates because participants are asked whether they want to participate

38
Q

What is Neyman bias?

A

a type of selection bias that results when incidence is estimated based on prevalence and data become skewed based on selective survival

39
Q

What is susceptibility bias?

A

a type of selection bias that arises when the therapy offered is dependent on the patient’s disease severity

40
Q

What is a Fisher’s exact test?

A

like a chi-squared test, comparing proportions, but with small sample size

41
Q

What must be true of a variable to make it a confounder?

A

it must have some properties linking it to both the exposure and outcome

42
Q

How is confounding controlled for?

A

randomization or matching as well as stratified analysis

STEP3 (1 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions