Skip to main content

Study Packages (Strategus)

Study Packages enable researchers to build and execute multi-analysis OHDSI study packages using the Strategus framework. Strategus orchestrates the execution of multiple HADES analysis modules (CohortMethod, PatientLevelPrediction, SCCS, etc.) as a single coordinated study against one or more CDM data sources. Each module runs independently and writes results to a shared output directory.

Parthenon provides a 5-step wizard for assembling study packages, a validation engine, and an execution interface with progress tracking.

Module Architecture​

StudyPackagePage
 |
 |-- Header (Package icon, Import/Export JSON buttons)
 |
 |-- Step Indicator (5 steps, clickable breadcrumb)
 |
 |-- Step 1: Study Info
 |    |-- Study Name (required)
 |    |-- Description (optional)
 |
 |-- Step 2: Select Modules
 |    |-- Grid of OHDSI analysis modules
 |    |-- CohortGenerator always included (required)
 |    |-- Toggle selection per module
 |
 |-- Step 3: Shared Cohorts
 |    |-- Assigned cohorts list (with role badges)
 |    |-- Role selector (target / comparator / outcome)
 |    |-- Cohort picker (searchable, from existing definitions)
 |
 |-- Step 4: Review & Validate
 |    |-- Study Package Summary
 |    |-- "Run Validation" button
 |    |-- Validation result (passed/failed, issues, warnings)
 |
 |-- Step 5: Execute
 |    |-- Target data source selector
 |    |-- "Execute Study Package" button
 |    |-- Progress indicator
 |    |-- Execution result (status, modules run, output dir)
 |
 |-- Navigation (Back / Next)

Feature Walkthrough​

Step 1: Study Info​

Provide a Study Name (required) and an optional Description. The name is used as the filename for JSON export and identifies the study package in execution results.

Step 2: Select Modules​

Choose which OHDSI analysis modules to include in the study package. Modules are displayed in a selectable grid with descriptions:

ModulePackageDescription
Cohort GeneratorCohortGeneratorGenerates cohorts from definitions. Always included.
Cohort MethodCohortMethodPopulation-level effect estimation using comparative cohort design
Patient Level PredictionPatientLevelPredictionBuilds prediction models for patient-level outcomes using ML
Self-Controlled Case SeriesSelfControlledCaseSeriesEstimates incidence rate ratios using SCCS design
Cohort DiagnosticsCohortDiagnosticsEvaluates phenotype algorithms and characterizes cohorts
CharacterizationCharacterizationComputes baseline characteristics across cohorts
Cohort IncidenceCohortIncidenceCalculates incidence rates of outcomes in target populations
Evidence SynthesisEvidenceSynthesisMeta-analysis across data sources

CohortGenerator is always included and cannot be deselected. Modules not installed on the R runtime server are shown as disabled.

tip

For a standard comparative effectiveness study, select CohortGenerator + CohortMethod + CohortDiagnostics + Characterization at minimum. Add PatientLevelPrediction if you also need risk prediction models.

Step 3: Shared Cohorts​

Add cohort definitions that are shared across all analysis modules. Each cohort is assigned a role:

RoleColorPurpose
TargetTealThe primary treatment or exposure cohort
ComparatorGoldThe comparison treatment cohort
OutcomeCrimsonThe clinical event being measured

Use the Add Cohort button to open a searchable picker that lists existing cohort definitions from your Parthenon instance. Already-assigned cohorts are grayed out. Cohorts can be removed with the X button.

warning

Most analysis modules require at least one target cohort. CohortMethod additionally requires at least one comparator and one outcome. The validation step (Step 4) will flag missing cohort assignments.

Step 4: Review & Validate​

A summary card displays the full study package configuration: study name, description, selected modules, and assigned cohorts with their roles.

Click Run Validation to check the analysis specification for configuration issues. The validator returns one of two statuses:

  • Passed (teal) -- the specification is valid and ready for execution.
  • Failed (red) -- issues were found that must be resolved.

Issues are displayed with severity levels (error or warning) and indicate which module is affected.

Step 5: Execute​

Select a Target Data Source from the dropdown (lists all configured CDM sources). Click Execute Study Package to begin.

During execution:

  • A progress bar and animated indicator show that the package is running.
  • The page displays an estimate that execution may take several minutes depending on dataset size and module count.

Upon completion, the result panel shows:

  • Status -- overall execution status
  • Modules Run -- list of successfully executed modules
  • Result Files -- count of output files generated
  • Output Directory -- filesystem path to results
  • Elapsed Time -- total execution duration

Import / Export​

The page header provides two buttons:

  • Import JSON -- load a previously exported Strategus analysis specification from a .json file. Module selections and cohort assignments are restored from the imported spec.
  • Export JSON -- download the current specification as a JSON file named after the study.
tip

Use Import/Export to share study packages between Parthenon instances or to version-control study designs alongside your research protocol documentation.

API Reference​

EndpointMethodDescription
GET /api/v1/strategus/modulesGETList available Strategus modules on the R runtime
GET /api/v1/strategus/sourcesGETList CDM sources available for execution
POST /api/v1/strategus/validatePOSTValidate an analysis specification
POST /api/v1/strategus/executePOSTExecute a study package against a source