From GRADE Evidence to Decision Factors for Making Strong Versus Weak Recommendations
Factor Relevant Question Notes
Priority of problem Is the problem addressed by the question important enough to make a recommendation? Many problems may not be identified a priori as high enough importance to justify strong recommendations when weighed against other problems.
Balance of benefits and harms Across outcomes, are the overall effects and confidence in those effects a net gain? Most interventions, prognostications, and diagnostic tests have positive and negative consequences. Confidence in these estimates must be viewed in aggregate—do positive effects outweigh negative ones? Consideration must weigh outcomes by importance.
Certainty in the evidence What is the overall certainty that these estimates will support a recommendation? More certainty supports stronger recommendations, and vice versa.
Values and preferences To what extent do the values and preferences of patients regarding outcomes or interventions vary? Minimal variation and a strong endorsement of the outcomes or the interventions based on patients’ values and preferences supports stronger recommendations. The lack of consistency in patients’ values and preferences or a weak endorsement of the outcomes or the interventions supports weaker recommendations
Costs and resources Are these net results proportionate to the expenditures and demands of the recommended measure? Factors such as manpower, time, distraction from other tasks, and monetary investment are viewed through local values. Lower costs of an intervention and greater cost-effectiveness support strong recommendations, and vice versa. Analysis should account for uncertainty in the calculated costs.
Equity Are the net positive effects of the measure distributed justly? Measures that improve disparities or benefit fairly may drive a stronger recommendation, and vice versa.
Acceptability Across stakeholders, is the measure tractable? To be strong, a recommendation ideally appeals to most.
Feasibility Can the recommendation be implemented from a practical standpoint? Something that is practical to achieve may support a strong recommendation, and vice versa.
Summary: To what extent do positive and negative consequences balance in the settings in question?
  • Negative clearly outweighs positive
  • Negative probably outweighs positive
  • Negative and positive consequences balanced
  • Positive probably outweighs negative
  • Positive clearly outweighs negative
  • Strong recommendation against
  • Weak recommendation against
  • Weak recommendation for
  • Strong recommendation for
Considerations: Are there important subgroups that might be treated differently? Are there important concerns for implementation?