Case study

Lost in summarisation

China's new plan for its drug industry has no full English text, so the English-language debate runs on Xinhua's summary instead. We measured 128 English articles about it. Six link to the plan; seven quote a Xinhua reporter's sentence as the plan's own words; and 84 of the 86 that cite its first-in-class target leave out what that target counts.

method -- measuring English-language coverage of China's 15th Five-Year Plan for the pharmaceutical industry against the plan's own text · 2026-10-04 · 15 sources

A sentence nobody translated

"The plan arrives at a pivotal moment as China's pharmaceutical industry shifts rapidly from its traditional identity as a supplier of active pharmaceutical ingredients and a large consumption market toward a new phase defined by the export of innovative drugs." So, according to BioSpace, "an English translation of the plan stated" on September 18th.1BioSpace, "China plans to transform biotech sector into $522B powerhouse by 2030", 18 Sep 2026, live page checked 4 Oct 2026 -- "an English translation of the plan stated". https://www.biospace.com/business/china-plans-to-transform-biotech-sector-into-522b-powerhouse-by-2030 (T3 -- trade press) Biostock, a Swedish life-sciences site, attributed the same words to "an English translation of the document".2Biostock, "China aims for a quarter of the world's first-in-class drugs", 21 Sep 2026 -- "says an English translation of the document". https://biostock.se/en/2026/09/kina-siktar-pa-en-fjardedel-av-varldens-first-in-class-lakemedel/ (T3 -- trade press) Pharmaceutical Technology's report, syndicated to Yahoo Finance, said "the government stated" them "in an English notice".3Robert Barrie, Pharmaceutical Technology, "Beijing unveils blueprint for biomedical dominance by 2030", 21 Sep 2026, as syndicated by Yahoo Finance -- "In an English notice, the government stated". https://finance.yahoo.com/healthcare/articles/beijing-unveils-blueprint-biomedical-dominance-153624619.html (T3 -- trade press, syndicated)

The plan says no such thing. A Xinhua reporter wrote the sentence, in the state news agency's English feature on the plan that day, and it appears in all 20 copies of that feature we found on state and municipal websites.4Xinhua, "China aims to cement its place at forefront of biomedical innovation with new 5-year plan", 18 Sep 2026 -- "The plan arrives at a pivotal moment as China's pharmaceutical industry is shifting rapidly ...". https://english.news.cn/20260918/1400518bcc1f426099496d4efd16b994/c.html Second edition ("China Focus"): https://english.news.cn/20260918/31b0bf89404141278cd0cb40e377e275/c.html (T2 -- state news agency) The plan itself, issued by the Ministry of Industry and Information Technology (MIIT) and nine other bodies as 工信部联规〔2026〕210号, exists in Chinese only.5Ministry of Industry and Information Technology and nine other bodies, 《医药工业发展"十五五"规划》, 工信部联规〔2026〕210号, Box 1 and notes -- "〔〕为5年累计数"; "首创新药(FIC)占全球创新药比例统计范围包含临床Ⅰ期、Ⅱ期、Ⅲ期、新药上市申请(NDA/BLA)、已获批上市药物,不包含临床前、终止研发、撤项的品种". https://www.miit.gov.cn/zwgk/zcwj/wjfb/tz/art/2026/art_9463adac4db040159f6cbf482675eeea.html (T1 -- issuing ministry) MIIT has published no official English text, and we found no unofficial one besides our own.6Searches for an English translation, Exa and Tavily, 3 Oct 2026, including the query "China Pharmaceutical Industry Development 15th Five-Year Plan English translation full text"; query log in the dataset. Our side-by-side text: 15th Five-Year Plan for Pharmaceutical Industry Development. (T4 -- our search, negative result)

That is the condition in which the English-speaking world is discussing a document that sets China's drug-industry targets to 2030. We set out to measure what that discussion rests on.

What 128 articles rest on

We collected every English-language article about the plan we could find between September 17th and October 3rd: 128 pieces, from wire copy and trade press to brokers' notes and LinkedIn posts. For each we fetched the publisher's own HTML, so that we could count links from the page as served rather than from a text extract. We then grouped near-identical texts and matched every sentence against two candidate sources: the plan's Chinese text and Xinhua's English feature.

Lost in summarisation

Of 125 English pieces we could read, six link the plan itself. Twenty copy Xinhua's English summary almost word for word, and seven present a Xinhua reporter's sentence as the plan's own text. The numbers mostly survive; what three targets count does not: 141 of 153 first-in-class claims drop the plan's stage scope, as Xinhua's own feature does.

"The plan arrives at a pivotal moment as China's pharmaceutical industry is shifting rapidly from its traditional identity as a supplier of active pharmaceutical ingredients and a large consumption market, toward a new phase defined by the export of innovative drugs."Xinhua English feature, 18 September 2026

Box 1 verdicts on 594 claims in 102 articles: a blind DeepSeek judge (temperature 0) plus auditable wording rules for kept versus dropped conditions; rule and judge agree on 530 of 567 (93%).

The numbers behind this figure
Lineage lanePieces
link to the plan6
copy Xinhua almost verbatim20
Xinhua's sentence called the plan7
reuse Xinhua sentences16
copy another source10
no shared source69
Box 1 targetPlanClaimsDropped conditionWrongExactUnverifiable
First-in-class share>=25%153141255
Innovative devices>=200 5-yr2417061
Industry revenue>=3.5 tn yuan84392412
First-trial pipelinesworld-leading61041
Innovative-drug growth>=20% avg91110764
100bn-yuan parks202120181
R&D intensity>=10% avg2510213
Billion-dollar products>=58710815
Firms over 10bn507500741
Innovative drug approvalsworld-leading2800280

The result is lopsided. Of the 125 pages we could fetch, six link to the plan on a Chinese government site. None of the six is a wire service, a mainstream outlet or a state-media English site; they are a crypto-news aggregator, a reprint site and its mirror, a trade newsletter, a LinkedIn post and a venture-capital blog.7Pages linking the plan: PANews https://www.panews.io/articles/01a0b20d-2f67-7671-a334-ed9b9d913434; News-Pravda https://news-pravda.com/world/2026/09/18/2602491.html; Telegraph (telegra.ph) https://telegra.ph/Chinas-Next-Industrial-Leap-Is-Medicine-09-18; AllSci https://allsci.com/news/china-pharmaceutical-five-year-plan-chinas-ten/; LinkedIn https://www.linkedin.com/posts/jeremymlevin_biotech-usa-biotechnology-activity-7507888863235715073-pbnJ; LifeSciVC https://lifescivc.com/2026/10/chinese-biopharma-the-next-five-years/ (T4 -- the pages measured) Xinhua's feature, by contrast, was copied almost word for word 20 times on the day of publication, and sentences from it run through at least 16 more pieces.8Copy families and sentence matches: dataset files cluster_raw.csv, pairs.csv, fidelity.csv at https://bigbio.ai/documents/china-15th-fyp-coverage/SOURCE.txt (T4 -- our measurement) Seven pages, spread across six outlets, present one of its sentences as the plan's text.

None of this makes Xinhua's summary inaccurate. On the numbers it is the most careful text in the corpus. Its English feature states six of the plan's ten targets, and five of the six survive with every condition the plan attaches, including the "above designated size" scope of the revenue target that many later reports dropped.4Xinhua, "China aims to cement its place at forefront of biomedical innovation with new 5-year plan", 18 Sep 2026 -- "The plan arrives at a pivotal moment as China's pharmaceutical industry is shifting rapidly ...". https://english.news.cn/20260918/1400518bcc1f426099496d4efd16b994/c.html Second edition ("China Focus"): https://english.news.cn/20260918/31b0bf89404141278cd0cb40e377e275/c.html (T2 -- state news agency) The trouble is not that the summary is wrong. It is that a summary has become the text, and whatever the summary left out has left the conversation.

A quarter of what?

The plan's Box 1 lists ten targets for 2030 and attaches conditions in its notes.5Ministry of Industry and Information Technology and nine other bodies, 《医药工业发展"十五五"规划》, 工信部联规〔2026〕210号, Box 1 and notes -- "〔〕为5年累计数"; "首创新药(FIC)占全球创新药比例统计范围包含临床Ⅰ期、Ⅱ期、Ⅲ期、新药上市申请(NDA/BLA)、已获批上市药物,不包含临床前、终止研发、撤项的品种". https://www.miit.gov.cn/zwgk/zcwj/wjfb/tz/art/2026/art_9463adac4db040159f6cbf482675eeea.html (T1 -- issuing ministry) We extracted every sentence in the corpus that states one of them, 594 claims in 102 articles, and scored each against the plan. For seven of the ten targets, coverage is faithful. Three lose a condition that changes what the number means.

The headline target is the one most often garbled. The plan wants "first-in-class" (FIC) drugs to reach at least 25% of the global total. Its note defines the count: drugs in phase I, II or III trials, under application, or approved, as a share of the world's innovative drugs, excluding anything preclinical, terminated or withdrawn.5Ministry of Industry and Information Technology and nine other bodies, 《医药工业发展"十五五"规划》, 工信部联规〔2026〕210号, Box 1 and notes -- "〔〕为5年累计数"; "首创新药(FIC)占全球创新药比例统计范围包含临床Ⅰ期、Ⅱ期、Ⅲ期、新药上市申请(NDA/BLA)、已获批上市药物,不包含临床前、终止研发、撤项的品种". https://www.miit.gov.cn/zwgk/zcwj/wjfb/tz/art/2026/art_9463adac4db040159f6cbf482675eeea.html (T1 -- issuing ministry) It measures the pipeline. Coverage turned it into "a quarter of the world's first-in-class drugs", which reads as approved medicines. Of the 86 articles that state this target, two keep the scope: China Daily Brief ("drug candidates") and Fierce Pharma ("across clinical programs and approved products").9China Daily Brief, "China 2030 Pharma Plan Targets 3.5T Yuan Revenue", 18 Sep 2026 -- "more than 25 percent of global first-in-class drug candidates". https://chinadailybrief.com/publications/china-pharma-five-year-plan-2030-targets-glw9 Fierce Pharma, "Fierce Pharma Asia", 25 Sep 2026 -- "a 25% global share of first-in-class drugs across clinical programs and approved products". https://www.fiercepharma.com/pharma/lilly-335b-innocare-deal-korea-breakthrough-china-5-year-biotech-plan (T3 -- trade press) Xinhua's feature drops it too, and so does every one of its 20 copies.

The second is industry revenue. The plan's target, at least 3.5trn yuan (about $520bn), covers only firms "above designated size", which China's statistics bureau defines as those with annual revenue of at least 20m yuan.10National Bureau of Statistics of China, FAQ "六、工业统计", question 14 -- "2011年开始至今,统计范围为年主营业务收入2000万元及以上的工业法人单位". http://www.stats.gov.cn/hd/cjwtjd/202302/t20230207_1902277.html (T1 -- national statistics agency) Of the 65 articles that state the target, 27 drop the qualifier. BioSpace went further. Its first version told readers the government wanted the "local biotech industry to grow from about 20 million yuan ($2.98 million) to 3.5 trillion yuan", and headlined the story "China plans to transform $3M biotech sector into $522B powerhouse".11BioSpace, first version, 18 Sep 2026, captured by Exa search on 3 Oct 2026 at its original address -- "grow from about 20 million yuan ($2.98 million) to 3.5 trillion yuan ($521.8 billion)". https://www.biospace.com/business/china-plans-to-transform-3m-biotech-sector-into-522b-powerhouse-by-2030 (T3 -- trade press; original address now redirects) The threshold that defines which firms count had become the size of the industry. BioSpace has since rewritten the page and changed its address, without a correction note; the page still attributes Xinhua's sentence to "an English translation of the plan".1BioSpace, "China plans to transform biotech sector into $522B powerhouse by 2030", 18 Sep 2026, live page checked 4 Oct 2026 -- "an English translation of the plan stated". https://www.biospace.com/business/china-plans-to-transform-biotech-sector-into-522b-powerhouse-by-2030 (T3 -- trade press)

The third is medical devices. The plan marks its target of at least 200 innovative devices as a five-year cumulative figure.5Ministry of Industry and Information Technology and nine other bodies, 《医药工业发展"十五五"规划》, 工信部联规〔2026〕210号, Box 1 and notes -- "〔〕为5年累计数"; "首创新药(FIC)占全球创新药比例统计范围包含临床Ⅰ期、Ⅱ期、Ⅲ期、新药上市申请(NDA/BLA)、已获批上市药物,不包含临床前、终止研发、撤项的品种". https://www.miit.gov.cn/zwgk/zcwj/wjfb/tz/art/2026/art_9463adac4db040159f6cbf482675eeea.html (T1 -- issuing ministry) Twelve of the 16 articles that state it give 200 without the period, which lets a reader take it as a yearly rate, five times the plan's ambition.

Why the summary wins

The likeliest mechanism is mundane. MIIT published no English version, so on September 18th the first English text with an official air was Xinhua's, and its most quotable sentence, the one about "a pivotal moment", reads like a ministry's preamble. A reporter who has not seen the plan can easily mistake one for the other. The record fits that reading: the misattributions quote Xinhua's wording exactly, none of the seven pages links to the plan, and the copies carry the feature's omissions along with its prose.

Nor is this new. When the previous plan, 工信部联规〔2021〕217号, appeared in January 2022, English readers got summaries: briefs from MERICS, a German think-tank, and Trivium China, a consultancy, and an English notice from China's medical-device association that linked back to the Chinese text.12State Council policy library, 工信部联规〔2021〕217号, published 31 Jan 2022. https://www.gov.cn/zhengce/zhengceku/2022-01/31/content_5671480.htm (T1 -- government mirror)13MERICS, "China Industries" brief, February 2022 https://merics.org/sites/default/files/2022-03/MERICS%20February%202022%20China%20Industries.pdf; Trivium China, "Pharma industry: target setting in the shadow of Covid-19", 9 Feb 2022 https://triviumchina.com/2022/02/09/pharma-industry-target-setting-in-the-shadow-of-covid-19/; China Association for Medical Devices Industry, English notice http://en.camdi.cn/news/5045 (T3 -- think-tank and consultancy summaries; T2 -- industry association) We found no full English translation of that plan either.

That matters because the targets already move markets: a Hang Seng biotech fund rose 4.8% in a day on news of the plan.14NewTimeSpace, "CSOP Hang Seng Biotech ETF (03174.HK) Rises 4.84% Intraday after Ten Departments Issue the Pharmaceutical Industry 15th Five-Year Plan", 21 Sep 2026. https://www.newtimespace.com/en/etf/1448403.html (T4 -- market aggregator) Headlines from Stockholm to Hong Kong report that China wants a quarter of the world's first-in-class drugs by 2030.2Biostock, "China aims for a quarter of the world's first-in-class drugs", 21 Sep 2026 -- "says an English translation of the document". https://biostock.se/en/2026/09/kina-siktar-pa-en-fjardedel-av-varldens-first-in-class-lakemedel/ (T3 -- trade press)15South China Morning Post, "How China plans to 'firmly rank at the global forefront' with 25% of first-in-class drugs", 18 Sep 2026. https://www.scmp.com/business/china-business/article/3367952/how-china-plans-firmly-rank-global-forefront-25-first-class-drugs (T3 -- newspaper) The plan's own target is a quarter of the world's clinical-stage and approved pipeline, a different measure, and the argument over which of the two China is chasing has not yet started.

Appendix: method and limits

Collection. Exa web search, 50 queries, window September 17th to October 3rd 2026; Tavily for five pages Exa missed. We kept a piece if it carried at least two sentences about the plan or a headline devoted to it. We left non-English coverage (128 further items) out of scope.

Links. We fetched each page's HTML on October 3rd and classified every hyperlink. A page links to the plan if an anchor targets the MIIT notice, its interpretation, a ministry copy, or a State Council or provincial mirror; we excluded site-registration footers. We could not fetch three pages and so exclude them from the denominator.

Copies and sources. Near-copies: MinHash on five-word shingles, 256 permutations, Jaccard similarity at least 0.5, thresholds fixed before we read the results. Thematic groups: HDBSCAN on bge-m3 document embeddings (model revision 5617a9f6). We matched sentences to the plan with the same multilingual model, and calibrated its threshold (cosine 0.748) on our side-by-side translation of the plan, at a 1% false-positive rate and 87% recall, before scoring any article. We drew the lines in the figure only from text evidence: a copy family, at least two matched sentences, or a framed quotation.

Claims. We extracted candidate sentences by rule for each target (673 sentences, 427 unique). A model from a different family to the one that drafted this article, DeepSeek at temperature zero, judged each unique sentence blind against Box 1 and its conditions. Run twice, it agreed with itself on 93% of claims; its disagreements sat on whether an article kept a condition, so published wording rules, which agree with the model on 530 of 567 claims (93%), make that call. We set one verdict by hand: BioSpace's first version of the revenue target, marked wrong.

Limits. The corpus is what two search engines returned; Reuters, Bloomberg and most mainstream outlets surfaced nothing about the plan, which may reflect the tools as much as the press. We cover LinkedIn and paywalled pages only partly. We count Xinhua's two English editions separately. Our own side-by-side translation is the only full English text of the plan we know of; we used it to calibrate the matching threshold, and it is not part of the corpus.

Dataset

The corpus index, link classification, copy families and every scored claim are published as CSV files, with the scripts that produced them, at /documents/china-15th-fyp-coverage/SOURCE.txt, which lists every file.

Footnotes

  1. BioSpace, "China plans to transform biotech sector into $522B powerhouse by 2030", 18 Sep 2026, live page checked 4 Oct 2026 -- "an English translation of the plan stated". https://www.biospace.com/business/china-plans-to-transform-biotech-sector-into-522b-powerhouse-by-2030 (T3 -- trade press) ↩ ↩2

  2. Biostock, "China aims for a quarter of the world's first-in-class drugs", 21 Sep 2026 -- "says an English translation of the document". https://biostock.se/en/2026/09/kina-siktar-pa-en-fjardedel-av-varldens-first-in-class-lakemedel/ (T3 -- trade press) ↩ ↩2

  3. Robert Barrie, Pharmaceutical Technology, "Beijing unveils blueprint for biomedical dominance by 2030", 21 Sep 2026, as syndicated by Yahoo Finance -- "In an English notice, the government stated". https://finance.yahoo.com/healthcare/articles/beijing-unveils-blueprint-biomedical-dominance-153624619.html (T3 -- trade press, syndicated) ↩

  4. Xinhua, "China aims to cement its place at forefront of biomedical innovation with new 5-year plan", 18 Sep 2026 -- "The plan arrives at a pivotal moment as China's pharmaceutical industry is shifting rapidly ...". https://english.news.cn/20260918/1400518bcc1f426099496d4efd16b994/c.html Second edition ("China Focus"): https://english.news.cn/20260918/31b0bf89404141278cd0cb40e377e275/c.html (T2 -- state news agency) ↩ ↩2

  5. Ministry of Industry and Information Technology and nine other bodies, 《医药工业发展"十五五"规划》, 工信部联规〔2026〕210号, Box 1 and notes -- "〔〕为5年累计数"; "首创新药(FIC)占全球创新药比例统计范围包含临床Ⅰ期、Ⅱ期、Ⅲ期、新药上市申请(NDA/BLA)、已获批上市药物,不包含临床前、终止研发、撤项的品种". https://www.miit.gov.cn/zwgk/zcwj/wjfb/tz/art/2026/art_9463adac4db040159f6cbf482675eeea.html (T1 -- issuing ministry) ↩ ↩2 ↩3 ↩4

  6. Searches for an English translation, Exa and Tavily, 3 Oct 2026, including the query "China Pharmaceutical Industry Development 15th Five-Year Plan English translation full text"; query log in the dataset. Our side-by-side text: 15th Five-Year Plan for Pharmaceutical Industry Development. (T4 -- our search, negative result) ↩

  7. Pages linking the plan: PANews https://www.panews.io/articles/01a0b20d-2f67-7671-a334-ed9b9d913434; News-Pravda https://news-pravda.com/world/2026/09/18/2602491.html; Telegraph (telegra.ph) https://telegra.ph/Chinas-Next-Industrial-Leap-Is-Medicine-09-18; AllSci https://allsci.com/news/china-pharmaceutical-five-year-plan-chinas-ten/; LinkedIn https://www.linkedin.com/posts/jeremymlevin_biotech-usa-biotechnology-activity-7507888863235715073-pbnJ; LifeSciVC https://lifescivc.com/2026/10/chinese-biopharma-the-next-five-years/ (T4 -- the pages measured) ↩

  8. Copy families and sentence matches: dataset files cluster_raw.csv, pairs.csv, fidelity.csv at https://bigbio.ai/documents/china-15th-fyp-coverage/SOURCE.txt (T4 -- our measurement) ↩

  9. China Daily Brief, "China 2030 Pharma Plan Targets 3.5T Yuan Revenue", 18 Sep 2026 -- "more than 25 percent of global first-in-class drug candidates". https://chinadailybrief.com/publications/china-pharma-five-year-plan-2030-targets-glw9 Fierce Pharma, "Fierce Pharma Asia", 25 Sep 2026 -- "a 25% global share of first-in-class drugs across clinical programs and approved products". https://www.fiercepharma.com/pharma/lilly-335b-innocare-deal-korea-breakthrough-china-5-year-biotech-plan (T3 -- trade press) ↩

  10. National Bureau of Statistics of China, FAQ "六、工业统计", question 14 -- "2011年开始至今,统计范围为年主营业务收入2000万元及以上的工业法人单位". http://www.stats.gov.cn/hd/cjwtjd/202302/t20230207_1902277.html (T1 -- national statistics agency) ↩

  11. BioSpace, first version, 18 Sep 2026, captured by Exa search on 3 Oct 2026 at its original address -- "grow from about 20 million yuan ($2.98 million) to 3.5 trillion yuan ($521.8 billion)". https://www.biospace.com/business/china-plans-to-transform-3m-biotech-sector-into-522b-powerhouse-by-2030 (T3 -- trade press; original address now redirects) ↩

  12. State Council policy library, 工信部联规〔2021〕217号, published 31 Jan 2022. https://www.gov.cn/zhengce/zhengceku/2022-01/31/content_5671480.htm (T1 -- government mirror) ↩

  13. MERICS, "China Industries" brief, February 2022 https://merics.org/sites/default/files/2022-03/MERICS%20February%202022%20China%20Industries.pdf; Trivium China, "Pharma industry: target setting in the shadow of Covid-19", 9 Feb 2022 https://triviumchina.com/2022/02/09/pharma-industry-target-setting-in-the-shadow-of-covid-19/; China Association for Medical Devices Industry, English notice http://en.camdi.cn/news/5045 (T3 -- think-tank and consultancy summaries; T2 -- industry association) ↩

  14. NewTimeSpace, "CSOP Hang Seng Biotech ETF (03174.HK) Rises 4.84% Intraday after Ten Departments Issue the Pharmaceutical Industry 15th Five-Year Plan", 21 Sep 2026. https://www.newtimespace.com/en/etf/1448403.html (T4 -- market aggregator) ↩

  15. South China Morning Post, "How China plans to 'firmly rank at the global forefront' with 25% of first-in-class drugs", 18 Sep 2026. https://www.scmp.com/business/china-business/article/3367952/how-china-plans-firmly-rank-global-forefront-25-first-class-drugs (T3 -- newspaper) ↩