Section 2 of the AKT content guide Covers research statistics and epidemiology 10 of the exam This element is designed to examine the candidates ability to use evidence and data to underpin clinical decision making and the possession of critical appraisal skills sufficient to recognise ID: 932818
Download Presentation The PPT/PDF document "Some Statistics for the AKT" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Some Statistics for the AKT
Slide2Section 2 of the AKT content guide
Covers research, statistics and epidemiology.
10% of the exam
‘This element is designed to examine the candidate’s ability to use evidence and data to underpin clinical decision making, and the possession of critical appraisal skills sufficient to recognise good evidence and adopt guidelines as appropriate’
Slide3What we are going to cover:
Some statistical terminology (not all of it )
Calculations for evidence based practice
Graphical representations
General advice fo
r the AKT
Slide4The contingency table
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide5The null hypothesis
No relationship between two measured phenomena.
Rejecting the null hypothesis leads to the conclusion that there is a relationship between the phenomena.
Slide6Type 1 vs Type 2 error
Slide7Sensitivity
Measures the proportion of positive patients that are correctly identified as such.
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide8Sensitivity
How would you calculate it?
Sensitivity =
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide9Specificity
Measures the proportion of negative patients that are correctly identified as such.
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide10Specificity
How would you calculate it?
Specificity =
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide11Positive predictive and negative predictive values
Proportions of positive and negative values that are true positives and true negatives.
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide12PPV and NPV
How would you calculate these?
PPV =
NPV =
True condition
Test condition
Positive
Negative
Positive
True positive (TP)
False positive
(FP)
Type 1 error
Negative
False negative
(FN)
Type 2 error
True negative
(TN)
Slide13Example
400 plants are selected from a field. 75 of them have round leaves, the rest have pointy leaves.
Scientists have developed a test to determine if a plant has round leaves or pointed leaves. 60 plants with round leaves were round positive. 20 plants with pointy leaves were round positive as well.
Calculate the sensitivity, specificity, PPV and NPV of the test.
Slide14Step 1
Leaves
Test
Round
Pointy
Round positive
Round negative
Slide15Step 1
Leaves
Test
Round
Pointy
Round positive
60
15
Round negative
20
305
Slide16Step 2
Plug in the numbers
Sensitivity =
=
=
Specificity =
=
=
= 95.3%
Leaves
Test
Round
Pointy
Round positive
60
15
Round negative
20
305
Slide17Step 2
PPV =
=
=
= 80%
NPV =
=
=
= 93.8%
Leaves
Test
Round
Pointy
Round positive
60
15
Round negative
20
305
Slide18Evidence Based Practice
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide19Absolute Risk Reduction
The change in the risk of an outcome of a given treatment/activity in relation to a comparison treatment/activity.
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide20Absolute Risk Reduction
ARR = EER-CER
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER=
CER=
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
Slide21Number Needed to Treat
The average number of patients who need to be treated to prevent one additional bad outcome
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide22Number Needed to Treat
NNT =
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide23Number Needed to Treat
For example a drug will treat a disease. P
A
is the probability the drug will treat the disease. P
B
is the probability the group still have the disease.
Slide24Number Needed to Treat
Description
P
A
P
B
NNT
Interpretation
Perfect drug
0.0
1.0
1.0
Everybody is cured with the pill; nobody without
Very good drug
0.1
0.9
1.25
Ten take the pill; 8 cured by the pill, 1 cured by itself, 1 still sick.
Satisfactory drug
0.3
0.7
2.5
Ten take the pill; 4 cured by the pill, 3 cured by itself, 3 still sick.
High placebo effect
0.4
0.5
10
Ten take the pill; 6 cured but 5 of those would be cured anyway.
Low cure rate
0.8
0.9
10
Ten take the pill, one is cured by the pill, one cured by itself, 8 still have the disease.
Goes away by itself
0.1
0.2
10
Ten take the pill and 9 are cured; but 8 would have been cured anyway.
Sabotages cure
0.9
0.8
−10
Ten take the pill, two would have been cured without it, but with the pill, only one is cured, so NNH=10.
Slide25Relative Risk
The probability of an event occurring in an exposed group to the probability occurring in a non exposed group.
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide26Relative Risk
Relative Risk =
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide27Odds Ratios
A measure of association between an exposure and an outcome
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide28Odds Ratios
OR=
Experimental group
(E)
Control
Group
(C)
Events
(E)
EE
CE
Non events
(N)
EN
CN
Total subjects
(S)
ES
CS
Event rate (ER)
EER
CER
Slide29Example
200 rabbits are randomly allocated to two groups. They have all been exposed to a virus. 100 of the rabbits are given the treatment . 95 rabbits are cured. Of the 100 rabbits given a placebo, 10 are cured.
Calculate the ARR, NNT, Relative risk and odds ratios.
Slide30Step 1
Experimental group
(E)
Control
Group
(C)
Events
(E)
Non events
(N)
Total subjects
(S)
Event rate (ER)
Slide31Step 1
Experimental group
(E)
Control
Group
(C)
Events
(E)
95
10
Non events
(N)
5
90
Total subjects
(S)
100
100
Event rate (ER)
0.95
0.1
Slide32Step 2
How would you calculate ARR?
ARR = EER-CER = 0.85=85%
Experimental group
(E)
Control
Group
(C)
Events
(E)
95
10
Non events
(N)
5
90
Total subjects
(S)
100
100
Event rate (ER)
0.95
0.1
Slide33Step 2
How would you calculate NNT?
NNT =
=
Experimental group
(E)
Control
Group
(C)
Events
(E)
95
10
Non events
(N)
5
90
Total subjects
(S)
100
100
Event rate (ER)
0.95
0.1
Slide34Step 2
How would you calculate relative risk?
Relative Risk =
=
Experimental group
(E)
Control
Group
(C)
Events
(E)
95
10
Non events
(N)
5
90
Total subjects
(S)
100
100
Event rate (ER)
0.95
0.1
Slide35Step 2
How would you calculate odds ratios?
OR=
=
=171
Experimental group
(E)
Control
Group
(C)
Events
(E)
95
10
Non events
(N)
5
90
Total subjects
(S)
100
100
Event rate (ER)
0.95
0.1
Slide36Graphical Representations
Slide37Normal distribution
Slide38Skewed distribution
Slide39Scatter diagrams
Slide40Box plots
Slide41Forest plots
Slide42Funnel plots
http://www.bmj.com/content/343/bmj.d4002
Slide43Cates diagrams
Slide44Kaplan Meier curves
Slide45Conclusion
We covered some terminology and calculations involved in the AKT
We also covered some diagrams and the interpretation of these diagrams
Slide46What else should I do?
Go over section 2 of the AKT content guide.
Do lots of practice questions.
Learn the statistical terminology (59+ terms)
Understand the principles of screening
Slide47What about the administration section?
Oxford Handbook of GP first few chapters tells you most of what you need to know about section 3
.
Slide48Learn (this is not an exhaustive list):
DVLA
fitness to drive
guidelines,
CAA
fitness to fly
guidelines,
Childhood
imms
schedule,
School
exclusion
guidelines,
Child
development
milestones,
UKMEC guidelines,Emergency management,
Consultation models,
Any relevant CKS/NICE guidelines of common conditions.
Slide49Questions?
Ideas?
Concerns??
Expectations???