Skip to content

Personas & Learning Paths

Not everyone arrives at this program with the same background or goal. Use this page to find the track that best matches your role, then follow the recommended sequence. The core four days are required for everyone; the persona-based guidance shapes how you engage and which extras to prioritize.

At a Glance

Persona Start Here Key Tools Skip / Skim
Vocabulary & Terminology Expert Day 1 → Day 2 → Day 3 Athena, Atlas Concept Sets Day 4 SQL depth, Day 6 R
Statistician / Data Analyst Day 3 → Day 4 → Day 5 Atlas Pathways, HADES, SQL review Environment deep-dives
Data Engineer Module 0 → Day 2 → Day 4 Databricks, DBeaver, SEARCH, GitHub Atlas GUI portions
Clinician / Research Analyst Day 1 → Day 3 Athena, Atlas Cohort Editor SQL labs (participate lightly)

Persona 1 — Vocabulary & Terminology Expert

Who you are: Clinical coders, medical informaticists, clinical terminology leads, or anyone whose primary responsibility is understanding and curating medical vocabularies and code sets.

Your goal: Build reliable, reproducible concept sets that capture the clinical phenomenon you care about — and understand how vocabulary structure affects every downstream analysis.

  1. Module 0 — Environment Setup · Verify Athena access and Atlas login.
  2. Day 1 — OMOP CDM · Focus on vocabulary tables (concept, concept_relationship, concept_ancestor). The table overview matters less than the vocabulary deep-dive.
  3. Day 2 — Vocabulary & Data Quality · This is your primary session. Spend extra time on the Standard/Mapped toggle and the DQD concept-set vs. data-quality distinction.
  4. Day 3 — Cohort Definition · Understand how concept sets plug into cohort entry events and inclusion rules. You don't need to build pipelines, but you need to validate what analysts build.
  5. Day 4 — Data Extraction · Participate lightly; focus on how extraction counts reflect your concept set choices.
Tool Why it matters for you
Athena Primary workspace — search, compare, inspect vocabularies
Atlas Concept Sets Build and validate the reusable concept sets used everywhere
SQL Client (Databricks / DBeaver) Confirm your sets against the CDM with the SQL Cheat Sheet
EHDEN Academy Deep-dives on standardized vocabularies and ETL
  • Day 5 (Optional): See how concept sets directly drive treatment pathway results — a powerful feedback loop for refining your sets.
  • White Rabbit / Usagi: Tools for source code mapping; relevant if your institution is doing an ETL.
  • Oncology-Specific Resources on the Resources page: key if your work involves cancer registries.

Persona 2 — Statistician / Data Analyst

Who you are: Epidemiologists, biostatisticians, outcomes researchers, or study designers who understand observational study design and need to operationalize it in OHDSI tools.

Your goal: Translate a study protocol into a cohort definition, run the analysis in ATLAS or HADES, and interpret the results with appropriate caution about confounding and data quality.

  1. Module 0 — Environment Setup · Confirm R, RStudio, and HADES install if you plan to do Day 6.
  2. Day 1 — OMOP CDM · Focus on how OMOP domains map to study design constructs (exposure, outcome, covariate).
  3. Day 2 — Vocabulary & Data Quality · Focus on concept set validity and DQD as a study-fitness check.
  4. Day 3 — Cohort Definition · This is your primary session — build the new-user cohort with careful temporal logic.
  5. Day 4 — Data Extraction · Understand what the extracted dataset looks like and how to validate counts.
  6. Day 5 (Optional) — Treatment Pathways · Directly applicable to treatment utilization and sequencing research questions.
  7. Day 6 (Optional) — HADES · Essential if you will run characterization, estimation, or prediction pipelines.
Tool Why it matters for you
Atlas Cohort Editor Operationalize inclusion/exclusion criteria with temporal logic
Atlas Pathways Treatment sequence visualization
HADES (R) CohortMethod, PatientLevelPrediction, FeatureExtraction
RWD Guide Bias, confounding, and study design for observational data
  • Book of OHDSI Ch. 12–14: Estimation, Prediction, and HADES — your primary reference.
  • OHDSI Community Calls: Great for hearing real-world study design trade-offs.
  • Day 6 HADES kit from Resources: includes patient-level prediction notebook and interpretation guide.

Persona 3 — Data Engineer (SQL-First)

Who you are: Data engineers, SQL developers, BI analysts, or ETL developers whose primary workflow is the command line, SQL editors, and code repositories — not GUIs.

Your goal: Understand the OMOP CDM well enough to build reliable extraction pipelines, validate results programmatically, and maintain reproducible analysis code.

  1. Module 0 — Environment Setup · Set up Git, confirm SQL client connectivity, and confirm GitHub repo access. This is your most important session.
  2. Day 1 — OMOP CDM · Focus on schema, table relationships, and SQL queries. Use the Day 1 Code Snippets throughout.
  3. Day 2 — Vocabulary & Data Quality · Focus on the SQL validation steps. The Atlas GUI is secondary; the SQL behind it is what matters.
  4. Day 3 — Cohort Definition · Export cohort SQL from Atlas and run it yourself. The SQL Validation Mini Lab is key.
  5. Day 4 — Data Extraction · Your primary session — work through every step in your actual SQL client.
  6. Day 6 (Optional) — HADES · Run HADES R packages via DatabaseConnector; relevant if you support statistical workflows.
Tool Why it matters for you
Databricks / DBeaver / SQL Client Primary workspace — run and validate every Atlas export
GitHub Version-control cohort definitions, SQL scripts, and notebooks
SEARCH Tool Site-specific extraction pipeline
OMOP SQL Cheat Sheet Quick reference for every CDM query pattern
SQL Validation Mini Lab Step-by-step export → validate → reconcile workflow
  • DatabaseConnector (HADES): Connects R directly to your CDM — highly relevant if you support analysts.
  • White Rabbit / Rabbit-in-a-Hat: Profiling and ETL design tools.
  • OHDSI on GitHub: Explore the CDM DDLs, HADES packages, and community pipelines.
  • Capstone Module (Optional, Module 10 in Syllabus): End-to-end mini study — build in your SQL client.

Persona 4 — Clinician / Research Analyst

Who you are: Physicians, nurses, clinical research coordinators, or patient advocates who need to understand how EHR data is organized and how to interpret OHDSI analyses — without necessarily writing code.

Your goal: Develop enough OMOP/OHDSI literacy to collaborate effectively with data teams, review cohort definitions for clinical accuracy, and interpret analytic results.

  1. Module 0 — Environment Setup · Focus on ATLAS login and Athena access. SQL client setup is optional.
  2. Day 1 — OMOP CDM · Focus on conceptual understanding: how EHR data maps to OMOP domains. The SQL exercises are optional for you.
  3. Day 2 — Vocabulary & Data Quality · Use Athena to explore concepts relevant to your clinical area. Focus on the Standard/Mapped distinction and what bad concept sets look like clinically.
  4. Day 3 — Cohort Definition · Your most important session. Review cohort logic for clinical accuracy — are the criteria capturing the right patients? Review characterization outputs.
  5. Day 4 — Data Extraction · Attend for context; extraction validation is primarily for the data team.
Tool Why it matters for you
Athena Explore concepts and hierarchies in your clinical domain
Atlas Cohort Editor Review cohort definitions for clinical face validity
Atlas Characterization Interpret who is in a cohort — demographics, prior conditions, medications
EHDEN Academy Self-paced intro modules for non-technical learners
  • "Introduction to OMOP" course (Tufts CTSI iLEARN) — accessible overview, no coding required. See Resources.
  • RWD Guide — written for clinicians and researchers new to real-world data.
  • Characterization outputs from Day 3: Bring a real clinical question and review the output with the data team.

Not Sure Which Persona Fits?

Start with Day 1 regardless of background — the OMOP CDM is the foundation everything else depends on. After Day 1, reassess: if you found yourself most interested in the vocabulary tables, you're probably Persona 1; if you were most engaged by the study design implications, you're Persona 2; if you immediately wanted to write the SQL yourself, you're Persona 3; if the clinical accuracy of concept sets was your primary concern, you're Persona 4.

Most participants end up combining elements of multiple personas. The paths above are starting points, not rigid rules.