Transcript Stats for MRCGP

Statistics for nMRCGP
Jo Kirkcaldy
Curriculum Condensed Knowledge
•
•
•
•
•
•
•
•
Incidence and prevalence
Specificity and sensitivity
Positive and negative predictive values
Absolute risk and absolute risk reduction
Relative risk and relative risk reduction
NNT and NNH
Odds and odds ratio
Hazard ratios
Incidence and Prevalence
• Incidence
– The number of new cases of a disease/condition
occuring in a population over a given time period
• Prevalence
– The number of existing cases of a
disease/condition in a population at a specific
point in time
Specificity and sensitivity
• Relevance is in screening and diagnostic tests
• Sensitivity = true positive rate
– i.e. chance of a positive test result for people with
the disease
• Specificity = true negative rate
– i.e. chance of a negative test result for people
without the disease.
Sensitivity and specificity
Disease
No
disease
Total
Test +ve
a
b
a+b
Test -ve
c
d
c+d
Total
a+c
b+d
a+b+c+d
Sensitivity = a/a+c (x100)
Specificity = d/b+d (x100)
Positive and negative predictive values
• Measure screening performance (i.e post test
probability)
• Positive predictive value
– Chance of disease for person with positive test result
• Negative predictive value
– Chance of no disease for person with negative test
result
PPV and NPV
Disease
No
disease
Total
Test+ve
a
b
a+b
Test -ve
c
d
c+d
Total
a+c
b+d
a+b+c+d
• PPV = a/a+b
• NPV = d/c+d
Risk
• Used in prospective studies
• Absolute risk – chance of developing
disease/death over a set time period
• Relative risk - compares chance of developing
disease/death in two groups e.g. with and
without treatment over set time period
Deaths
Survival
Total
Control
32
67
99
New
drug
21
79
100
Total
53
146
199
• Absolute risk of death for
condition = 53/199 =
0.266 (x 100)
• Relative risk of death(risk
ratio) = 21/100/32/99 =
0.21/0.323 = 0.65
• Relative risk reduction =
1-0.65 = 0.35 = 35%
NNT and NNH
• Related to risk
• Calculate how many patients need to save one
life/harm one person
Deaths
Survival
Total
Control
32
67
99
New
drug
21
79
100
Total
53
146
199
• NNT = 1/risk difference
(=ARR)
• Risk difference = 32/9921/100 = 0.323-0.21 =
0.113
• NNT = 1/0.113 = 9
• Need to treat nine people
to save one life
Odds and odds ratio
Deaths
Survival
Total
Control
32
67
99
New
drug
21
79
100
Total
53
146
199
• Used in retrospective
studies
• Odds of death
– control = 32/99/67/99=
0.478
– drug = 21/100/79/100=
0.266
• Odds ratio =
0.266/0.478 = 0.55
Hazard ratios
• Reflects time survived to an invent i.e. survival
data.
• Complicated log calculation but essentially
calculates ratio of rates at which an event
happens.
• E.g if 1000 patients have an MI and 500 are
thrombolysed hazard ratios could be used to
describe the time to death in each group.
Curriculum Condensed Knowledge
•
•
•
•
•
•
•
•
Incidence and prevalence
Specificity and sensitivity
Positive and negative predictive values
Absolute risk and absolute risk reduction
Relative risk and relative risk reduction
NNT and NNH
Odds and odds ratio
Hazard ratios
Further useful references
• ‘How to read a paper’ Trish Greenhalgh –
pubilshed in serial form in BMJ or available as
a book.