EBM

Question 1

What does the PICO framework stand for

Answer 1

PICO framework helps define key elements in a research question
P: population
I : intervention / E : exposure
C : control or comparator
O : outcome

Question 2

What does E in PECO framework mean and when is it used

Answer 2

Exposure, and is when studies are looking at the effect of an expose on the risk or progression of a disease

Question 3

Ecological study

Answer 3

Looks at data for large groups or populations not individual

E.g investigating if a population that smokes more has higher rates of heart disease

Question 4

Cross-sectional survey

Answer 4

Collects data from a group at a single point in time

E.g measuring the average height of 3 yr olds today

Question 5

Case-control study

Answer 5

Looks back at the past to find potential causes for a condition

E.g comparing the past habits of cancer patients to people without cancer to see what increases cancer risk

Weakness: it finds associations but can’t prove that one causes the other

Question 6

Cohort study

Answer 6

Follows a group of people over time to see who develops a condition and compares outcomes between those exposed to a factor and those not

E.g observing smokers and non-smokers over years to see who develops lung cancer

Question 7

Randomised controlled trial

Answer 7

“Gold standard” for resting treatments, participants are randomly assigned to either receive a treatment or placebo and monitored

E.g testing if a new drug reduces heart attacks compared to a sugar pill (placebo)

Question 8

Cross-over trial

Answer 8

Each participant gets multiple treatments one after the other with a break (washout period) in between and evaluated

E.g testing two diets by giving each person both diets at different times

Question 9

Systemic review

Answer 9

Combines the results of many studies to summarize what is known about a topic

E.g reviewing all research on whether excersise reduces depression

Question 10

Meta-analysis

Answer 10

An analysis that combines data from multiple studies to produce a single clear result or statistic

E.g combining results from many studies on a drug to determine its overall effectiveness by giving a percentage, for example, 7%

Question 11

Sensitivity definition and formula

Answer 11

The ability of the test to correctly identify people who have the disease

True positive / true positive + false negatives

True positive are people who 100% have the disease

Question 12

Specificity its definition and formula

Answer 12

The ability of the test to correctly identify people who have don’t the disease

True negative / True negatives + False positives

True negatives are people who 100% don’t have the disease

Question 13

Absolute risk definition and formula

Answer 13

The chance or probability that someone will develop a disease over a certain period

Number times someone developed the event / number of individuals

If 3 people out of 100 patients developed diabetes then absolute risk is
3/100 = 3%

Question 14

Relative risk definition and formula

Answer 14

Compares the risk of a disease in one group compared to another (smokers vs non-smokers

RR = risk in treatment group (or exposure group) / risk in control group

If RR is 2 it means that the group smoking has twice the amount of risk of getting disease then people who weren’t and vice versa for a score of 0.5

Question 15

Case report

Answer 15

A detailed description of a single patients medical history and treatment

Question 16

Case series

Answer 16

A collection of case reports involving several patients with the same condition or outcome

Question 17

Reliability

Answer 17

The extent to which a procedure yields the same result in repeated trials

Question 18

Validity

Answer 18

How well a study or test measures what it is supposed to measure

Question 19

What is used to summarize categorical data

Answer 19

N(%)
Relative risks
Odds ratios
Risk differences

Question 20

What is used to summarize numerical data that has a normal distribution

Answer 20

Mean and SD
Median and IQR
Mean difference

Question 21

What is used to summarize numerical data that has no assumptions about distribution

Answer 21

Mean and SD
Median and IQ
(Both used for normal distribution)
(Not mean difference)

Question 22

What graphs are used to show categorical data

Answer 22

Bar charts
Pie charts
Tables

Question 23

What graphs are used to show numerical data

Answer 23

Histograms
Box and whisker plots

Question 24

What statistical tests are used to to compare categorical data

Answer 24

Chi-squared test
Fishers exact test (2 groups and no assumptions)

Question 25

Chi squared test results meaning and examples

Answer 25

A result (x^2) greater than critical value (0.05) would indicate that the difference was unlikely due to chance (it wasn’t change that created a difference) and the opposite is true for a value of less than critical value (0.05)

E.g
Supposed you flip a coin 100 times, you’d expect 50 tails and 50 heads, if you get 60 heads and 40 tails a chi-squared test will tell you if the difference was up to chance or not

Question 26

What statistical test would you use to compare groups of numerical data

Answer 26

Normal distribution:
T-test (only 2 groups)
ANOVA (more than 2 groups)

No assumptions about distribution:
Mann-Whitney U test

Question 27

T-test results meaning and example

Answer 27

A t-test produces a p-value in which you compare to the common threshold (0.05). If the p-value is less than 0.05 then the result is significant and wasn’t up to chance, if the p-value is less than 0.05 then the result is not significant and was up to chance.

This is the opposite to the x^2 value compared to the critical value where if the value is >0.05 then it is significant

Imagine Chi-square test is like chi from dragon balls, you want that value to be as high as possible (over 9000!) so u want x^2 value to be over 0.05 (to be significant) while in a T-test you have tea in small cups so you want the values to be <0.05 to be significant (to be significant)

Question 28

How to remember what is significant in t-test compared to chi squared test

Answer 28

Imagine Chi-square test is like chi from dragon balls, you want that value to be as high as possible (over 9000!) so u want x^2 value to be over 0.05 (to be significant) while in a T-test you have tea in small cups so you want the values to be <0.05 to be significant (to be significant)

Question 29

How can you calculate the reference range for standard distribution data

Answer 29

Reference range will be:
mean +/- (1.96 x the standard deviation)

Question 30

When do you use regression and what are the two types

Answer 30

When you want to adjust for multiple factors in an experiment like including confounders such as age, diet, excersise etc

Logistic regression
Linear regression

Question 31

Difference between linear and logistic regression

Answer 31

Logistic regression is used for binary outcome measure (yes no) (categorical)

Linear regression is used for numerical outcome measures (numerical)

Makes sense as linear is usually to do with numbers and logistic sounds like logic which links with yes and no (binary/categorical)

Question 32

What must you have to conduct a t-test

Answer 32

A normal distribution of data and 2 groups of data only