‹ Home
Exported

CSTE OD Spike Alert Advisory Group

CSTE has established the Overdose Spike Alert Advisory Group to focus on addressing three key needs and provide corresponding guidance on: 1. What data spikes mean (and what spikes are meaningful) 2. What actions should be taken 3. Variation at state and at local levels. This activity will principally address analysis and interpretation of surveillance data to identify overdose clusters/spikes, while also accounting for aspects of alert guidance that intersect with epidemiologic investigations, partnerships with key stakeholders, and prevention. Membership in the Advisory Group is intentionally limited to maintain a an effective size for resource development. CSTE will seek input outside of the group from a wide range of stakeholders. The group will develop an online toolkit that will be publicly available by late 2021. While CSTE staff will monitor this page and activity regularly, anyone with the link can access these resources. Please do not share any proprietary information or upload private documents. Email Danielle Boyd (dboyd@cste.org) if you have questions about the advisory group or about this basecamp site for the group.
Megan Toe, CSTEHayleigh McCall, Program Analyst III, Syndromic Surveillance at CSTEDanielle Boyd, Program Analyst at CSTESandra GonzalezAlyaa Altabbaa, Program Analyst at CSTECailyn Lingwall, Program Analyst at CSTELawrence SchollKrystal Collier, Syndromic Surveillance Program CoordinatorBarbara MassoudiKara ManchesterKenan ZamoreJennifer KretAmy BoardJoseph Spadafinohoward burkom +20 others
Posted by Megan Toe ·

Initiating Thread for Toolkit Layout/Visual/Functional Comments

Creating a message thread as a parking lot for any and all comments and ideas on key design related elements for the online toolkit as we continue our primary focus on the substantive content through September, we are also initiating backend layout.  There will be concrete future opportunities to respond to upcoming demos and formally provide feedback.  This is an place for suggestions and ideas as they arise and as we share some current CSTE toolkit examples.

Comments & Events

Danielle Boyd, Program Analyst at CSTE Danielle Boyd, Program Analyst
Toolkit Comments: Desired Functions
--> Search Feature 
--> Content Directory listed on the side
--> Intro sections for the topics with ability to expand for more information

Toolkit Comments: Desired Links
--> Relevant Articles 
--> Code
--> Visualizations: Dashboards?
Notified 30 people
Krystal Collier, Syndromic Surveillance Program Coordinator Krystal Collier, Syndromic Surveillance Program Coordinator
Thinking about content for the toolkit,  I have a question about possible code using ESSENCE API, R, SaTScan for the group.  I am testing some code with Roseric from NSSP, based on the RNSSP Package templates and some other R Code that has been shared for overdose cluster detection and other cluster detection.

Would anyone be interested in testing this?  Right now we have a version that's purely spatial and easier to test without needing to download SatScan. I am asking, because if we can get a template report that can be added to the RNSSP Package Templates, then we would have a tool that could support the community, while also having some content that can be added to the toolkit.  The code I have right now that is easiest to adopt into a template is not purely spatial and not using SaTScan.  So it is missing that time/space cluster detection.  However we have some working code that does this but you would need to download SaTScan for your desktop.
Notified 30 people
Krystal Collier, Syndromic Surveillance Program Coordinator Krystal Collier, Syndromic Surveillance Program Coordinator
One things a group of us working on overdose cluster detection for the CSTE DSTT project is identifying parameters for the cluster.  Have we identified some good examples of this in the toolkit?  This is one of the hardest things to find resources on.  Wondering if the toolkit would benefit from a focus on adding something with the code or visualizations?  One question that needs assistance from overdose cluster detection is how do you know what parameter to choose for the location of the cluster like city block, etc?  Is there a gold standard for this?  Is there documentation that would be helpful to share with others to make these determinations?
Notified 31 people
howard burkom howard burkom
This is not a short discussion, Krystal, It could be helpful to schedule another “happy hour” session on this at the upcoming symposium, but good planning with shareable examples would help.
When you say “city block”, do you mean you have address data for clustering? Right now my projects are restricted to the zip code level.
Sharon Greene in NYC has had and published success at the census tract level. With ODMAP data, I tried census tract and census block group granularity, and the block group tests didn’t give any advantage over the census tract tests. In a city like Phoenix or western Portland OR, block group granularity might be informative, but not at the state level.
Information about what health departments can do with cluster details could help answer your questions about parameter choices. Glad to discuss further.

Best,
Howard
Notified 31 people
Jennifer Kret Jennifer Kret
Krystal Collier, Syndromic Surveillance Program Coordinator Krystal Any chance that the RNSSP Package templates and other R Code  for overdose cluster detection (including spatial components) would be translated into SAS code? I don't use R but this sounds exciting!

Also, I'm also very interested what the toolkit will include re: examples for cluster detection and parameters. I have seen several variations presented at the OD2A DOSE quarterly calls/annual meeting, and CSTE Overdose subcommittee. Everyone uses different criteria it seems though...  what, in your mind would make a "good example"? :) 
Notified 31 people
Krystal Collier, Syndromic Surveillance Program Coordinator Krystal Collier, Syndromic Surveillance Program Coordinator
Jennifer Kret Jennifer I don't think it will be translated into SAS code but you might want to talk to Aaron Kite-Powell with NSSP about options.  The one thing you might want to try is exploring the RNSSP Package and the templates in R.  Its intended to be a tool that you don't need to do the R coding to run the templates.  That's why I like it.  I also like the flexibility for this purely spatial code that can be used for overdose cluster detection or other CCDD categories.  I wanted the flexibility to have a standard report if possible that I could just select CCDD Categories with ease and not have to rewrite the code for every CCDD category if we had to do this on our own. 

I have heard like you different criteria.  I am wondering if folks have examples to reference then they will have the ability to determine the parameters easier if there isn't a gold standard for example.  I don't have "good example" in mind, just examples.  Thinking of the audience using the toolkit, I want to see example of what is being in practice, so if I want to implement I can by referencing an  example in the toolkit.  Just my thought.
Notified 31 people
howard burkom howard burkom
I just saw this thread, which reminded me of a couple of wrappers for SaTScan written in SAS and made publicly available. If you want to see them and have trouble finding them, I may have them.
Notified 31 people
Danielle Boyd, Program Analyst at CSTE Danielle Boyd, Program Analyst
Add a button for users to Suggest and Edit (could be used to notify when a link isn't working or new resources are available.)
Notified 32 people