Using this catalogue | Design rules that apply to every scheme | Family A — National statistical standards (Australia) | Family B — International statistical standards | Family C — Vaccination & immunisation guidance | Family D — Surveillance-system conventions | Family E — Clinical & developmental staging | Family F — Disease- and research-specific bands | Verifying illustrative schemes

Overview | Why "brood"? | Two population models | Model 1: Pre-aggregated denominator | Multi-vaccine product comparison | Model 2: Cohort -- single time point | Window 2a: Vaccination status at exit | Window 2b: Baseline (pre-entry) vaccination status | Window 2c: Vaccination within 1 year of cohort entry | Window 2d: Post-intervention vaccination | Model 3: Cohort -- birth cohort with person-time | Model 4: Cohort -- time series (interrupted time series analysis) | Stratified time series | Passing to bowerbird::brood_plot() | The denominator_notes requirement | Session info

What clean_the_nest() does | The three data types | Worked examples | Cases (notifiable disease linelist) | Hospital admissions | Vaccination data — long to wide | Birth-cohort studies | Australian Medicare numbers | Structure | Which column to use for linkage? | The drop_eggs argument | What comes next

What corncrake() does | Three ways to get a factor | Synthetic data | User-supplied factors | Uncertainty: two different questions | Deriving a factor instead: the ratio (multiplier) method | Severity anchor: the CFR/IFR-anchor inversion | A worked example, reproducing the paper's own numbers | The bounds invert — and corncrake() handles that for you | A stratified or time-varying reference rate | Rates alongside corrected counts | A note on missing factors

What flyway() does | Synthetic linked data | One call, several event columns | Observed rates, directly | Why a wrapper, not a change to roost() | Feeding straight into corncrake() | Unnamed events

Why mudnester? | The pipeline at a glance | Synthetic data | Step 1 — clean_the_nest(): standardise and prepare | Step 2 — preening(): age categorisation | Step 3 — roost(): aggregate to a time unit | Step 4 — corncrake(): correct for under-ascertainment | Step 5 — molting() + homing(): de-identification | Complete pipeline summary | Further reading

When to use homing() | The basic relink workflow | Controlling the hash column name | Removing the hash after relinking | Partial matches and unmatched records | Security and governance checklist | What comes next

Why de-identify? | What gets removed | The output list | Hash algorithm selection | Controlling which columns are hashed | Adding columns to the removal list | Irreversible de-identification | Hash collisions | What comes next

Why a dual-code strategy? | Conditions detected | Basic usage | The conditions_category summary | No-decimal code format | Including AR-DRG codes | Prefixing output columns | Lean output with drop_eggs | After plumage(): what comes next | Notes for SCPHU practice

The problem preening() solves | Three ways to specify a scheme | 1. Exact name | 2. Filter to a single match | 3. Let list_age_schemes() guide you | The scheme families | Focus tags — cross-cutting filters | Neonatal schemes: age_unit = "days" | Custom schemes | Catch-all bands and the full-lifespan rule | After preening: what comes next

What roost() does | Synthetic data | Available time units | Examples | Monthly counts by pathogen | Epidemiological weeks | Seasons (southern hemisphere) | Biannual — half-year aggregation | Event columns — counting outcomes | Stratified aggregation after preening() | The roost_tbl object | Zero-filling matters | What comes next

Overview | Step 1: Pre-linkage audit with preflight() | Step 2: Medicare checksum validation with check_medicare() | Step 3: Blocking variable construction with flock() | Step 4: Threshold sensitivity analysis with perch() | Step 5: Probabilistic linkage with murmuration() | Step 6: Visualise the score distribution with murmuration_plot() | Step 7: Post-linkage validation | Summary: the complete starling workflow | Session information

Why pre-linkage quality matters | preflight() — structured pre-linkage audit | check_medicare() — Modulus 10 checksum validation | What to do with invalid Medicare numbers | flock() — blocking variable construction | Choosing blocking variables for Australian datasets | Putting it together | Session information

Why the threshold matters | Australian linkage authority benchmarks | Example: running perch() on scored pairs | Example: visualising the score distribution | Using perch_before_linking = TRUE inside murmuration() | Clerical review zone | Session information