Explore decision tree tools for data sharing in public health emergencies

[Daniel-Mietchen]

The Global Research Collaboration for Infectious Disease Preparedness (GloPID-R) has a Data Sharing Working Group, which is working on a decision tree to help address issues around data sharing over the course of a public health emergency. While progress has been made on the kinds of questions that need to go into the tree, it is still unclear how best to visualize and share the tree. It could thus be a nice project for this hackathon to look into this. Some free and open-source options that are being considered:

• Twine
• Silver decisions
• Gambit
• various R packages

[Daniel-Mietchen]

Here are some of the questions that should be addressed in the decision tree:

• How does the disease present? How do we recognise it?
  • output case description disseminated widely case definition
• How does the disease progress over time? Who is at risk of severe disease, how do we recognise them?
  • severity is included here e.g. helps to define who the high risk groups are for a vaccine programme
  • Outcome measures
  • impact on ICU beds and materials
• How do we prevent transmission in?
  • mode of transmission inform key IPC interventions
  • animal - human interface understood for community IPC and animal husbandry guidance public engagement
• Which drugs to use and what dosage to give?
  • study protocols launched supportive care may have to be included
  • clinical trials prepared and started
• What samples are required?
  • sampling protocol agreed to include serial sampling from multiple sites
  • feeds into IPC and public health messaging
• What is the best treatment strategy?
  • e.g. ventilator strategy
• What is the pathogenesis?
  • biomarker identification
  • feeds into sampling protocols
  • influences interventional basic science and study protocols
• How do we best treat children?
  • outcomes may differ to adults
  • vaccine strategy

Another aspect to keep in mind is that the decision tree may differ for different stakeholders (e.g. clinicians, policy makers, modelers, public health) and that it may be adapted for different purposes (e.g. different outbreaks, different countries, languages or years).