Epic Cosmos

Epic Cosmos User Guidelines for Medical Students

What Is Epic Cosmos?

Epic Cosmos is a large, de-identified electronic health record (EHR) research database maintained by Epic. It allows users to explore clinical data across participating health systems to support hypothesis generation, observational research, and quality improvement projects.

Available data include:

  • Diagnoses (ICD-9/10- CM, problem lists)
  • Procedures (CPT)
  • Medications (RxNorm)
  • Demographics (limited, de-identified)
  • Encounters (dates, provider specialty)
  • Social determinants of health (e.g., Area Deprivation Index)
  • Limited clinical notes (in development), etc.

Cosmos data are de-identified and primarily intended for aggregate analysis and research exploration.

Who Can Use Cosmos?

  • Clinical Researcher and investigator
  • Medical students engaged in approved research projects under the PI of DH/Geisel/VA
  • Students must complete required onboarding and training
  • Access is granted only after all prerequisites are met

 

Access Requirements and Logistical (Required for All Students)

To be granted Epic Cosmos access, students must complete of the following:

  • First, reach out to the Office of Medical Student Research. You must have sponsor PI and research project.
  • Second, complete the user training. Student must complete required onboarding and training. (Students are encouraged to pursue the Super User Badge to build advanced data skills. The badge requires access to Cosmos and offers additional training opportunities; while optional, it can be valuable for students planning to work with clinical or research data).
  • The Office of Medical Student Research will then review and confirm the user and will request the approval access from Application System Analyst at Dartmouth Health .
  • Access is granted only after all prerequisites are met

Note: Student accounts may be deactivated if the student is no longer an active user, if the research project has been completed, or after the student has graduated from Geisel.

  1. Epic UserWeb Account (Required)
  • Log in at: https://userweb.epic.com using your Dartmouth Health credentials
  • If prompted with an Employment Confirmation message:
    • Select “Dartmouth Health employee”
    • Do NOT select “Third-Party Contractor” (this will delay access)

⚠️ This step maybe the most common cause of access delays.

  1. Local eDH Epic Account (Required)
  • Many medical students already have an eDH account
  • If you do not have one:
    • A request will be submitted through DHSM to the Users and Provisioning team
    • This account is required even for Cosmos-only access

 

Training Options (Choose Based on Your Research Needs)

Option 1: SlicerDicer (Aggregate-Level Data)

Recommended for:

  • Hypothesis generation
  • Exploratory research
  • Projects not requiring line-level data

Key features:

  • Graphical interface (no coding required)
  • De-identified aggregate counts only
  • No IRB oversight required
  • Limited data export

Recommended training (via Epic weLearning):

  • COS100
  • COS200
  • COS300
  • COS310

Option 2: Data Science Virtual Machine (DSVM – Line-Level Data)

Required for:

  • SQL-based queries
  • Advanced data analysis
  • Python, R, or Excel workflows

Access requires Epic certification:

  • COS305 (Data Architect prerequisite)
  • COS500 (Data Architect Certification)

Important notes:

  • SQL proficiency exams required
  • Data remain de-identified (zip code and dates obscured)
  • Includes access to EDDI (Cosmos database)
  • Higher training burden; not required for most student projects

See more information and details on this document: Cosmos Training

Super User Training (Optional but Encouraged)

  • Asynchronous, free Epic training (~8 hours)
  • Designed to create first-line Cosmos support
  • Increases institutional Cosmos licenses
  • Recommended that each department has at least one Super User

Responsible Use Guidelines (Do’s and Don’ts)

✅ DO

  • Use Cosmos only for approved academic or research purposes
  • Ensure your project aligns with de-identified, aggregate data use
  • Use the Cosmos Data Dictionary to understand data availability and quality
  • Follow publication and acknowledgment guidelines
  • Ask for guidance if unsure about appropriate data use

❌ DO NOT

  • Attempt to identify individuals or institutions
  • Download or store restricted data locally
  • Use Cosmos for clinical decision-making
  • Share access credentials
  • Misrepresent Cosmos data as institutional-specific results

Publications and Project Approval

  • Some projects may require completion of:
    • Cosmos Publication Checklist
    • SYNERGY Consult Request Form (for funded projects or advanced analysis)
  • Always confirm publication requirements before submission

For more details:  Cosmos Publication Checklist

Getting Help : Reach out to the office of medical student research

  • Use Epic’s eLearning resources for training support
  • Consult your project mentor or research office
  • For advanced analytics or funded projects, SYNERGY Informatics support may be available

More information: Cosmos Rules of the Road