Human Genome Epidemiology Network

1
Human Genome Epidemiology Network
2
Capture and appraise the evidence on the evolving
'big picture' across whole fields

• Develop widely accepted criteria for appraising
  evidence
• Causal inference

3
Assessment of cumulative evidence on genetic
associations
4
Content 4d

• DISCUSSION
• Key results
• Limitations
• Generalizability
• Interpretation lt gene dose-response effect
  consideration along with other studies data on
  functional effects of polymorphism and
  consideration as to relevance to putative genetic
  association
• APPENDIX
• Funding

5
Assessment of cumulative evidence on genetic
associations

• overlap with assessment of single studies
• schemes differ from one another

6
IARC Evaluations of Carcinogenic Risk to Humans

• human evidence, animal experiments evaluated
  separately classified as
• Sufficient
• Limited
• Inadequate
• Evidence suggesting lack of carcinogenicity
• strength of evidence that any carcinogenic effect
  observed is due to a particular mechanism
  assessed as weak, moderate or strong assessment
  if that particular mechanism likely to be
  operative in humans

7
IARC Evaluations of Carcinogenic Risk to Humans

• Group 1 The agent (mixture) is carcinogenic to
  humans.
• The exposure circumstance entails exposures that
  are carcinogenic to humans.
• Group 2A The agent (mixture) is probably
  carcinogenic to humans.
• The exposure circumstance entails exposures that
  are probably carcinogenic to humans.
• Group 2B The agent (mixture) is possibly
  carcinogenic to humans.
• The exposure circumstance entails exposures that
  are possibly carcinogenic to humans.
• Group 3 The agent (mixture or exposure
  circumstance) is not classifiable as to its
  carcinogenicity to humans.
• Group 4 The agent (mixture) is probably not
  carcinogenic to humans

8
Assessment of cumulative evidence on which
healthcare recommendations based
Atkins et al., 2004

• 6 systems
• prominent
• included features not captured by other prominent
  systems
• assessed by 12 assessors

9
Assessment of cumulative evidence on which
healthcare recommendations based

• American College of Chest Physicians (ACCP)
• Australian National Health and Medical Research
  Council (NHMRC)
• Oxford Centre for Evidence-Based Medicine (OCEBM)
• Scottish Intercollegiate Guidelines Network
  (SIGN)
• US Preventive Services Task Force (USPSTF)
• US Task Force on Community Preventive Services
  (USTFCPS)

10
Assessment of cumulative evidence on which
healthcare recommendations based

• Poor agreement
• Stimulated development of new assessment system

11
GRADE Evidence profile

• study design (RCT/observational)
• study quality (serious limitations Y/N)
• consistency
• directness, the extent to which the people,
  interventions, and outcome measures are similar
  to those of interest. (uncertainty Y/N)

12
GRADE Evidence profile

• Type of evidence
• Randomised trial high
• Observational study low
• Any other evidence very low

Decrease grade if Serious ( - 1) or very
serious ( - 2) limitation to study quality
Important inconsistency ( - 1) Some ( - 1) or
major ( - 2) uncertainty about directness
Imprecise or sparse data ( - 1) High
probability of reporting bias ( - 1)
Increase grade if Strong evidence of
associationsignificant relative risk of gt 2 ( lt
0.5) based on consistent evidence from 2 or more
observational studies, with no plausible
confounders (1) Very strong evidence of
associationsignificant relative risk of gt 5 ( lt
0.2) based on direct evidence with no major
threats to validity (2) Evidence of a dose
response gradient (1) All plausible
confounders would have reduced the effect (1)
13
Hill - 9 issues to be considered when judging
whether an observed association might be a causal
relationship
Experimentation
Cancer Epidemiology Consistency Strength Dose-res
ponse Biological plausibility
Epidemiologic methods Strength Temporality Consis
tency Biological plausibility Dose-response
Specificity
Coherence
Analogy
Potischman Weed 1999
Weed Gorelic 1996
14
SIGN Levels of evidence
15
SIGN Levels of evidence