Grading the quality of evidence and the strength of recommendations

1
Grading the quality of evidence and the strength
of recommendations

• Equity

2
Equity

• For disadvantaged groups are there different
• Effects?
• Values?
• Thresholds for costs?
• Distributions of disease?
• Settings?
• Baseline risks?

3
Different effects

• Quality across studies for specific outcomes
• Are there different effects?
• Overall quality of evidence across outcomes
• Balance of benefits and harms
• Balance of net benefits and costs
• Recommendations

4
1. Are there different effects for disadvantaged
groups?

• If there are inconsistent results that can be
  explained by social inequalities
• Different balance sheets for different groups
• Tobacco taxation
• If there is uncertainty about the directness of
  evidence for disadvantaged groups
• Lower quality of evidence
• Pneumococcal vaccine

5
Different values

• Quality across studies for specific outcomes
• Overall quality of evidence across outcomes
• Are there different values?
• Balance of benefits and harms
• Are there different values?
• Balance of net benefits and costs
• Recommendations

6
2. Are there differences in the relative
importance of the health outcomes for
disadvantaged groups?

• If an outcome is critical for a disadvantaged
  group and not for others
• This may change the overall quality of evidence
• If evidence regarding return to work is critical
  for disadvantaged groups and there is lower
  quality evidence for this outcome, the overall
  quality of evidence might be lowered
• If there are differences in the relative
  importance of health outcomes
• This may change the balance of benefits and harms
• Giving more weight to returning more quickly to
  work may shift the balance in favour of ARVs for
  HIV

7
Different thresholds for costs or distributions
of disease

• Quality across studies for specific outcomes
• Overall quality of evidence across outcomes
• Balance of benefits and harms
• Balance of net benefits and costs
• Are there different thresholds for costs?
• Are there differences in the distribution of
  disease?
• Recommendations

8
3. Is the cost of an intervention high for
disadvantaged groups?

• If the cost of an intervention is high for
  disadvantaged groups
• This may change the balance of net benefits and
  costs
• Cholesterol screening

9
4. Are there differences in the distribution of
the burden of illness that might affect
judgements about whether the intervention is
worth the costs?

• Differences in the distribution of the burden of
  illness
• May change the balance of net benefits and costs
• It may be more important to use resources for HIV
  than for screening for colorectal cancer

10
Differences in settings or baseline risks

• Quality across studies for specific outcomes
• Overall quality of evidence across outcomes
• Balance of benefits and harms
• Balance of net benefits and costs
• Recommendations
• Are there differences in settings?
• Are there different baseline risks?

11
5. Are there factors that would affect the
translation of the evidence into practice in
disadvantaged settings?

• If there are important factors that could be
  expected to modify the size of the expected
  effects, such as proximity to a hospital or
  available expertise
• This may change a recommendation
• Magnesium sulphate

12
6. Are there differences in baseline risk for
disadvantaged groups?

• If disadvantaged groups are at higher risk
• This will increase the benefits
• Penicillin for strept throat to prevent Rheumatic
  fever
• If there is uncertainty about the baseline risk
  for disadvantaged groups
• This may change a recommendation
• BCG for tuberculosis

13
How do judgements about equity affect
recommendations?

• For disadvantaged groups different
• Effects
• Inconsistant results (tobacco taxation)
  explicit value judgment
• Uncertainty about directness (pneumococcal
  vaccine) recommend evaluation
• Values
• Different values (ARV) different recommendation
• Thresholds for costs
• Different thresholds (cholesterol screening)
  different recommendation
• Distributions of disease
• Different distribution (colorectal cancer
  screening) different priorities
• Settings
• Different settings (magnesium sulphate)
  different recommendations
• Baseline risks
• Different baseline risks (BCG) different
  recommendations
• People may choose to recommend interventions that
  are less cost-effective because of equity

14
Equity should also be considered in formulating
questions, implementing recommendations and
evaluating the implementation of recommendations

• Formulating and prioratising questions
• Quality of studies for specific outcomes
• Quality across studies for specific outcomes
• Overall quality of evidence across outcomes
• Balance of benefits and harms
• Balance of net benefits and costs
• Recommendations
• Implementing recommendations
• Evaluation