TL;DR — While organizing Taiwan catastrophic illness data for cross-border work, I found that the problem was not a complete absence of English government information. The information was spread across explanatory pages and downloadable tables, so finding the right condition, interpreting it, and sharing it still took work. The official table already includes English disease names. My tool adds a Chinese-English-Japanese search entry point on top of that foundation; it does not define eligibility or replace an official determination.

The original plan and the direction the project later took were quite different.

In 2024, a Taiwanese medical technology company wanted to work through Sunpoint International to bring its cell-factory construction and operating services to Japan. The idea was to combine facility design, process validation, equipment coordination, and construction experience into a deliverable system. That was the proposed collaboration at the time. Completing a facility would not, by itself, secure authorization for a treatment.

After more than a year of exploration, the two sides shifted direction in mid-2025 and began planning a Japanese-affiliated subsidiary in Taiwan to pursue the local introduction of a cell-therapy approach for a hard-to-treat disease. Sunpoint handled coordination in Taiwan. Chairman Jerry Lee was the main contact, and I participated as Chief Strategy Officer and partner. This is a record of the work process, not a statement that the approach has been approved, shown effective, or included in NHI reimbursement.

The project’s medical adviser grew up in Japan and works in Japanese and English but does not read Chinese. We therefore often prepared Taiwan institutional information for him: how catastrophic illnesses are classified, which coding version a dataset uses, and how certificate conditions and validity periods should be read. These questions help explain the system, but they do not replace a pharmaceutical application or reimbursement review.

I opened the National Health Insurance Administration’s Catastrophic Illness Categories page and found classification, validity-period, and coding tables. The agency also has an English overview page explaining the application process and copayment exemption system.

One point needed more precise wording when I checked the material again: the official detailed table already includes a column of English disease names. A download link appearing on the Chinese website does not mean the detailed data exists only in Chinese.

At the time, I did not find an official Japanese entry point that met the needs of our work. I would download the file, locate the relevant item, and then reorganize it in English or Japanese for the collaborator. What slowed me down was the path after finding the source: restructuring the information, translating the applicable conditions, and sharing it in a usable form.

Public Data Has a Foundation, but the Interface Can Still Remove a Step

The NHI page lists 30 major categories and provides tables of categories and certificate validity periods along with coding updates. The table checked for this revision is marked as effective January 1, 2025 and uses the 2023 edition of ICD-10-CM/PCS. The page separately lists an updated rare-disease cross-reference, so one table should not be assumed to contain every later update.

The six download groups on the page are not six historical versions of the same table. They consist of two coding cross-reference updates and four versions of the category and validity-period tables. The version, effective date, and cross-reference updates need to be read together.

PDF is convenient for reading, while ODT and ODS can be easier to reorganize. Those are useful foundations. But an openly specified format does not mean a reader already knows which file to choose, and seeing an English disease name does not explain every eligibility condition.

What I wanted to add was an entry point for searching, cross-checking, and sharing. It can help a reader locate an item that still needs to be checked against the source. It cannot reduce all conditions to a disease name or code.

Turning Data into an Interface While Keeping the Source Visible

We increasingly use search, question-answering tools, and AI to find institutional information. A direct answer is convenient for the user, but institutional data still needs its version, source, and applicable scope. A fluent summary cannot be allowed to replace them.

That was the part I cared about most when designing the interface. After finding a disease, a reader should be able to see its conditions, validity period, and source, and distinguish original data from my organization and translation.

The English and Japanese interfaces contain author-organized translations, not authority-approved wording. Applications and medical or legal determinations still need the official text and formal procedure. Making information easier to use should not make its limits disappear.

I have not performed a complete comparison of availability or update speed across countries. I therefore no longer describe Taiwan as being near the top of Asia or claim that another country has already solved the problem. What this article can establish is the barrier I encountered and the small part I tried to address.

ICD Codes, Catastrophic Illness Certificates, and Reimbursement Are Separate Questions

I studied life sciences at university and have a medical laboratory background. That made medical classifications and data tables more familiar to me, but familiarity with a classification structure does not mean that one code can establish a diagnosis, determine eligibility, or authorize reimbursement.

The original article connected my earlier medical work too directly to the current ICD-10 version. Without checking the system and adoption timeline from that period, I should not claim that I necessarily used today’s version then. The relevant point here is my experience reading classification data, not an unverified institutional history.

When reading Taiwan’s data, at least four questions need to remain separate: how a physician makes a diagnosis, what the classification table lists, whether a catastrophic illness certificate is approved after review, and whether a particular treatment meets reimbursement requirements. These questions are related, but none can substitute for another.

According to the NHI’s English overview, after a certificate is approved, care for the catastrophic illness or a related condition can be exempt from the corresponding outpatient or inpatient copayment during the certificate’s validity period. Unrelated conditions continue through the ordinary process. This does not mean every treatment is free, nor can a certificate’s validity period establish that a drug or regenerative-medicine intervention has regulatory approval or reimbursement.

Finding a classification does not mean a catastrophic illness certificate has been issued.

A certificate’s validity period is not a guarantee that a specific treatment will be reimbursed.

Starting with One Table and Adding an Entry Point That Can Return to the Source

I started with catastrophic illness data because I had already spent time reading and organizing it for work. The understanding had previously remained in my own notes and explanations to collaborators. I wanted the next person who encountered the same barrier to have a shorter path.

The Catastrophic Illness Lookup Tool places the classifications I organized into a multilingual interface. Its design supports searching by name or code, filtering by validity period, and sharing a link to an individual item. It is a data-organization tool, not a diagnostic system or an official application portal.

A tool result still needs to be checked against the conditions and original file. This article revision did not revalidate every code, condition, and translation in the tool, nor does it claim that the dataset has remained fully synchronized with every official update. An actual application should use the official version in effect at that time.

Translation provides an entry point for understanding; it does not replace the official text.

The data version and effective date should remain visible with the answer.

Why I Began with Regenerative Medicine

Cross-border work has brought me into more questions about regenerative-medicine institutions. The Regenerative Medicine Act and the Regenerative Medicine Preparations Act were promulgated on June 19, 2024 and took effect on January 1, 2026. Those dates can be checked in the legislative history and the Executive Yuan submission to the Legislative Yuan. The system has entered a new stage, and readers need knowledge entry points that keep rules, sources, and procedures distinct.

Since 2024, I have worked with services related to exosomes and cell therapy and have considered the field from the perspectives of a user, family member, and industry participant. Personal experience is not evidence of efficacy, and a collaboration proposal is not clinical or pharmaceutical authorization. Those boundaries should remain visible alongside the industry’s opportunities.

Life sciences, medical laboratory work, AI, and cross-border collaboration were once separate parts of my experience. They gradually connected around this subject. I want this series to organize what has been useful in my work into material that other people can trace, compare, and use.

Readers interested in organizing data at an individual scale can continue with Personal Health Data Infrastructure: What Can Ten Years of Records Tell Us?. Bringing data together can be useful, but each conclusion it supports still needs to be checked.

Glossary: Formats and Institutions Have Different Boundaries

  • ODT / ODS: OpenDocument text and spreadsheet formats. An open format does not automatically grant an open license to the file’s contents; the terms governing the content need to be checked separately.
  • PDF: A document format commonly used to preserve page layout. A visible table is not necessarily convenient for automated lookup.
  • ICD-10-CM / PCS: Diagnostic and procedure classification systems encountered in this set of Taiwan NHI data. They need to be read with the version identified by the dataset. A disease name, code, and eligibility determination are not interchangeable.
  • Catastrophic illness certificate: A certificate applied for and reviewed under the applicable rules. Its validity period and scope of copayment exemption do not determine whether a particular treatment will be reimbursed.
  • Taiwan’s two regenerative-medicine laws: The two laws address the regulation of regenerative medicine and related preparations. They should not be reduced to the claim that possessing a cell factory authorizes treatment or market entry. This article does not address case-specific application or product classification.
  • Turn-key system: The concept of delivering an integrated system. The actual scope, acceptance criteria, and applicable regulations still need to be defined.
  • GMP: Good Manufacturing Practice. Manufacturing-quality requirements, product authorization, clinical efficacy, and NHI reimbursement are separate validation questions.

The tool and this series add one step for understanding and retrieval. Data can help people prepare better questions, but formal medical, legal, and application decisions still belong to the relevant professionals and competent-authority procedures.

Four-panel comic: What I wanted to add was an entry point for searching, cross-checking, and sharing. Finding a classification does not mean a catastrophic illness certificate has been issued. A certificate's validity period is not a guarantee that a specific treatment will be reimbursed. Translation provides an entry point for understanding; it does not replace the official text. The data version and effective date should remain visible with the answer. Data can help people prepare better questions;

November 2026 revision: This revision checks the official download page, English overview, and detailed table; corrects the claims that detailed information existed only in Chinese and that the six downloads were historical table versions; and separates classification, certification, copayment exemption, and treatment reimbursement. The tool’s code and data were not changed during this article revision, and the official sources remain distinct from the author’s organization and translations.