Connectathon Planning is beginning - started a confluence to managing this. If you want to be able to edit
Webinar for the track, we will hold two: one next week for people who have never done a connectathon. Another one towards the end of this month in prep for the Connectathon
Will use one our Friday call on 4.17.20 to hold the first one.
BU Resiliency Challenge - 9-week hackathon being held by BU Spark! SANER Project is participating.
Rush University work?
Trying to find out from both sides
City of Chicago put out a public health order
Hospitals in their jursidiction to provide a copy of the CDC/NHSN reports to them as well
Right now, there are a few options for submission, but currently all manual
Working on a FHIR-based approach as well
Hans has Rush contacts and can try to reach out. Keith to check in with Abigail Watson.
Want to make sure that the work they do aligns with what we are doing
Introduced Brian Harris and Sandeep Savarala from the Ai team.
Renamed the saner-ig to saner-ig-old to make sure we keep things straight and there is a redirect on the old one.
Hot Topics
Testing Commitments
Checked in with the team on testing commitments for connectathon, will continue to follow up on following meetings.
Sensitive Information Sharing
Security updates from John Moehrke are listed here.
SANER starting to appear in the news, we just want to identify any risks to organizations (# of deaths in facility), regions (# of deaths in the region), ramifications of these risks like being targeted or being bullied (not explicitly expressed but also could man targeting by scams and advertising).
We want to make sure the regions we are reporting on is not too small. "Too small" was not defined. K-anonymity starts at a value of 3. This means you don't have any data in your data set that has groups smaller than 3. Implying that if you have less than 3 incidences you should report it as NA until you have 3 - but this is potentially problematic.
Some things we should be thinking about while designing IG: not grouping all reports into one document even though it may seem efficient. As soon as you do that you create vectors where you are reporting on a region/age/gender in one report and people can see correlations.
Operational Mitigations we should express: Having to do a risk analysis at each location, making regions as big as they need t be to not have data grouped, if you need it more fine-grained locations then you have sensitive data and you have to be careful who you share that data with.
The information being reported about beds available/occupied is sensitive information because it indicates how well you are doing as a business. That is only shared in situations like this one or in emergencies.
Summary reports can have risk but the risk discussions in the de-id handbook are not as easily applied towards sensitivity on a summary report. We may just have t explain in our own security considerations our specific points.
Cohort vs. Measure Population - if you read the quality reporting section of FHIR if you are doing ratio or measure Cohort isn't really the right answer. May want to reach out.