samples, ratios, risk, confounding stuffff Flashcards
what 3 mains things do we want our sample population to be?
- representative
- ubiased
- precise
what 2 errors can happen?
- chance (random) - this is due to sampling variation - it reduces as sample size increases
- bias (systematic) - the difference between the true and expected value - doesnt chnage with sample size
what does increasing the sample size do?
reduces the random errors
increases precision
reduces uncertainty
you want a precise and non bias study
what are examples of bias/systematic error?
selection biases:
information biases
- study sample - study sample is not representative (external validity)
- group selction within a study - groups are not comparable (internal validity)
- healthy worker effect - woerkers have a lower mortality rate
- recall error - difference in recollection from participants regarding past events
- observer/interviewer error - preconceieved expectations that influences result
- meaurement error - different tools or rounding when measuring participants
- misclassification - person is put in the wrong group - rises from meaurement or obsverer error
what is prevalence?
equation?
the proprtion of people with a disease at a given time (new and existing cases)
is a snap shot at a given time
decribes burden of disease - how many have it - funding allocation for areas in NHS
number of people with disease/total population
what is incidence?
equation
how to caluclate total patient time at risk?
the number of NEW cases in a given time frame e.g. month
focuses on new cases only
useful for monitoring disease spread e.g. epidemic
it is a rate
number of new cases/patient time at risk
no. patients X time they were at risk
what is incidence rate ratio?
equation?
what is an example of what its used for?
compares the incidence rate in one group to another - relative eaurement between 2 groups
IRR a/ IRR b
group of intertest = numerator
mortality rates at differnent hospitals
what are odds?
equation?
what number woulf you get if the probabilitys are even?
the probability of the even happening over the probalilty of it not happening
p = probability of it happening
P/ 1-P
1
what is the odds ratio?
equtation?
comparing the odds of 2 different groups - relative comparision
odds of A/ odds of B
odds is a ratio therefore the odds ratio is a ratio of a ratio
what is absolute risk?
equation
what is relative risk?
the risk of disease happneing to you - its a proportion
diseased/ diseased +nondiseased
its a ratio of proportions - it compares risk of 2 groups
risk of A/ risk of B
what does it mean if the odds ratio is less than 1?
what does it mean if the risk ratio is less than 1?
when is this desireable?
what does it mean if the odds ratio is more than 1?
what does it mean if the risk ratio is more than 1?
when is this desireable?
it means the numerator has a lower odds of the event
numerator has a lower risk of the event
if the event is a bad outcome e.g. death or disease
the numerator has a greater odds of the event
the numerator has greater risk of the event
when the event/outcome is good e.g. quitting smoking
what is risk difference?
how is to calculated?
it is the absolute difference in risk between 2 groups
absolute risk A - absolute risk B
what is confounding ?
how do we adjust for confounding?
a confounder is a variable that influences both the dependent variable and independent variable, causing a spurious association - links the relationship the interest
standardisation
howeerve a lot ar eunknown therefore is difficult to address
