Meta Analysis

1
Meta Analysis

• OT 653
• Fall 2003

2
Meta-Analysis

• Definition
• A specific subset of systematic reviews that
  statistically combine data from many studies in
  order to find a common effect(p.117).

3
Meta-Analysis

• Definition
• Meta-analysis refers to the analysis of
  analysesthe statistical analysis of a large
  collection of analysis results from individual
  studies for the purpose of integrating findings
  (Glass, 1976, as cited in Law, 2002, p.117).

4
Meta-Analysis

• Definition
• Meta-analysis is a name that is given to any
  review article in which the results of several
  independent studies are combined statistically to
  produce a single estimate of the effect of a
  particular intervention or health care situation
  (Jadad, 1998, as cited in Law, 2002, p.118).

5
Meta-Analysis

• A type of systematic review
• Uses statistical summary
• Synthesizes findings from several studies on the
  same topic.

6
Meta-Analysis

• Select a proper outcome measure
• Choose a variable in the data is present in all
  studies and
• Is a good indicator for the study
• Manipulate the data from several studies to
  organize the data in a unified fashion
• Choose a standard or measure
• Format data so it fits within that standard of
  measure
• Dialogue with researchers
• Diagram a Forest Plot

7
What is a Forest Plot?

• A forest plot is a vertical line with a number
  of horizontal lines running across it. it
  represents the pooled odds ratio of all the
  studies in the review (p.118).

8
Odds-ratios are primarily used in epidemiological
studies cohort and case-comparison studies
9
An odds ratio is used to estimate the odds of
membership in an identified group (i.e. persons
with stroke) from a sample given the presence of
specific independent variables
10
Odds-ratio values of less than one means an
individual is not in the target group (persons
with stroke) A value of 1 means the individual
could be in either group(with or without
stroke)A value of gt1 means the individual is in
the target group (is a person with a stroke).
11
Forest Plot

• Lines show the confidence Interval - 95
  certainty of the effects of a study
• Shorter the line, the more certain the results
• Diamond combined results
• Shows the treatment in question has had an effect
  because the diamond is to the left of the
  vertical line

Attempted reproduction of the Cochrane
Collaboration Logo an example of a Forest Plot
12
Remember confidence intervals???
13
And just in case you forgot confidence
intervals..
Confidence intervals are a range of possible
values in which the population mean is likely to
fall. A 95 confidence interval indicates that
the true mean of whatever you are trying to
measure is likely to fall in this range 95 of
the time.
14
(No Transcript)
15
Cochrane Collaboration

• Named for Archie Cochrane a British
  epidemiologist
• He believed that systematic reviews were
  important and necessary to help stay on top of
  the rapid changes in health care and to change
  health care practice for the better.
• He also believed that they should be organized
  and distributed to make sure knowledge about the
  best health care is readily available.
• Cochrane Collaboration, named after him, does
  systematic reviews using meta-analysis.

16
Questions and Answers

• ?

17
So how do systematic reviews differ from
metaanalyses??
18
Systematic Analysis

• Usually carried out by 2 or more people
• Criteria for study inclusion are decided either
  in advance, or after doing a search and seeing
  what evidence exists.
• Stronger systematic reviews are those carried out
  on RCTs, but this is not always possible

19
More on systematic reviews

• Once types of research designs to be included in
  the review are agreed upon
• Reviewers analyze articles independently, scoring
  them on whether or not they include specific
  criteria

20
Some Specific Criteria

• Sampling strategies
• Sample size
• Treatment variables
• Outcome measures
• Duration of study/study phases
• Length of follow-up
• Significant or non-significant findings

21
Article Scoring

• Reviewers then compare scores
• Often, articles that do not meet a pre-set score
  (5 out of 10 scored criteria) are then discarded
• Articles where authors disagree on criteria
  scores are either discussed or given to an
  outside party to score

22
Remaining Articles

• The articles left in the analysis are then
  written up in a format which includes multiple
  tables which compare the scoring criteria across
  studies.
• A summary is then written which compares the
  articles across criteria for strengths and
  weaknesses

23
A final summary draws conclusions about the kinds
of evidence and the strength of that evidence to
support the topic under review.
24
Critically Appraised Topics (CATs)

• OT 675
• Fall 2003

25
Definition

• A Critically Appraised Topic (CAT) is a one- or
  two page summary of a search and critical
  appraisal of the literature related to a focused
  clinical question, which should be kept in an
  easily accessible place so that it can be used to
  help make clinical decisions (Center For
  Evidence-Based Medicine, n.d.a., as cited in Law,
  2002, p.186).

26
Definition

• A CAT is a one- or two page summary of all the
  preceding steps involved in your evidence-based
  approach to the literature,. It provides
  immediate access to your method and results
  (Center For Evidence-Based Medicine, n.d.b., as
  cited in Law, 2002, p.186).

27
CATs and the Steps of the Evidence-Based Practice
Process

• To do evidence-based practice you need to
• 1. Have a focused question
• 2. Categorize the evidence found
• 3. Allow for evaluation of that evidence
• 4. Produce a clinical bottom line to be developed
  into practice (Law, 2002, p.186)
• Note the above 4 steps are part of the CAT
  process
• 5. Review CATs regularly and analyze successes
  further.

28
Five Major Types of CATs

• 1. Diagnosis
• 2. Prognosis
• 3. Evaluating risk and harm in a case-control
  study
• 4. Evaluating risk and harm in a cohort study
• 5. Intervention Studies (treatment, prevention,
  and screening). (Law, 2002, p.186).

29
Six Necessary Elements of CATs

• 1. Date of completion (of the CAT)
• 2. Question
• The person or problem being addressed
• The intervention or exposure being considered
• The comparison of the intervention or exposure,
  when relevant
• The outcomes of interest.
• 3. Clinical Bottom Line (CAT summary should
  include applicability of results to your client)
• 4. Evidence (CAT summary should include a summary
  of evidence)
• 5. Gold Standard (For Diagnosis or Screening -
  compare to best test out there for Risk and Harm
  - compare to existing treatments.
• 6. Notes (important issues, your reflections).

30
Drawbacks of CATs

• 1. Subject to bias, error, and other limitations
• 2. Become dated fairly quickly as new research is
  published

31
Strengths of CATs

• 1. Brief
• 2. Informative
• 3. Useful
• 4. Quick
• 5. Easy
• 6. Intuitive tool
• 7. Can use old CATs as a starting point when
  looking again at a clinical question.

32
Questions and Answers

• ?