Implementing GRADE: calculating the risk difference from the baseline risk and the relative risk

Area of Science:

• Medical decision making
• Biostatistics
• Health economics

Background:

• The Grading of Recommendations Assessment, Development and Evaluation (GRADE) system is crucial for evidence-based practice.
• Accurate estimation of risk differences is essential for implementing GRADE recommendations.
• Existing methods may not fully account for all sources of uncertainty in risk difference calculations.

Purpose of the Study:

• To present a simple and effective method for calculating confidence intervals (CIs) for risk differences.
• To address the specific situation where baseline risk is estimated from independent sources.
• To ensure CIs for absolute risk differences account for all relevant uncertainties.

Main Methods:

• Described a method for calculating confidence intervals (CIs) for risk differences.
• Emphasized the use of independent sources for estimating baseline risk.
• Incorporated all sources of uncertainty into the CI calculation for absolute risk differences.

Main Results:

• A straightforward and effective method for computing CIs for risk differences was developed.
• The method allows for the incorporation of baseline risk estimates from independent sources.
• Confidence intervals for absolute risk differences are calculated considering all uncertainties.

Conclusions:

• The proposed method provides a reliable way to calculate confidence intervals for risk differences within the GRADE framework.
• Utilizing independent sources for baseline risk estimation enhances the accuracy of risk difference calculations.
• Accounting for all uncertainties is vital for robust GRADE implementation and informed clinical decision-making.