Implementing GRADE in Guideline Development: RealWorld Experiences

1

• Implementing GRADE in Guideline Development
  Real-World Experiences
• NIAID Guidelines for the Diagnosis and Management
  of Food Allergy
• Dr. Matthew Fenton
• Asthma Allergy Inflammation Branch
• DAIT, NIAID, NIH
• September 15, 2009

2
Food Allergy Guidelines Background

• July 2007 workshop assessed the need for clinical
  guidelines that could be used by health care
  providers across various medical specialties.
• More than 30 professional organizations, federal
  agencies, and advocacy groups unanimously agreed
  that NIAID should develop the guidelines.
• Guidelines should be developed using both
  evidence-based data and expert opinion.
• Wide dissemination of the guidelines by
  participating organizations would be needed.

3
Food Allergy Guidelines Process Overview

• Literature Review and Evidence Grading (RAND
  Corp.)
• Comprehensive literature search
• Prepares evidence tables
• Assesses quality of body of evidence using GRADE
• Expert Panel (Chair Dr. Joshua Boyce)
• Drafts the guidelines based on RANDs literature
  review and expert clinical opinion
• Identifies both knowledge gaps and areas of
  agreement
• Drafting, Review and Final Guidelines
• Guidelines draft prepared by the Expert Panel to
  be edited based on Coordinating Committee review
  and public comment
• across medical specialties
• Committee comprised of professional
  organizations, Federal agencies, and advocacy
  groups

4
Guidelines Timeline

• July 2007 Meeting highlighting the need for
  Guidelines
• September 2008 First meeting of the Coordinating
  Committee award of RAND contract for evidence
  based review
• July 2009 First draft of RAND report given to
  Expert Panel
• October 2009 Expert Panel meets to review first
  draft of the guidelines
• Dec 2009 Final draft of Expert Panel guidelines
  to be completed and forwarded to the Coordinating
  Committee
• Jan-Feb 2010 60 day public comment period
• May 2010 Final guidelines document ready for
  release

5
Food Allergy Guidelines Implementing GRADE

• Training the Expert Panel Members
• Dissemination of papers from the GRADE Working
  Group
• GRADE presentation at March 18, 2009 Expert Panel
  Kick Off Meeting
• Ongoing discussion to refine recommendations for
  grading/strength
• RAND provides additional training

6
Food Allergy Guidelines Implementing GRADE
(Contd)

• Preliminary Scoring plan
• 5 distinct writing groups to prepare guidelines
  chapters
• Each small group will propose the strength of a
  recommendation based on RAND evidence and expert
  opinion using GRADE
• Vote by entire EP upon integration will determine
  final grading strengths
• Tentative plan to assess agreement by Panel
  members
• Each section of the guidelines will contain a
  agreement score to compliment the evidence
  grade and recommendation strength
• Uniform (100 agree), Substantial majority (80),
  majority (50-80), No agreement (lt50)
• Agreement will be assessed upon integration of
  the 5 guideline chapters using the GRADE grid

7
Food Allergy Guidelines GRADE Potential
Difficulties

• Nature of food allergy and state of the science
• Few DBPC clinical trials (ethical issues)
• No existing treatments for food allergy
• Relatively few population-based studies
• Reported trials use different endpoints
• Potential population environmental
  differences

8
Food Allergy Guidelines GRADE Potential
Difficulties

• Difficult to rate some evidence highly, leading
  to a lack of strong recommendations
• Example Oral food challenge as a study endpoint
• Oral food challenges, the gold standard for food
  allergy diagnosis, can be very risky for those
  with severe food allergies
• Many clinical trials use oral preparations to
  desensitize patients, but, to minimize risk, do
  not use oral food challenge as the endpoint to
  test the success of the intervention
• Such studies are not considered the highest
  quality, due to lack of the gold standard
  endpoint
• Does GRADE work where the body of evidence is
  considered weak, but expert opinion is strong?