measures of association

Question 1

what are measures of strength

Answer 1

• risk ratio
• odds ratio

Question 2

what are measures of impact

Answer 2

• risk difference
• attributable fraction
• number needed to treat

Question 3

outcomes

Answer 3

disease, cure

Question 4

exposures

Answer 4

risk factors

Question 5

interventions

Answer 5

treatments

Question 6

vet research compares association between

Answer 6

outcomes and exposure or intervention

Question 7

odds

Answer 7

• ratio of one outcome to that of an alternative outcome
• calculated as number of outcomes divided by number on non outcomes
• 51 boys in 100 births, ODDS any birth is a boy is 51/49=1.04

Question 8

risk

Answer 8

• probability an outcome will happen
• calculated as number of outcomes divided by total number
• 51 boys in 100 births; RISK of having a boy is 51/100=0.51

Question 9

risk to odds

Answer 9

odds= risk/ (1-risk)

Question 10

odds to risk

Answer 10

risk= odds/ (odds+ 1)

Question 11

risk of disease in exposed and risk of disease in non exposed

Answer 11

just like normal thinking
either a/(a+b) or c/(c+d)

Question 12

odds of disease in exposed or non-exposed

Answer 12

either a/b
or c/d

Question 13

risk ratio

Answer 13

a/ (a+b) OVER c/(c+d)
Risk of exposed / risk of non exposed

ratio of the risk of disease in the exposed group to risk of disease in the unexposed group

Question 14

odds ratio

Answer 14

a/b OVER c/d

which = ad/bc

the ratio between the odds of disease of exposed to odds of disease in unexposed group

Question 15

RR interpretation

Answer 15

RR=1 then no exposure between exposure/ treatment and disease/cure

RR < 1 then exposure/ treatment reduces risk of disease/cure

RR>1 exposure/ treatment increases risk of disease/cure

Question 16

risk ratio

Answer 16

• provides an estimate of how many times more likely exposed/ treated animal experience disease/ cure compared w exposed/ treated animals
• no units, ranges from 0-infinity
• greater departure of RR from 1 the stronger/weaker the association
• RR cannot be estimated in case-control studies, as these studies do not allow calculation of risks
• recall its only relative; no indication to absolute risk, even if RR very high, even though overall disease risk is quite low

Question 17

odds ratio interpretation

Answer 17

OR= 1 no association between association between exposure/ treatment and disease/cure

OR<1 exposure/treatment reduces odds of disease/cure

OR> exposure/ treatment increases odds of disease/cure

Question 18

odds ratio

Answer 18

• provides an estimate of the odds of disease in the exposed/ treated group to the odds of disease in the unexposed/ not treated group
• OR can be calculated in any study design
• no units, ranges from 0 to infinity
• the greater departure of OR from 1 the stronger/weaker the association
• symmetrical

Question 19

when odds ratio and risk ratio are v different what does that mean

Answer 19

disease is common

if they are v close then disease is NOT common

Question 20

measure of strength

Answer 20

• risk ratio and odds ratio
• measure the extent to which risk of disease/ success differs between exposed/ treated and unexposed/ not treated groups
• they are useful for identifying risk factors, but do not provide info on abolsute risk

Question 21

measures of impact

Answer 21

• risk difference, attributable factor, number needed to treat
• express the impact (effect) of a risk factor on a disease/success
• quantify the consequences in terms of absolute risk resulting from exposure to a risk factor
• give an indication of the importance to the risk factor
• these are the measures most common and most useful for clinicians

Question 22

risk difference

Answer 22

• exactly what is sounds like, difference is risk of exposed minus risk of non exposed
• a/(a+b)- c/(c+d)
• reflects the increase or decrease in the disease/cure in the exposed/ treated group that is due to exposure/treatment
• describes the absolute quantity of outcome measure that is associated w the exposure/treatment
• clinical effects often expressed as RD

Question 23

number needed to treat

Answer 23

• 1/RD
• describes how many cases have to be treated to cure (or prevent) one additional case of the disease
• ideal is 1, where every animal improves/ is cured w new treatment and none improve/ cured with control
• so effective treatments have low NNT

Question 23

attributable factors

Answer 23

• is proportion of successes for for the new treatment that is due to the new treatment
• = RD/ risk in exposed group
  -= (RR-1)/ RR
• = (OR-1)/ OR
• percentage reduction or increase in risk of outcome in the exposed/ treated attributable to exposure/ treatment
• relative measure of risk reduction