Cosmos Onboarding Guide

You will need:

•  SUNet ID
•  SHC SID (Stanford ID) or SMCH Network Login
•  Epic UserWeb account
•  Review and follow Cosmos "rules of the road" (more info)

When publishing Cosmos-enabled findings, you must acknowledge Cosmos in your publication and, twice a year, provide the Governing Council with a list of your citations and any other works referencing your Cosmos-based research (per Section 2.3.1 of the Cosmos Terms).

1. SID Request:
   Have someone with an SID request one for you 
       In the Onboarding Request form, choose:
          • On-Site User
          • SOM User Type “Other” (most cases)
          • Organization "School of Medicine"
          • For "Requested Epic Access,” choose “Other”
          • For “Please describe ‘Other’:” write “Cosmos User; No patient-facing activity and does not need to do FY25 SMHC New Hire Employee Training Curriculum HealthStream training”
          • For “Application Access Options” choose “Epic”
          • For “Other Access” choose “Other” and write “Cosmos and SlicerDicer Access” “Minimal Cosmos User (SHC ANALYTICS ONLY [T00275] template)”
          • Enter existing SoM email for SID request
          • Enter “Cosmos User” for the “Business Justification” for the Epic request
   

          • Enter “Cosmos User” for the “Business Justification” for the Other Access request

2. Epic UserWeb Account:
   Create an account at Epic UserWeb (select Stanford Health Care).
3. Cosmos Access Request:
   Complete the Stanford Cosmos Access Request form. Ensure that the URL of your Epic UserWeb profile is included.
4. Training:
   Once Cosmos access has been granted, you can complete the applicable Epic Cosmos trainings. Sign up at Epic Training Catalog (search "cosmos" under "all applications").

• Aggregate-level access:
  Use Slicer Dicer in the Cosmos Portal for cohort discovery and feasibility analysis.
  
  
• Data Science Virtual Machine (DSVM) access:
  Line-level access (SQL):
  Requires Epic COS305 ($200) + COS500 ($600), ($800 total) (Data Model for Data Architects certificate).
     
  Line-level data science tools access (R and/or Python):
  Requires Epic COS305 ($200) + COS550 ($200), ($400 total) (Data Scientist certificate).
  
  
• Read more about Cosmos training here

• After approval, log in to the Cosmos Portal with your Stanford credentials. 
• Slicer Dicer: Explore cohorts, diagnoses, medications, and outcomes with built-in filters and visualizations.
• DSVM: Use R, Python, or SQL for advanced analysis and data mart creation. Data marts can be exported as SQL tables, CSV, or parquet files. You must send an email request with proof of certification to cosmosteam@stanford.edu to request DSVM access if certified or accredited in Cosmos Data Model for Data Architects or Cosmos Data Model for Data Scientists.

• Cosmos includes demographics, diagnoses, labs, medications, procedures, SDOH, and more.
  
• Dates are shifted per patient; ages may be truncated (all birthdays set to Jan 1).
  
• No direct identifiers; LDS excludes direct dates and other PHI.

• Define your study or project.
• Use Slicer Dicer for feasibility, then request a data mart from a Stanford data architect for complex analysis (contact Kameron Black, if needed).
• Notify Epic of your publication intent.
• Import data marts directly into R/Python via SQL connection for analysis.
• Validate, analyze, and publish results per Stanford and Epic guidelines.

• Before publishing, clearly define your base population using recommended filters (e.g., "Has any encounters?" and "Base Patient") and avoid drawing conclusions from random samples ("Sneak Peek" mode).
• Review known data limitations, biases, and data quality issues for all relevant domains (see Cosmos Data Science White Papers and Data Encyclopedia).
• When ready to publish or share, mask small counts (<11), request file transfer via your Cosmos representative, notify Epic of your publication or presentation (CosmosPublications@epic.com), and include required Cosmos attribution.
• Information about your publication, including basic details about the associated Cosmos project and a link to the abstract, will be made available for other users in the community on the Cosmos Publication and Funded Studies dashboard within the Cosmos portal.
  
  For externally funded studies, submit your publication, presentation, or abstract to the Cosmos community, ensuring transparency regardless of publication outcome.
  More info

• Before reaching out for help, we recommend that you first review the Cosmos documentation.
• We also recommend that you consult the Cosmos Data Dictionary.
  
  
• If you are stuck, you can ask for help from other Stanford Medicine researchers on the cosmos_users Slack channel. 
• For help from the wider Cosmos community, you can also post your question in UserWeb.
• To provide suggestions for Cosmos enhancements, post them in the Cosmos section of Ideas.
  
  
• For issues with your Stanford Cosmos account, please contact cosmosteam@stanford.edu
• Note: If you do not log in to Cosmos for 180 days (~6 months), you will be blocked for inactivity. If this happens, please contact us at cosmosteam@stanford.edu to request that your Cosmos account be reactivated. You can avoid this issue by logging in to Cosmos at least once every 180 days (~6 months).
  
  
• For questions about Cosmos-related publications and collaborations at Stanford, please contact Kameron Black