Stats

Question 1

type of data

Answer 1

1. Forrest plots (blobbograms)
2. Histogram
3. Stem and leaf diagram
4. Pie charts

pie and bar are discrete + categorical
stem+ leaf, histogram + box plot show continuous data

Question 2

Critical appraisal RAAMbo

Answer 2

R – Representative?
A – Allocated or Adjusted?
A – Accounted for?
Mbo – Measurement blind or objective?

Question 3

CASP checklist for RCT

Answer 3

1. Did the study ask a clearly focussed question?
2. Was it an RCT, and appropriately so?
3. Were participants appropriately allocated to control and intervention groups?
4. Were all persons blind to participants study group?
5. Were all participants accounted for?
6. Was there consistency between groups?
7. Did the study have enough participants?
8. How well are results presented, what is main result?
9. How precise are the results?
10. Were all important outcomes considered?

Question 4

Concepts of an RCT

Answer 4

Randomization, confounding,bias

ethical issues, cost, attrition

Question 5

ways of randomising

Answer 5

simple, block, stratified

Question 6

blinding types:

Answer 6

participants, investigators and/or assesssors unaware of group allocation

Question 7

bias

Answer 7

systemic disposition of certain trial designs to produce results consistently better or worse than other trial designs

Question 8

Cohort study:

Answer 8

incidence study, follow a group of people over a period of time

Question 9

case-control:

Answer 9

looks at people with a disease and compares with a control

Question 10

cross-sectional

Answer 10

prevelence study

Question 11

ecological studies

Answer 11

population based data rather than individual data

Question 12

types of data

Answer 12

categorial/qualitative

quantitative

Question 13

Standard deviation

Answer 13

the average distance of the observations from the mean value. It is used to find abnormal results or “outliers”

Question 14

The normal distribution

Answer 14

This is a bell-shaped curve where 2/3rds of the data lies within 1 standard deviation of the mean and 95% lies within 2 standard deviations of the mean. The median and the mean will be the same in a normal distribution.

Question 15

reference range

Answer 15

A reference range gives limits within which we would expect the majority of data to fall. For normally distributed data we would use 2 standard deviations above or below the mean (95% of the data).

Question 16

what do larger samples reduce?

Answer 16

the standard error of the mean. SE’s quantify how good an estimate a sample result is likely to be.

Question 17

types of random sampling

Answer 17

• Simple random: all within a population are equally likely to be picked
• Stratified random sampling: population divided into groups and then randomly sampled within those groups
• Cluster sampling: rather than sample individuals, groups or clusters of individuals are samples.

Question 18

Standard error

Answer 18

standard deviation of all the sample means

Question 19

CIs

Answer 19

The sample mean must be assessed to see how good it is as an estimate of the true population mean.

A confidence interval can be constructed from the sample mean and standard error.
The 95% confidence interval is found between two standard errors (1.96 standard errors) above and below the mean. only 5% chance than the range excludes the true mean of the population. It is smaller with larger sample sizes. If the size of the sample is squared, the confidence intervals are half the size.

Question 20

estimation and hypothesis testing

Answer 20

• Set null hypothesis H0 and study hypothesis H1
• Carry out significance test
• Obtain test statistic
• Compare test statistic to hypothesised critical value
• Obtain P-value
• Make a decision

looking to disprove the null.

Question 21

test statistic

Answer 21

reduces the data to a single value.

TS = (observed value - hypothesised value) / standard error of the hypothesised value

Question 22

significance test

Answer 22

The test statistic is compared to a hypothesised critical value (using a distribution we expect if the null hypothesis is true) to obtain a P value.

Question 23

p value

Answer 23

P is the probability of obtaining the test statistic from the data, assuming that the null hypothesis is true. It is the probability of committing a false positive error, i.e. rejecting the null hypothesis when it is actually true.
If the P value is very small (less than 0.05) the null hypothesis can be rejected. The P value ranges from 0-1.

Question 24

power of a study

Answer 24

The probability of rejecting the null hypothesis when it is actually false is called the power of the study. It is the probability of concluding there is a difference when a difference truly exists.

Question 25

statistical and clinical significance

Answer 25

A clinically significant difference is one that is big enough to be worthwhile. It is important that the size of the sample is adequate to detect the clinically significant result, at the 5% significance level with at least 80% power.

Question 26

risk

Answer 26

The risk is the incidence divided by the population. This is also known as absolute risk as is the probability that an event will occur. I.e. person A had a 10% risk of getting cancer

Question 27

relative risk

Answer 27

Relative risk is the risk of an event in an exposed group divided by the risk in the not exposed group

Question 28

Absolute risk difference/risk reduction/risk excess

Answer 28

absolute additional risk of an event following an exposure.

ARD = risk in an exposed group - risk in unexposed group

i.e 5% increase in risk in developing cancer if you take the pill

Question 29

number needed to treat to benefit

Answer 29

This is the additional number of people you would need to treat in order to cure one extra person compared to the old treatment.

number needed to treat = 1/absolute risk reduction

Question 30

number needed to treat to harm

Answer 30

For a harmful exposure the number needed to harm is the additional number of individuals who need exposure to the risk in order to have one extra person develop the disease compared to an unexposed group.

number needed to harm = 1/absolute risk difference

Question 31

odds

Answer 31

The odds of an event is the ratio of the probability of an occurrence compared to the probability of a non-occurrence.

Odds = probability/(1-probability)

Question 32

odds ratio

Answer 32

(P exposed/ 1 - Pexposed)/(p unexposed/ (1- punexposed))

for case control studies it is not possible to calculate the relative risk so the odds ratio is used

Question 33

whats used for cross-sectional and cohort studies?

Answer 33

odds ratio and relative risk can be used if its not clear which is the IV and the DV because it is symmetrical

Question 34

critical appraisal

Answer 34

The purpose of critical appraisal is to assess and consider validity, reliability and applicability. It should be done so that you can apply results to your own patients, provide your patient’s the best possible evidence when communicating risk and to remain professional.

Question 35

validity

Answer 35

Validity is how close to the truth something is. I.e. is a study testing what it says it’s testing or are there confounding variables which are in fact the reason for the results.

Question 36

reliability

Answer 36

Reliability is how consistent results are. If the experiment was repeated again, would the same/similar results be seen?

Question 37

applicability

Answer 37

how relevant a study is to clinical medicine