Statistics Flashcards

1
Q

How do you calculate the standard error of the mean?

A

Standard error of the mean (SEM) is an indicator of how the true mean of the population would differ from the mean of your sample

Standard error of the mean = SD / square root of (n)

SD = standard deviation
n = sample size

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

What is the 95% confidence interval

A

The range of values within which you are 95% certain the true value lies within.

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

How do you calculate the 95% confidence interval?

A

Upper limit = mean + (1.96 x SEM)
Lower limit = mean - (1.96 x SEM)

1.96 is the critical value

Note - often easier to change this to 2x SEM to make calculation easier

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

How do you calculate the positive predictive value of a test?

A

Positive predictive Value = Number of true positives / number of true positives + number of false positives

PPV = TP / TP + FP

Positive predictive value is the probability that a positive result is correct

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

How do you calculate the negative predictive value of a test?

A

Negative predictive value is the probability that a negative result is truly negative

Negative predictive value = the number of true negatives / number of true negatives + number of false negatives

NPV = TN / TN + FN

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

How do you calculate the sensitivity of a test?

A

The sensitivity of the screening test is how much of the true disease cases are detected by the the test.

Sensitivity = Number of true positives from the test / total number of people found to have the disease

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

How do you calculate the specificity of a test?

A

The specificity of the screening test is the probability of a negative test result being truly negative

Specificity = the number of true negatives / the total number of disease-free people

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

What is a cohort study?

A

A cohort study is a prospective study that follows people up that are exposed to an intervention or exposure PROSPECTIVELY and see what outcomes they have

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

What is a case-control study?

A

A case-control study is a RETROSPECTIVE study that looks back to see what exposures / interventions people had in the past that ended up with a certain outcome

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

What are the different levels of evidence that are referred to in evidence-based medicine?

A

Level I is the best

Level I - evidence from systematic review or meta-analyses or 3+ RCTs with similar results or evidence-based guidelines using SR / MAs.

Level II - from a single good RCT

Level III - from a well-designed but non-randomised trial

Level IV - from case-control or cohort study

Level V - expert opinion

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

How to calculate the odds ratio ?

A

Odds ratio - the odds of having the disease in the exposed group compared to the odds of having the disease in the control group.

OR = (exposed with the disease / exposed without the disease) / (control/non-exposed with the disease / control without the disease)

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

In a normally distributed dataset, what percentages of the data lie within 1SD, 2SD and 3SDs of the mean?

A

In normal distribution,
68% of the value should lie within 1 standard deviation of the mean
95% within 2 SDs
99% within 3 SDs

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

How to calculate the number needed to treat from an intervention?

A

(1 / absolute risk reduction) x 100

For example if risk of adverse outcome is 10% in untreated group and 5% in the treated group, the absolute risk reduction is 5% (10% - 5%).
So the NNT = 1/5 x 100 = 20

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

What r values would indicate weak, moderate and strong correlation respectively?

A

R < 0.3 = weak correlation
R 0.3 - 0.7 = moderate correlation
R > 0.7 = strong correlation

(If R is positive = positive correlation, if R is less than 0 = negative correlation. R of 1 or -1 = perfect linear relationship)

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

What is the Kaplan Meier estimate used for?

A

A statistical tool used to estimate the survival rates from lifetime data

Used in medical research to estimate the survival rates at certain time points after treatment

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

What is a censored survival time?

A

A censored survival time is a time point where you know a participant was definitely alive at that point in time but you don’t know how much longer after that time point they survived for

17
Q

What is middle ground research?

A

Middle ground research is research focused on the healthcare needs of patients themselves.

It is normally more locally ran and over shorter timeframes so that findings can be implemented in practice sooner.

18
Q

What is the standardised mortality rate?

A

The baseline standardised mortality rate is expressed as either 1 or 100

Ie an SMR of 100 is the same as would be expected of the standard population

If an exposure leads to an SMR > 100 (or >1 if expressed in that unit) then it carries a higher mortality

If less than 100 (<1) then a lower mortality

19
Q

What is heterogeneity in the context of a meta-analysis and what is an agreed % above which heterogeneity is considered significant?

A

Heterogeneity is how different the studies included in the MA are. Ideally they should be as similar as possible

Higgins value (I^2) is a marker of heterogeneity

I2 = 0 all studies are homogenous

I2 of 50% or higher is considered significant