Article type
Abstract
Background: Grading of Recommendations, Assessment, Development and Evaluations (GRADE) as a support system for health systems and public health decision-making is currently recognized and adopted by several guideline development bodies and organizations. However, the application status of GRADE approach in public health guidelines (PHGs) remains unclear.
Objective: To analyze the characteristics, hot spots, and gaps of GRADE used in the fields of public health, taking PHGs issued by the World Health Organization (WHO) as an example.
Methods: The WHO website was searched in March 2022 to identify PHGs applying GRADE. Researchers independently extracted and analyzed the data using Excel 2019, plotted radar plots using Origin 2020, and implemented keyword contribution matrix plots using VOS Viewer 1.6.1. Descriptive analyses were performed on the included PHGs, focusing on publication characteristics, topic characteristics, strength of recommendation, certainty of evidence, and characteristics of downgrading/upgrading domains of GRADE.
Results: 203 PHGs applied GRADE methods, reporting 8715 (91.19%) outcome indicators using the GRADE rating, which mainly focused on sex and reproductive health (24.14%). Only 31.69% of the outcomes reported moderate or high certainty of evidence. The main clustering results were dominated by adverse events. A total of 3635 recommendations were formed, of which 1939 (53.34%) were strong recommendations and 1278 (35.15%) were weak/conditional recommendations. Among these strong recommendations, 907 (54.44%) were based on low-confidence (30.19%) or very low-confidence (24.25%) estimates. For inconsistent recommendations, 119 recommendations fit one of the 5 paradigm scenarios. Among the 891 inappropriate recommendations, 316 recommendations were more suitable for weak/conditional recommendations. Around 41.8% of the recommendations were related to drug interventions. Among the 12,567 instances of downgrading, 41.44% were due to the risk of bias and 36.18% were due to imprecision. Only the magnitude of effect size (0.26%) was used as an upgraded factor.
Conclusion: GRADE is widely used in PHGs, but the quality and grading criteria of existing evidence may be confusing for its application in the field. More suitable grading methods for the characteristics of public health decision-making need to be explored.
Objective: To analyze the characteristics, hot spots, and gaps of GRADE used in the fields of public health, taking PHGs issued by the World Health Organization (WHO) as an example.
Methods: The WHO website was searched in March 2022 to identify PHGs applying GRADE. Researchers independently extracted and analyzed the data using Excel 2019, plotted radar plots using Origin 2020, and implemented keyword contribution matrix plots using VOS Viewer 1.6.1. Descriptive analyses were performed on the included PHGs, focusing on publication characteristics, topic characteristics, strength of recommendation, certainty of evidence, and characteristics of downgrading/upgrading domains of GRADE.
Results: 203 PHGs applied GRADE methods, reporting 8715 (91.19%) outcome indicators using the GRADE rating, which mainly focused on sex and reproductive health (24.14%). Only 31.69% of the outcomes reported moderate or high certainty of evidence. The main clustering results were dominated by adverse events. A total of 3635 recommendations were formed, of which 1939 (53.34%) were strong recommendations and 1278 (35.15%) were weak/conditional recommendations. Among these strong recommendations, 907 (54.44%) were based on low-confidence (30.19%) or very low-confidence (24.25%) estimates. For inconsistent recommendations, 119 recommendations fit one of the 5 paradigm scenarios. Among the 891 inappropriate recommendations, 316 recommendations were more suitable for weak/conditional recommendations. Around 41.8% of the recommendations were related to drug interventions. Among the 12,567 instances of downgrading, 41.44% were due to the risk of bias and 36.18% were due to imprecision. Only the magnitude of effect size (0.26%) was used as an upgraded factor.
Conclusion: GRADE is widely used in PHGs, but the quality and grading criteria of existing evidence may be confusing for its application in the field. More suitable grading methods for the characteristics of public health decision-making need to be explored.