3. Biostatistics Flashcards

1
Q

Define the 2 types of study data (broad) and their respective two categories.

A

Continuous data: logical values that increase/decrease by equal units (ie. HR, BP).
- Interval = has no impactful zero (ie. 0 Celsius just means freezing point of water)
- Ratio = has meaningful zero (ie. HR of 0 means no pulse, aka cardiac arrest)

Discrete data: categorical data (not on a continuous spectrum)
- Nomial = data sorted into arbitrary categories w/ names (ie. Male/female, yes/no)
- Ordinal = numerical ranked data, but difference between ranks is not equal (ie. 0-10 pain scale)

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

What three values are included in the “measures of central tendency”? Define each

A

MEAN: aka the average
MEDIAN: the middle value when numerical values are lined up lowest to highest
MODE: the value that occurs most frequently

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

What two values describe the spread of data? Define each

A

RANGE: the difference between the highest and lowest values
STANDARD DEVIATION: indicates dispersal from the mean, higher SD = more dispersed data

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

What is a Gaussian distribution?

A

When large sets of continuous data form a symmetrical, bell-shaped curve. Two-tailed
Mean, median and mode are all the SAME value
68% of the data falls within 1 standard deviation

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

Described “skewed data”? What kind of sample set and value is likely to cause this?

A

Skewed = non-symmetrical data
Usually occurs when sample size is SMALL, often with OUTLIERS

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

What is the null hypothesis (H0)?

A

The statement that there is NO significant difference between two study groups. AKA the statement researchers are trying to disprove.

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

What is the name for the opposite hypothesis from the null hypothesis?

A

The alternative hypothesis = states there is a significant difference between two study groups, and is what researchers are trying to PROVE.

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

What is the alpha level? What is it most commonly set to?

A

The maximum permissible error margin
Commonly set to 0.05 (5%)

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

How does the alpha value differ from the P value?

A

Alpha value is a permissible limit of error set BEFORE the study
P value is calculated AFTER with the data, and COMPARED to the alpha

AKA, if P < 0.05, the null hypothesis is rejected and the results are significant

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

What is the confidence interval? How is it calculated?

A

Represents the precision of the results along w the significance of the data (like the P value)
CI = 1- alpha [aka if alpha is set to 0.05, the CI will be 0.95 or 95%]

*NOTE: When using 95% CI for…
DIFFERENCE data = if the CI includes ZERO the data is non-significant
RELATIVE RISK, OR, HR = if the CI includes ONE the data is non-significant

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

What is the difference between Type I and Type II errors?

A

TYPE I: False positive (alpha error)
- When alpha is 0.05 and P value is falsely reported as <0.05, thereby falsely disproving the null

TYPE II: False negative (beta error)
- Occurs when null is accepted when it should have been rejected

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

What is study POWER?

A

Power = the probability that the null will be rejected CORRECTLY. Power to avoid type 2 error.
Aka = 1 - (beta)
Determined by the # of outcome values collected, difference in outcome rates, and alpha level.

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

Relative risk equation

A

(X/total subjects in TREATMENT group)
RR = ———————————————————
(Y/total subjects in CONTROL group)

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

Relative risk reduction equation

A

1 - RR
OR

(% risk in control - % risk in treatment)
——————————————————
(% risk in control)

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

Absolute risk reduction equation

A

ARR = (% risk in control group) - (% risk in treatment group)

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

Number needed to treat (NNT) equation

A

NNT = 1/(risk in control - risk in treatment)
OR
NNT = 1/ARR

17
Q

NNH equation

A

1/NNT

18
Q

Odds ratio equation

A

(+exposure+outcome X -exposure-outcome) [desired outcomes]
————————————————————— AKA ————————————
(+exposure-outcome X -exposure+outcome) [unexpected outcomes]

19
Q

Hazard ratio equation

A

(# incidents/total treatment group) [Hazard rate TREATMENT]
HR = ————————————————— —————————————
(# incidents/total control group) [Hazard rate CONTROL]

20
Q

What type of statistical tests can be used on CONTINUOUS data?

A

T-tests: used when data is normally distributed, ≤2 groups of data
Analysis of variance (ANOVA): used when there are 3+ groups of data.

21
Q

Describe the different types of T-tests

A
  • When data is compared to a known data set, a one-sample t-test is used.
  • When data is used as pre/post-measurement, a paired t-test is used.
  • When 2 independent samples (treatment & control), a student t-test is used.