Using evidence in practice and Basic Statistics

1
Using evidence in practice and Basic Statistics

• Nick Price 21st Oct 2008

2
Objectives

• To have an overall understanding of the concept
  of evidence based practice and its application
• To be able to apply simple statistics in practice
• To develop some of the skills in this area
  required to pass the AKT

3
Schedule

• 1400-1500 Lecture
• Types of evidence
• Hierarchy of evidence
• Practical considerations
• Some simple statistics
• 1500-1515 Tea
• 1515-1615 Doing some numbers
• 1615-1645 Summary and evaluation (plenary)

4
What is evidence based practice?
5
Evidence based medicine?

• Maybe it is the application of the best available
  scientific knowledge to the patients or
  populations clinical problem?
• Or
• Evidence-based medicine is the integration of
  best research evidence with clinical expertise
  and patient values Sackett et al (2000)

6
(No Transcript)
7
A model of influences inclinical decision making

• Science Social Science
• Scientific Method
• Counting differences
• Understanding differences
• Technical judgement

• Arts and humanities
• Ethics
• Adaptability, broad perspective
• Humane Judgement

Clinical Judgement
8
Principles of a health care system
Systematic review Meta-analysis

• equity
• effectiveness
• efficiency
• appropriateness
• accessibility
• responsibility
• humanity

RCT Cohort studies Case control
studies Descriptive studies
quantitative
qualitative
9
Your practice is considering reviewing the
treatment for patients with moderate depressive
illness.How would you gather evidence?
10
First define your question

• What Population
• What Intervention(s)
• What Condition
• What Outcome

11
Levels of evidence

• (I-1) a well done systematic review of 2 or more
  RCTs
• (I-2) a RCT
• (II-1) a cohort study
• (II-2) a case-control study
• (II-3) a dramatic uncontrolled experiment
• (III) respected authorities, expert committees,
  etc..
• (Good Old Boys Sitting Heroically At Tables?)
• (IV)....someone once told me...

12
Gathering evidence

• Thorough
• Practical

• Medline search
• SR and Meta-analyses
• RCTs
• Review articles
• Observational Studies
• Clinical Evidence
• Cochrane
• DTB / MeReC
• RCGP/RCPsych guidelines
• NICE guidelines
• NSF
• Local Guidelines
• Bandolier
• Text books / GP notebook etc
• Ask the consultant or others?

13
Mini Test

• You are a member of a health care organisation
  commissioning committee. You must decide which of
  these four different cardiac rehabilitation
  programmes (for people who have suffered a heart
  attack) you would prefer to fund. Each programme
  costs the same to implement
• Programme A reduced the rate of deaths by 20
• Programme B produced an absolute reduction in
  deaths of 3
• Programme C increased patients' survival rate
  from 84 to 87
• Programme D meant that 33 people needed to enter
  the programme to avoid one death.
• Which would you choose?

14
Terms used to describe differences (in the
results)

• Relative risk (risk ratio)
• What is the chance of the endpoint occurring in
  the treatment group compared with the control
  group? EER / CER
• Absolute risk reduction (risk difference)
• How many fewer patients experienced the endpoint
  in the treatment group than in the control
  group? CER - EER
• Relative risk reduction
• By how much did treatment reduce the chance of
  the endpoint occurring in the treatment group
  compared with the control group? CER EER / CER
• Number needed to treat
• How many people, on average, do we need to treat
  for one of them to benefit? 100 / ARR
• Odds ratio
• Hazard ratio

15
Risks and benefitsSome basic numbers

• 20 additional protection from atherothrombotic
  events with Plavix (clopidogrel) over and above
  standard therapy (including aspirin).
• Proven Safety and tolerability profile.
• Only 1 increase in major bleeding.
• No statistically significant increase in life
  threatening or fatal bleeds when used with
  aspirin.
• Would you prescribe?
• How are they presenting benefit?
• How are they presenting risk?

16
The risks of aspirin with and without Clopidogrel

• Clopidogrel and Aspirin
• Major bleed 3.7
• Life threatening bleed 2.2
• Minor bleed 5.1

• Aspirin
• 2.7
• 1.8
• 2.4
• So what are the relative risks here?

17
The risks of aspirin with and without Clopidogrel
18
So they could have said

• 112.5 additional risk of minor bleeding with
  Plavix (clopidogrel) over and above standard
  therapy (including aspirin).
• Proven unsafe and intolerability profile.
• Only 37 increase in major bleeding and 22
  increase in life threatening bleeds but this is
  not statistically significant.
• Would you prescribe?

19
So lies damn lies and statistics?

• Absolute risk / benefits often sound small
• Relative risk / benefits can sound big
• e.g.
• Your chance of winning the lottery with 2 tickets
  as opposed to one is increased by 1 in 14million
• Your chance of winning the lottery with 2 tickets
  as opposed to one is increased by 100

20
Numbers Needed to Treat

• The average number of people from a defined
  population you would need to treat with a
  specific intervention for a given period of time
  to achieve one beneficial outcome.
• NNT 1 / ARR
• Can you calculate this for winning the lottery
  jackpot?
• Assume chance of winning is 1 in 14m per ticket
  per draw.

21
NNTs 2

• ARI 1/14m per ticket per draw
• 1/14m x 2 per week with 1 ticket for each
• draw
• 1/14m x2 x52 x5 for 5 years
• 520/14m
• NNT 14m/520 27,000 approx
• i.e. 27,000 people have to buy 2 tickets a week
  for 5 years for one person to win the jackpot

22
Summary

• Trials are there to inform clinical decision
  making.
• Identifying appropriate literature needs to
  consider quality / rigour but also practicality.
• Some stats are really quite easy but need to be
  used with care if you are to be honest.

23
Some reading / resources

• Greenhalgh T (2000) How to Read a Paper The
  basics of evidence based medicine. London BMJ
  Publishing
• McGovern D P B, Summerskill W S M, McManus R
  (2001) Evidence-based medicine in General
  Practice. Oxford BIOS Scientific. OR
• Sackett DL, Strauss S, Richardson WS, Rosenberg
  W, Hayes RB (2000) Evidence-based Medicine How
  to practice and Teach EBM. Edinburgh Churchill
  Livinstone

24

• Apples were 3 a bag, now only 2 a bag
• Amount saved is 1 per bag (Original rate new
  rate)
• Saving is one third or 33. (original rate new
  rate / original rate i.e. 3-2 1, 1/3 one
  third, 1/3 x 100 33

25

• Apples were 3 a bag, now only 2 a bag
• Amount saved is 1 per bag (Original rate new
  rate)
• Saving is one third or 33. (original rate new
  rate / original rate i.e. 3-2 1, 1/3 one
  third, 1/3 x 100 33

• Apples 30p a bag, now 20p a bag
• Saving is 10p a bag
• Saving is STILL one third

Ratio of new price to old price? 23 or 2/3 0.66
26
Now in pictures..
40
30
27
Now in pictures..
40
30
10
7.5
28
Describing differences in clinical studies

• In the first trial, 40 of people taking the
  control died. Only 30 of people taking the
  treatment died.
• The difference is 10.
• Control rate experimental rate 40 - 30
  10
• Absolute risk reduction (ARR) or risk difference
• The difference is 25.
• Control rate experimental rate / control rate
• 40 - 30 10 / 40 1/4 25
• Relative risk reduction (RRR)
• 3. The relative risk is EER / CER 30 / 40
  0.75

29
Describing differences in clinical studies

• In the second trial, 10 of people taking the
  control died. Only 7.5 of people taking the
  treatment died.
• The difference is 2.5.
• Control rate experimental rate 10 - 7.5
  2.5
• Absolute risk reduction (ARR) or risk difference
• The difference is 25.
• Control rate experimental rate / control rate
• (10 - 7.5)/10 2.5 / 10 1/4 25
• Relative risk reduction (RRR)
• 3. Relative Risk EER / CER 7.5 / 10 0.75

30
Numbers needed to treat (NNT)

• In the first trial, 40 of people taking the
  control died. Only 30 of people taking the
  treatment died.
• ARR 10
• RRR 25
• Number needed to treat (NNT)
• the number of people who need to take the
  treatment rather than the control for one to
  benefit
• NNT100/ARR()
• In this case 100/10 10
• For every 10 people who takes the treatment, one
  benefits who wouldnt have done had they all
  taken control.

31
Same medicines, different people

• In the second trial, 10 of people taking the
  control died. Only 7.5 of people taking the
  treatment died.
• The absolute risk reduction (ARR) is 2.5
• Control rate experimental rate 10 - 7.5
  2.5
• The relative risk reduction (RRR) is still 25
• Control rate experimental rate / control rate
• 10 - 7.5 2.5 / 10 1/4 25
• The number needed to treat is now 40
• 100/ARR
• 100/2.5 40

32
Same medicines, different people again

• In RCT 1
• 40 of people taking the control died
• Only 30 of people taking the treatment died
• ARR 10, RRR 25, NNT 10

• In RCT 2
• 10 of people taking the control died
• Only 7.5 of people taking the treatment died
• ARR 2.5, RRR 25, NNT 40

In the higher risk population, we would only need
to treat 10 people to save one.
But in the lower risk population we would need to
treat 40 people to save one.
33
(No Transcript)