Calculations Flashcards
Risk calculations
-when to use
-what can you find out
Cohort studies - to find out whether a specific exposure is a risk factor for developing an outcome
Experimental event rate
EER = outcome number in this group/everyone in this group
Control event rate
CER = outcome number in this group/everyone in this group
Relative risk ratio
EER/CER
1+ - the rate of the event is increased compared to the control
U1 - the rate of the event is decreased compared to the control
Relative risk change - how much does the risk change with a certain exposure
EER-CER / CER
Absolute risk change -
CER - EER
Number needed to treat
-meaning
-calculation
The number of patients who need to be treated to prevent the occurrence of 1 adverse event or promote the occurrence of 1 beneficial event
NNT = 1/ARR
Odds
-when to use
-what can you find out
Case control studies - to find out what the ratio between an exposure/non exposure and an outcome
Differs from probability
1/6 probability of rolling a 6
1/5 odds of rolling a 6
Total no of patients on paracetamol = 60
No of patients achieving pain control on paracetamol = 40
Odds of pain relief on paracetamol = 40/20 = 2
Total no of patients on placebo = 90
No of patients achieving pain control on placebo = 30
Odds of pain relief on placebo = 30/60 = 0.5
So the odds ratio of achieving pain control on paracetamol = 2/0.5 = 4
SD calculations
-variance vs SD
-SEM and CI
-SDs
Variance - spread of values away from the mean
Variance = SD^2
SEM = SD/square root of n
Lower limit of 95% CI = mean - (1.96 x SEM)
Upper limit of 95% CI = mean + (1.96 x SEM)
1SD = 68
2SD = 95
3SD = 99.7
Sensitivity
Specificity
Positive predictive value
Negative predictive value
Likelihood ratio +
Likelihood ratio -
Sensitivity
TP/All people with disease
Specificity
TN/All people without the disease
Positive predictive value
TP/All people with a positive result
Negative predictive value
TN/All people with a negative result
Likelihood ratio +
Sen/1-spec
Likelihood ratio -
1-sen/spec
