Biostats primer

Question 1

What is meant by epidemiology?

Answer 1

the study of health and illness in human populations, or, more precisely, to the patterns of health or disease and the factors that influence these patterns

Question 2

Where does the word epidemiology come from?

Answer 2

Greek words for “upon” (epi) and “people” (demos).

Question 3

What is epidemiology used for?

Answer 3

to understand the cause of the disease, determine public health policy, and plan treatment

Question 4

What is the number needed to treat (NNT)?

Answer 4

The number of patients that need to be treated with a proposed therapy in order to prevent or cure one individual; it is the reciprocal of the absolute risk reduction (1/ARR).

Question 5

What is the absolute risk reduction?

Answer 5

The reduction in risk with a new therapy compared with the risk without the new therapy; it is the absolute value of the difference between the experimental event rate and the control event rate (|EER – CER|).

Question 6

What is the control event rate (CER)?

Answer 6

The number of subjects in the control group who develop the outcome being studied

Question 7

What is the experimental event rate (EER)?

Answer 7

The number of subjects in the experimental or treatment group who develop the outcome being studied

Question 8

What is sensitivity (use your words)?

Answer 8

Refers to how well a given test does at identifying a disease

Question 9

How is sensitivity calculated?

Answer 9

Test positive/true positives + false negatives

Question 10

What is specificity (use your words)?

Answer 10

Refers to how well a given test identifies people who are well (i.e. they don’t have the condition)

Question 11

How is specificity calculated?

Answer 11

Test negatives/true negatives + false positives

Question 12

What is positive predictive value (use words)?

Answer 12

the chance that a positive test result will be correct; to calculate this, you take the number of true positives divided by the test positives which will give you a percentage. 38 people test positive on a given test, 24 of them actually have the condition, thus PPV=63%

Question 13

What is negative predictive value?

Answer 13

chance that a negative test result is correct; 62 people test negative and 56 of them truly don’t have the condition then the NPV=90%

Question 14

Do sensitivity and specificity values change if the prevalence of a disease changes?

Answer 14

No, but the PPV and NPV will change if the prevalence of the disease changes. The PPV will fall when prevalence falls, while the NPV rises as prevalence falls. This isn’t surprising if you think about prevalence as the probability that someone has the disease before we actually do the test.

Question 15

What is effect size?

Answer 15

The magnitude of the difference between groups in a study. Useful when measurements have no intrinsic meaning (e.g. score on Likert scales), when studies have used different scales so no direct comparison is possible, or when effect size is examined in the context of variability in the population being studied

Question 16

What is the absolute effect size?

Answer 16

The difference between the average (mean) outcomes in two different intervention groups. This is a useful measure when variables under study have intrinsic meaning (e.g. # of hours of sleep)

Question 17

Why should studies report effect size?

Answer 17

Because it is the main finding of a quantitative study. P values tell you if there was an effect, but doesn’t tell you how big the effect was. Effect size is the substantive significance, while P value is statistical significance.

Question 18

What is type II (or beta) error?

Answer 18

Probability of concluding that there is no effect when actually there is.

Question 19

What is power?

Answer 19

Refers to whether or not a study has enough participants to ensure that there won’t be a type II error. In other words, the sample was too small to detect an effect that actually would show up if the study population were bigger.

Question 20

What is a odds ratio and when is it used?

Answer 20

Used when reporting binary outcomes and compares the odds of outcome occurring from one intervention vs another

Question 21

How is odds ratio calculated?

Answer 21

Odds of outcome in group 1/odds of outcome in group 2

OR=1 then it’s equally likely in both groups

Question 22

Explain statistical significance.

Answer 22

Stat sig=probability that an observed difference between 2 groups is due to chance.

Question 23

How is P value related to statistical significance?

Answer 23

If P value is larger than the alpha chosen (e.g. 0.05), then any observed differences between the groups is assumed to be due to sampling variability. Level of significance is alone does not predict effect size.

Question 24

Why is reporting P value not enough to make sense of a given result?

Answer 24

Because with a sufficiently large sample, a statistical test will almost always show statistical significance even when these differences are clinically meaningless; i.e. even when an effect size is very small, you can see statistical significance when the study sample is big; thus you need effect size to understand clinical significance

Question 25

How is effect size calculated?

Answer 25

Difference in outcomes between intervention and control groups divided by the standard deviation