Pharmacoepidemiology · cumulative exclusions

Evaluation of the washout period for identifying new users in Taiwan's NHIRD

Cumulative exclusion of prevalent users at washout windows ranging from 3 to 120 months, measured across 22 disease cohorts and 182 disease–drug class combinations drawn from the Taiwan National Health Insurance Research Database (2021 index dates).

22 disease cohorts 182 drug-class rows 13 washout windows Index year 2021

Methods

How cumulative exclusions are computed, suppressed, and bounded by drug availability.

Cumulative exclusion at washout T = number of patients with at least one dispensing in the T months preceding the 2021 index date (would be excluded as non-new under a T-month washout). Reported as n (% of 120-month exclusions).

Cells suppressed by the data centre for privacy reasons are filled with the immediately preceding (shorter) washout's value — a conservative lower bound. Cells beyond the drug class's plausible Taiwan NHI availability are left blank instead of extrapolated.

Disease cohorts

Drug-class rows

182

Washout windows

Late-introduction drugs (†)

Interactive explorer

Pick a disease cohort, the drug classes you want to compare, and a lookback range. The chart redraws live.

Cumulative exclusion

% of 120-month exclusions, vs washout period (months). Hover a point for the %.

Your candidate-cohort size for estimating excluded patients (per drug class)

Enter N = patients eligible for the drug. Tooltips and CSV will show estimated excluded ≈ N × frac.

Late-introduction drug classes (†)

Drug classes whose Taiwan NHI reimbursement began after 2011, with the maximum plausible lookback window we can observe for 2021 index dates. Curves for these classes in the explorer truncate at that maximum.

Direct factor Xa inhibitors †

First reimbursed2012-01

Max plausible lookback108 months

Abatacept †

First reimbursed2012-06

Max plausible lookback108 months

Direct thrombin inhibitors †

First reimbursed2012-06

Max plausible lookback108 months

Ivabradine †

First reimbursed2014-05

Max plausible lookback84 months

Janus-kinase inhibitors †

First reimbursed2014-12

Max plausible lookback72 months

Sodium-glucose-cotransporter-2 inhibitors †

First reimbursed2016-05

Max plausible lookback60 months

Interleukin-17 inhibitors †

First reimbursed2016-09

Max plausible lookback48 months

Direct-acting antivirals (HCV) †

First reimbursed2017-01

Max plausible lookback48 months

Sacubitril/valsartan †

First reimbursed2017-03

Max plausible lookback48 months

Tenofovir alafenamide †

First reimbursed2019-05

Max plausible lookback24 months

How to cite

JAMA-style suggested citation. Copy the formatted text, or download a reference file for your reference manager.

Data Infrastructure Working Group, PHDc, NCKU; Tsai DH-T, Lai EC-C. Washout Period Atlas. Population Health Data Center, National Cheng Kung University. Published 2026. Accessed …. https://taiwan-nhird-washout-evaluation.pages.dev/