# Stats in EBM Flashcards

What is incidence?

A rate measuring the number of new cases of a particular disease arising in a population at a certain time period (the tap in the bath analogy)

What is prevalence?

A ratio of all cases of disease existing in a given population at a given time (the bath in the bath analogy)

What is risk ratio?

Ratio of incidence rate in group A vs incidence rate in group B

eg 25/100 get disease in A, 10/100 people get disease in B

Risk ratio = 0.25/0.01 = 2.5x more risk of death in group A

What is odds ratio?

Odds of outcome in exposed group vs odds of outcome in unexposed group

Eg:

In group A 25/100 got disease so thats 75 that didn’t. So 25:75 = 1:3

In group B 10/100 got disease so thats 10:90 or 1:9

Odds ratio = 1:3 / 1:9 = 3

(because 1/3 x 9/3 = 9/3 = 3)

What is the absolute risk?

Risk of acquiring given disease over given period of time eg 35/100 = 35%

What is absolute risk difference?

Difference between each risk - eg if group A 25% got disease and group B 10% did, risk difference is 15%

Why is risk difference useful?

Can express as relative risk so eg 25/100 vs 10/100

Group A is 2.5x more likely to get the disease - ‘more than doubles your risk’

Can seem more scary

What is P value?

Measure of statistical significance - probability that the effect could have occured due to change

Therefore p value less than or equal to 0.05 is enough to reject the null and be statistically significant

Problems with P values

(3)

Can **never rule out** chance

Gives no indication of **size** of the effect which is statistically significant

P value gives **no idea of the range of uncertainty **- this is what confidence intervals are for

What is a confidence interval?

Gives range of uncertainty around the estimate you have found

When is the confidence interval statistically significant?

If it does **NOT contain 1 - it is significant**

If it does not contain 1 p value will always be less than or equal to 0.05

What is the line/point of no effect in confidence intervals for relative risk ratio?

1 is the point of no effect - if ratio is 1 this means outcomes are same

What is the ‘mean absolute difference’? aka MAD

Comparison between two outcomes

Eg if group A had weight loss of 1kg

Group B had weight loss of 0.9kg

Absolute difference is 0.1kg

What would be the line of no significance (eg what would the confidence intervals have to cross) for a MAD to be non-significant?

0 - if the CI cross 0 this shows there is no difference between some of the two groups

How to interpret Forrest Plot?

- Dot/square is odds ratio usually
- Lines extending out are confidence interval
- Line down centre is line of no significance (eg 1 if ratio, 0 if MAD)
- Diamond at base is sum of all data and width is confidence interval