Derive

The derive module is the main phenotype construction layer. It converts raw UKB fields and multi-source records into analysis-ready covariates, disease outcomes, event dates, timing variables, and follow-up durations.

Scope

Group	Functions
General variables	derive_missing(), derive_covariate(), derive_cut()
Disease sources	derive_selfreport(), derive_hes(), derive_first_occurrence(), derive_cancer_registry(), derive_death_registry()
Case definitions	derive_icd10(), derive_case()
Survival variables	derive_timing(), derive_age(), derive_followup()

Workflow Role

Use derive_missing() early to normalise UKB special missing values. Build source-specific disease variables next, combine them with derive_icd10() or derive_case(), and finish by deriving timing, age, and follow-up variables for downstream association models.

Review Focus

• consistent date parsing across source types;
• unambiguous prevalent and incident case definitions;
• stable naming conventions for status, date, age, timing, and follow-up columns;
• clear warnings when requested source columns are absent.