Changelog

CohortPrevalence v1.1.0

Feature Enhancements

• Improvements to standardization
  • Allow for standardization by other weighting schemes acs vs 2020 US census.
  • Automatically resolves differences in age groups between standardization references
• Addition of Lead-in Window - the minimum duration of continuous observation required before a patient can be considered at risk and included in the denominator.

Breaking Changes

• Occurrence Mode Removed: Package now uses era pattern exclusively for all analyses. The calculationMode and circeJsonPath parameters have been removed from createTargetCohort() and CohortPrevalenceExperiment$addCohorts(). All SQL files with “_occurrence” pattern have been deleted.
  • Migration: Remove calculationMode and circeJsonPath arguments from any calls to createTargetCohort()
  • All analyses now use interval overlap (era) pattern by default
• CIRCE JSON Validation Removed: Cohort validation against CIRCE patterns is no longer performed. The TargetCohort$validateCirceJson() method has been removed.

Method Changes

• add calcuation types of formal vs rough to determine is using either any occurrence of the disease or the formal start in the lookback period
  • createPrevalenceType() when mode = “rough” then counts based on the cohort_end_date
  • createPrevalenceType() when mode = “formal” then counts based on the cohort_start_date

API Changes

• CohortInfo R6 class has been removed. It has been replaced by two purpose-specific classes:
  • TargetCohort — for the prevalence numerator and incidence target cohort.
  • PopulationCohort — for the denominator population cohort.
• createPrevalenceCohort() has been renamed to createTargetCohort().
• createPopulationCohort() now constructs a PopulationCohort object instead of a CohortInfo object. The function signature is unchanged.
• createTargetCohort() signature simplified: createTargetCohort(cohortId, cohortName) (removed calculationMode and circeJsonPath parameters)
• CohortPrevalenceExperiment$addCohorts() tibble columns simplified: now requires only cohortId and cohortName (removed optional calculationMode and circeJsonPath columns)
• CohortPrevalenceExperiment$setCommonParameters adds minimumObservationLength and useOnlyFirstObservationPeriod as common parameters to the experiment.

CohortPrevalence v1.0.1

Main Bug fix

• add R6 as import for package
• correct bugs in prevalence

CohortPrevalence v1.0.0

Refactor of Sql

• Cohort Types:
  • Era pattern: first occurrence end of continuous observation
  • Refine Prevalence type options
  • point_prevalence –> pn1 and pd1 with a lookback period
  • period_prevalence_pd2 –> pn2 and pd2 with all time in poi
  • period_prevalence_pd3 –> pn2 and pd3 at least 1 day in poi
  • period_prevalence_pd4 –> pn2 and pd4 at least n days in poi

Major New Features

• PrevalenceResults Class
  • Unified results container with dual crude/standardized storage
  • Export/import with manifest.json tracking
  • SQL query audit trail (SHA256 checksums)
  • .addExecutedQuery() for provenance tracking
• Standardization Integration
  • Add stored reference data for census, who and japanese population for standardization
  • functions to build StandardizationReference for ACS on the fly
  • standardizePrevalence() method (modifies in-place, returns self)
  • Separate stdPrev slot preserves crude prevalence
  • Calls internal standardize_prevalence() function (snake_case)
• CohortPrevalenceExperiment Class
  • Multi-dimensional specification builder (fluent builder pattern)
  • Cartesian product expansion: cohorts × types × constraints × periods of interest
  • Expansion via tidyr::expand_grid() with automatic analysisId generation
  • define() materializes analysis objects from specifications
  • viewDesign() displays specification (console text or reactable HTML)
  • CSV-exportable specification for reproducibility and version control
  • Full validation gates at each step (fail-fast)
• Drug Output Type Support
  • drugConceptSets slot added to CohortPrevalenceAnalysis
  • Validation: required if “drugs” in outputTypes, otherwise ignored
  • Each item validated as Capr ConceptSetItem with @id and @Name slots
  • Marked experimental with warning in documentation: “Use with caution”
• Unified API Simplification
  • generatePrevalence() accepts only prevalenceAnalysisList (single input)
  • Composition pattern: exp$define() → generatePrevalence()
  • Removed dual-input routing logic (cleaner separation of concerns)
• New Dependencies
  • Imports: rlang, lubridate, digest, jsonlite, checkmate
  • Suggests: shiny, plotly, reactable

Deprecated Features

• All: generateSinglePrevalence(), generateMultiplePrevalence(), runPrevalence() (removed in prior session)

CohortPrevalence v0.1.0

• bug fix pn2 uses events in the current poi. ensure pn1 only uses prior events
• change sql render for pn1

CohortPrevalence v0.0.4

• bug fix span results - see #11
• bug fix pd2

CohortPrevalence v0.0.3

• add demographic constraint to population: limit by age range or specific gender
• fix pd4 bug
• fix prevalence results collect bug
• add rassen incidence calculator
• add meta info to results output

CohortPrevalence v0.0.2

• add ability to specify a multi year span as an analysis: ie 2020-2024
• add ability to specify demographic strata: none, age, gender, race
• Now able to execute multiple prevalence analyses in sequence…run different specs of prevAnalysisObj
• updates to docs, vignettes and minor bug fixes

CohortPrevalence v0.0.1

• Initial Release
• Single Yearly Prevalence functionality following Rassen et al 2018