Jurisdictional Decision Support Tool by Mark Hamblin, Carnegie Mellon University

Slide 1 - Jurisdictional Decision Support Tool

Mark Hamblin

Carnegie Mellon University

Slide 2 - Public Policy Project Course

• Tissue Engineering
• Navigating the FDA approval process
• Social and ethical issues
• Financial and marketing issues
• Jurisdictional determination
  • Review of current process
  • Review of inter-center agreements
  • Creation of web-based decision support tool

Slide 3 - Current Jurisdiction Process

• Intercenter Agreements provide rules for classifying combination products, but are too focused in scope
  • Only cover existing technologies
• Intercenter Agreements may not apply to new technologies
• Jurisdiction determination is then based only on the “primary mode of action”, of which no clear definition exists
• Some subjectivity is necessary to reach a decision, yielding a lack of consistency, predictability, and transparency

Slide 4 - Purpose of Decision Support Tool

• Create a rule based system that classifies medical products based on product characteristics
• Incorporate previously established jurisdiction rules from Intercenter Agreements
• Add additional criteria for determining jurisdiction to fill in the gaps
• Allow for easy adaptability and variability to accommodate current FDA regulatory requirements and trends
• Make the tool widely available (web-based) to allow for greater transparency and predictability in jurisdictiondetermination process

Slide 5 - Details of Decision Support Tool

• Each product has 3 “pools of points”, one each for each of the three FDA regulatory centers
• List of 88 yes/no questions pertaining to product characteristics
  • If “yes” for a question, X points go to pool 1, Y points go to pool 2, Z points go to pool 3
  • Each question has a weight from 0 to 1 based on how important that question is in the overall classification scheme
  • Points for each pool get scaled based on the weight for that question
• Product gets classified into the respective center based on the “pool” that has the most points
• Easy to change classification scheme by changing respective weights of questions and point distributions
  

Slide 6 - Creating Model Inputs

• Extracted 67 questions from the rules in the Intercenter Agreements
• Conducted a survey of tissue engineering experts
  • Sent to 205 members of the Pittsburgh Tissue Engineering Initiative
  • Proposed 21 different product characteristics
    • Surveyed how the experts thought the presence of these characteristics should affect classification
    • Created 21 questions for the model based on these product characteristics
    • Assigned points for the model based on the survey responses
  • Also gathered responses regarding experiences and opinions of FDA jurisdictional decision process and approval process
    

Slide 8 - How It Helps

• Intercenter Agreements form a precedent based decision model by looking only at specific characteristics of previously developed products
• Precedent based decision models typically are not optimal for classifying new types of products
  • Too subjective
• Proposed decision support tool is a rule based model that looks at a product’s general characteristics
  • More applicable to future products and technology
• Widely accessible rule based decision model will provide a consistent, predictable, and transparent method for classification problems

Slide 9 - Final Presentation and Report

• Final presentation to be given in Washington DC on Wed. December 4, 2002
• Members of review panel include senior management from FDA, academic and industry researchers, other stakeholders
• Publication of written report expected in January 2003
• Contact mhamblin@andrew.cmu.edu for details on final presentation or written report