Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for epidemiologists

Data Sharing Agreement Compliance Agent

Analyses and outputs that stay within the data sharing agreement, with every issue fixed before release

Data Sharing Agreement Compliance Agent: what goes in, what the agent does and what you get

What it does

Shared health data comes with strict terms: allowed uses, small cell suppression, no linkage to other datasets. This agent reads the agreement and turns its terms into a checklist. It checks the analysis plan against allowed purposes and variables, and reviews draft tables and outputs for small cell counts and disallowed linkages. Problems are listed with a suggested fix, such as combining categories. After the analyst revises outputs, it rechecks them. The data steward approves release. Edge case: a table of 8 cases in a small county is cell-suppression safe in total, but a breakdown by age creates cells of 2, so the agent flags the breakdown.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Analysis plan or draft output submitted 2 USES A TOOL Read the agreement and build a terms checklist 3 DOES Compare the plan's purpose and variables to alloweduses 4 CHECKS THE RESULT Are purpose, variables and linkages all allowed? If not: list the disallowed items and suggest a change.Back to step 3. 5 USES A TOOL Read draft tables and figures 6 DOES Scan every cell for counts below the minimum 7 CHECKS THE RESULT Does every cell meet the minimum count and rule? If not: suggest combining categories or suppressingcells, then rescan. Back to step 5. 8 DOES Draft the compliance note for the data steward 9 YOU APPROVE Data steward approves the release 10 RESULT Release-ready outputs and checklist
Read the steps as a list
  1. Analysis plan or draft output submitted
  2. Read the agreement and build a terms checklist
  3. Compare the plan's purpose and variables to allowed uses
  4. Are purpose, variables and linkages all allowed?If not: list the disallowed items and suggest a change. Back to step 3.
  5. Read draft tables and figures
  6. Scan every cell for counts below the minimum
  7. Does every cell meet the minimum count and rule?If not: suggest combining categories or suppressing cells, then rescan. Back to step 5.
  8. Draft the compliance note for the data steward
  9. Data steward approves the releaseThe agent waits here for your OK.
  10. Release-ready outputs and checklist

How it decides

An output passes when its purpose, variables and linkages are allowed and every cell meets the minimum count.

  • Suppress any cell below the agreement's minimum, usually 5 or 11
  • Block linkage to any dataset not named in the agreement
  • Check complementary suppression in totals
  • Flag any new variable not on the approved list

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Minimum cell size
  • Allowed variables list
  • Agreement expiry date
  • Steward contact

What keeps you in control

It always asks you first

  • Release of outputs
  • Any request to amend the agreement

Hard limits

  • Never releases data
  • Never interprets the law beyond the written terms

It stops when

  • Done: outputs pass the checklist
  • Stop: agreement is unclear or expired

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA draft table by county and age had a cell with 3 cases, so the small cell check failed against a minimum of 11. The agent suggested combining ages into three groups, then rescanned and found the group totals still allowed backing out a suppressed cell, so it flagged complementary suppression. After the analyst revised, all cells passed and the steward approved release.

More agents for epidemiologists