What does this package do? | Step-by-step usage | Step 1: Install and load the package | Step 2: Check which registers are needed | Step 3: Prepare the data | Step 4: Run the classification | Step 5 (optional): Collect the results into R | Understanding the output | About stable_inclusion_date vs raw_inclusion_date | Step 6: Saving the results | Working with real register data | Example of end-to-end classification pipeline from Parquet | Click here to show full example | Pre-process input data | Verify duckdb package version | Background population data | Health Service Register: sssy & sysi | Lab data: laboratorieproevesvar | Patient Register: lpr_adm, lpr_diag, lpr_a_kontakt, lpr_a_diagnose | LPR2 | LPR_A | Prescription data: lmdb | Execute osdc and save to disk | Join registers spread across multiple tables | Run the classification | Getting help

Principles | Use cases | Core functionality | Function conventions | Naming | Input | Output | Interface | prepare_lpr*() | classify_diabetes()

General description | High-level flowchart | Classifying type 1 diabetes | Classifying type 2 diabetes | Detailed and technical description | lpr_diag | lpr_adm | diagnoser | kontakter | lab_forsker | lmdb | Across register logic | References

Specific changes since the original validation (version from the paper) | Version 1.0 | Validity | Validity in 2019 | Stratified by diabetes type and age at onset | Bootstrapped metrics | Validity in 2025 | Potential future changes | References

Data required from registers | Expected data structure | Getting access to data

Identifying type 1 and 2 diabetes cases in Danish healthcare registers | Danish register data infrastructure | Current Danish register-based diabetes classifiers | Challenges in current classifiers | Diabetes classification algorithms | References