measurement of association Flashcards
why do we measure risk and benefit:
-Describe the – of the population
-Identify – that characterise groups that at are increased or decreased risk
-Measure the — of these factors
-What would happen to the health of the — if this factor were to be eliminated ?
-What would happen to the health of the — if this factor were to be eliminated?
-These measures are known collectively as — measures
heath
factors
impact
individual
population
effect size
to understand ad communicate risk we need:
- — and – risk are both needed
- Absolute risk provides the —
-If the absolute risk is low, the chances of it happening are – , so it may not be —
-Relative risk is a key to the importance of a risk factor
But a – increase in a big risk may be more important than a – increase in a small risk
-Smoking raises lung cancer risk 9-fold, but absolute risk remains –
-It raises risk of heart attack by 2½ but the absolute risk is —
-Food for thought: risk of adverse reaction to vaccines vs risk of the disease itself
relative and absolute
context
low
big
not important
small
low
high
measuring the relative exposure effect:
1- pervelancr ratio is = — / — which shows how many of u would test +ve for covid rn
2- odds ratio is = — / — which is +/- in the room
3- risk ratio show = — / — which is how many of u would test +ve for covid this semester
- — often is the group not exposed to the risk factor
prevalence in exposed group/prevalence in baseline group
odds of disease in exposed groups/odds of disease in baseline group
incidence in exposed groups/ incidence in baseline group
baseline
- For any ratio, a value of 1 means that the risk/rate is the – in both groups, so the factor has – effect on risk
-If the value > 1, risk/rate is – in “exposed” group = the factor associated with – risk
Eg a ratio of 1.5 in Group C means that the rate is 50% higher
-If the value < 1, risk/rate is – in “exposed” group = the factor associated with – risk
-Note that ratio measures are —
-If smokers have a prevalence ratio of 2 compared with non-smokers, non-smokers have a prevalence ratio of 0.5 (½) compared with smokers.
All of this is true for any ratio measure - asthma is associated w socioeconomics factors and physical disability w depression
same
no effect
higher
higher
lower
lower
reversible
relative risk is based on comparing —- so we need – data to — a sample of ppl
incidence
cohort study
follow ups
example:
risk of dementia and death after delirium:
1- OBJECTIVES: to examine the relationship between an episode of delirium
and subsequent dementia and death over 3 years.
2- DESIGN: prospective cohort study.
3- SETTING: patients (n = 203) were aged 65 years or older at baseline and survivors of the index admission.
4- METHODS: We followed 38 inpatients diagnosed with delirium and 148 patients with no delirium for a median of 32.5 months.
5- RESULTS: The incidence of dementia was 6% per year over 3 years for those without delirium and 18% per year for those with delirium. The relative risk of dementia for those with delirium was 3.0 (95% confidence interval 1.86-5.63)
-Sometimes we are not interested in comparing numbers of people but numbers of — occurring in a period of time
-This is where an individual can experience the event – such as:
1- —- : Heart attacks, strokes
2- — : Episodes of diarrhoea, Covid or flu infections
3- – Injury
event
repeatedly
chronic
infectious
acute
- —- is the amount of risk (in a specified group of people) that can be attributed to the exposure
-We can never know whether an individual’s cancer was “due to exposure” – there is no way to examine a person and determine that they got cancer because they smoked
-However, with some — , it may be possible to calculate how many such people there are (but never who they are)
Attributable risk
assumptions
2 types of questions:
1- Questions about the — :
-What proportion of cases of disease in the — are attributable to the risk factor? Eg what proportion of low birth weight babies are attributable to maternal smoking, and
2- Questions about — with the risk factor
-What proportion of that — total risk is attributable to the risk factor
-In mothers who smoke, what proportion of their risk of having a low birth weight baby is attributable to their smoking?
population
population
people
persons
in attributable risk:
- — attributable risk:
percentage of the cases in the population which can be attributed to the risk factor,
-attributable risk in the – :
percentage of the cases among those exposed to the risk factor which can be attributed to the risk factor.
population
exposed
check slide 28
if we want to measure the outcome rate its — over the – by which if we want to know the risk difference we simply need to know the difference between 2 – which is referred to as —- but we need to introduce a number needed to treat to know if the rate was added or multiplied
remission / total
events
absolute risk reduction
NNT ( number needed to treat) is the number of people you need to treat to produce — extra — outcome compared with — .
-The treatment produces 26 remissions per 100 patients, while the control produces 13 per 100 patients, so there are 13 extra remissions per 100 patients.
If treating 100 people gets us 13 extra remissions, then treating (100/13) patients gets us one extra responder.
100/13 = 7.7 patients
More formally, the number needed to treat is 100/(Risk difference %) which is 100/(26%-13%).
As above, 100/13 =7.7 patients
Always round up – patients are whole people
one
positive
control
- relative risk RR = event rate in the — divided by the event rate in the —
-The event rates, the number needed to treat and the relative risk have — interpretations.
In this case :
The treatment increases the remission rate from 13% to 26%, an increase of 13%.
(Event rates and their difference)
The remission rate in the treated group is twice as high as in the control.
(Relative risk)
Treating 8 patients results in one extra remission
(NNT)
treated
control
intuitive
