Late Lessons, Jensen Huang and AI
Markdown version

LL1-16 hindsight check: Twelve late lessons (Editorial team), Late lessons from early warnings (EEA, 2001), Ch. 16, pp. 168–191#

Checked 25 September 2026. This check covers what happened between 2001 and September 2026 that bears on the synthesis chapter’s claims. Page numbers refer to the 2001 report. The chapter is cross-cutting, so the check looks at two things: developments in the cases it cites, and how its framework and recommendations were received, critiqued or taken up in policy. Bracketed numbers such as [S12] refer to the Sources list at the end.

Overview#

The chapter is the editors’ synthesis. It draws on the 14 case studies, but its framework came first (Stirling’s ESTO work, p. 168), and its evidence is second-hand from the case chapters. Twenty-five years on, the lessons fall into three groups.

Held up well, in places strengthened - “No evidence of harm” is not “evidence of no harm” (p. 172). BSE is the flagship example. - EU-wide active testing from 2001 screened about 50 million cattle and found about 7,000 cases. It “confirmed that the clinical surveillance had a poor capacity to detect cases” (Ducrot et al., 2008 [S79]). - The Phillips Inquiry (October 2000) judged the UK “campaign of reassurance” “a mistake” [S76]. - Statistical bodies later warned in general terms against reading non-significance as “no effect” (ASA statement, 2016 [S43]; Amrhein et al., 2019 [S44]). - Dependence on applicants’ own data (p. 179). It persisted, and the EU’s own later reforms acknowledge it: - The 2019 Transparency Regulation keeps the burden of proof on applicants but adds study notification, disclosure and exceptional verification studies [S32]. - The Court of Justice told authorities “not to give in all cases preponderant weight to the studies provided by the applicant” (Blaise, 2019 [S34]). - Independent re-analysis of one industry developmental-neurotoxicity study found effects the sponsor’s summary had not reported (Mie et al., 2018 [S35]). The substance lost EU approval in 2020 [S37] and was listed under the Stockholm Convention in 2025 [S20]. - Institutional variation in precaution rather than an EU–US divide (p. 168). The largest systematic study found no significant average difference over 1970–2004, with a “mixed and diverse pattern” (Hammitt et al., 2005 [S13]; Wiener et al., 2011 [S15]). Vogel (2012 [S16]) qualifies this with a documented shift toward European stringency after about 1990. - Second-generation halocarbon substitutes and illegal production (pp. 175, 177). - The Parties themselves accelerated the HCFC phase-out in 2007, citing climate as well as ozone [S50]. - The Kigali Amendment (2016) then targeted the third-generation substitutes (HFCs) [S52]. - Unreported CFC-11 production after 2012 was detected by atmospheric monitoring, traced to eastern China and largely stopped by 2019 [S55–S58]. It delayed polar ozone recovery by up to about three years [S53]. - Persistence as a warning sign (pp. 170–171). This was strongly vindicated; the post-2001 PFAS experience is the main case in point [S23–S26].

Partly held up, or needing qualification - The pairing of persistence with bioaccumulation. - Lipid-based bioaccumulation criteria proved too narrow for mobile persistent substances. The EU therefore added PMT/vPvM hazard classes in 2023 [S22]. - Some later surprises had nothing to do with persistence. One example is the acutely toxic transformation product of a rubber antioxidant (Tian et al., 2021 [S28]). - No one has tested whether P/B screening made surprises “smaller”. - Innovation and competitiveness (p. 182). - The “weak” claim, that regulation induces innovation, is well supported. - The “competitive edge” claim is mixed. A meta-analysis of 103 studies found the most likely effect statistically insignificant (Cohen and Tubb, 2018 [S64]). A review found induced innovation real but not large enough to offset costs for regulated firms (Dechezleprêtre and Sato, 2017 [S65]). - “Diversity as insurance” (p. 187) remains largely conceptual. - Public reasoning about uncertainty and ignorance (pp. 185–186, 188). - Some later experimental work supports the claim that publics can cope with uncertainty (van der Bles et al., 2020 [S73]). - The Phillips Inquiry’s account of BSE “sedation” supports the claim that reassurance backfires [S76]. - The claim that upstream engagement fixes mistrust has weaker evidence. Engagement advocates themselves later conceded “over-promising” (Stilgoe et al., 2014 [S71]). - The historical claims that rest on hindsight. - The case set contains no false alarms. The 2013 volume’s attempt to fill that gap (Hansen and Tickner [S3]) came from the same network. - Hindsight bias is a known hazard for “could have been deduced” claims (Fischhoff, 1975 [S8]).

Small errors or overstatements (see “Minor factual checks”) - Maternal transmission of BSE “did occur” (p. 172) is stated more firmly than the cohort study supports. - The rendering hypothesis (p. 180) was rejected by the Phillips Inquiry. - The 1974 US reversal on DES (p. 179) was a court ruling on procedure. - I could not independently verify the mid-1970s US scrapie decision (pp. 174, 181).

Reception of the framework - The 2013 second volume reaffirmed that “the 14 case studies and 12 key lessons from the 2001 report remain highly pertinent today”. It also acknowledged a gap: the 2001 volume had no false-positive analysis (EEA 2013 Summary, p. 9 [S2]). - Critics of the precautionary principle have not engaged with the twelve lessons one by one. They attack its logic instead: - Risks lie on all sides, and precaution creates risks of its own (Sunstein, 2005 [S7]). - The transatlantic “precaution gap” is rhetoric rather than reality (Wiener et al., 2011 [S15]). - Many of the chapter’s institutional recommendations were enacted in some form after 2001. The report cannot be credited with causing them: - independent food-safety agencies (EFSA, 2002 [S31]); - a reversed burden of proof and PBT/vPvB criteria for chemicals (REACH, 2006 [S21]); - class-based restrictions meant to prevent like-for-like substitution ([S27], [S41]); - explicit uncertainty guidance (EFSA, 2018 [S48]); - legal protection for whistle-blowers (Directive (EU) 2019/1937 [S42]).

2013 update. Annex 3 of the 2013 volume updates nine of the fourteen 2001 cases: fisheries, benzene, asbestos, PCBs, halocarbons, DES, antimicrobials, MTBE and hormones (pp. 717–734 [S4]). It does not update BSE, radiation, SO2, the Great Lakes or TBT. The halocarbon “update” (p. 728) only summarises WMO’s 2006 Twenty Questions. Its one observation relevant here is that HCFCs were “still increasing”. The annex’s introduction states the cross-cutting lesson that harm “expands over time … to other kinds of harm”, often at exposures “previously considered ‘safe’” (p. 717). Three other parts of the 2013 volume bear directly on this chapter: - Table A2.1 (after Gee, 2009; p. 702) puts “years of substantial inaction” per case: asbestos 101, PCBs c. 100, benzene 81, radiation 65, DES 30–50, antimicrobials 30, SO2 25–55, MTBE 40+, Great Lakes 45+, hormones 16+, halocarbons 10–30, BSE 10–17, TBT 5–30, fisheries “hundreds”. - Table A2.15 (p. 716) says the Phillips Inquiry’s conclusions “do not seem sufficiently rigorous on judging government actions over time”. - Chapters 26 (Grandjean) and 27 (Gee) extend this chapter’s Box 16.1 and Table 16.1. They add explicit error-direction tables and probability bands for strengths of evidence (pp. 635, 658 [S5], [S6]).

Weight for the lens. The chapter’s strongest transferable claims are about mechanisms and have gained support since 2001: - absence of search presented as absence of harm; - dependence on evidence produced by interested parties; - reassurance that undermines implementation and trust; - persistence and irreversibility as reasons for early action; - substitutes that share the hazard they replace.

Its claims about innovation benefits, diversity as insurance, and what publics “correctly” understand are plausible but rest on thinner or partisan evidence and should be cited with that caveat. The chapter’s own hedges are fairer than its reputation, for example “Not all of these cases demonstrate the delaying or distorting effect” (p. 179), and later evidence mostly supports those hedges.

Claim 1. The gap between first credible warning and effective action was “many years or decades, and in some cases over a century”; unequivocal precautionary action remained “relatively scarce” even after the principle was articulated (p. 168)#

Original claim. - For many cases, the gap between problem identification and effective action “was long, certainly many years or decades, and in some cases over a century”. - Even after the need for the precautionary principle was identified in the 1970s and 1980s, “examples of unequivocal precautionary action were relatively scarce” (p. 168). - The chapter offers dates for early warnings: benzene 1897, asbestos 1898, PCB effects on workers by the late 1930s. It says significant restriction came only in the 1960s–70s (p. 179).

What happened since. - The EEA’s own quantification. The 2013 volume’s Table A2.1 (p. 702 [S4]) codes the fourteen cases. Asbestos (101 years), PCBs (c. 100), benzene (81) and radiation (65) are the century-scale cases. Halocarbons (10–30), BSE (10–17) and TBT (5–30) are the shortest. These are the editors’ own codings, not independent measurements. - The pattern continued after 2001 for some of the same agents. - Asbestos (US). The US Environmental Protection Agency issued its first comprehensive ban on ongoing uses of chrysotile asbestos on 28 March 2024 (89 FR 21970 [S9]), 126 years after the 1898 warning the chapter cites. (A 1989 US ban had been largely vacated in 1991; that is background, not re-checked here.) Asbestos illustrates the chapter’s point, but the 2024 rule covers chrysotile only. A “Part 2” evaluation of legacy uses was finalised in December 2024 (Federal Register notice, 3 December 2024 [S9]). - PCBs. A document-based history drawn from litigation discovery reports that the sole US producer’s internal files show knowledge of PCBs’ toxicity and environmental spread decades before production ended in 1977 (Markowitz and Rosner, 2018 [S11]). The authors disclose that they have testified for plaintiffs. - A persistent-chemical parallel outside the 2001 set. An analysis of manufacturers’ internal documents on PFAS reports that producers knew by 1970 that they were “highly toxic when inhaled and moderately toxic when ingested”, “forty years before the public health community” (Gaber, Bero and Woodruff, 2023 [S10]). - Precautionary action became less scarce in the EU after 2001. - The Court of Justice now states that “protective measures may be taken without having to wait until the reality and seriousness of those risks become fully apparent” (Bayer CropScience v Commission, C-499/18 P, 6 May 2021, para. 80 [S49], citing Blaise, 2019 [S34]). - The restrictions on neonicotinoid seed treatments were upheld on that basis. - The EU banned BPA and similarly classified bisphenols in food-contact materials in December 2024 (Regulation (EU) 2024/3190 [S41]). - A group-wide PFAS restriction was proposed in February 2023 [S27]. - None of these was “unequivocal” in the sense of being uncontested, but the claim of scarcity describes 1970–2000 better than it describes the EU after 2001. - Selection and hindsight. Every 2001 case is a confirmed hazard. The 2013 volume responded to this gap with Hansen and Tickner’s review (Ch. 2, pp. 17–25 [S3]): - Of 88 cases that critics had called regulatory false positives, they judged only four genuine: US swine flu (1976), saccharin (1977), food irradiation, and Southern corn leaf blight. - They conclude that “fear of false positives is misplaced” (p. 17). - That review comes from authors within the same precaution-oriented network, and its classification rules are its own. - The broader critique is logical rather than empirical. Precaution applied to one risk creates others, so a principle that says “act on uncertain harm” cannot tell you which harm to prevent (Sunstein, 2005 [S7]). - Hindsight bias inflates judgements that an outcome “could have been deduced” (Fischhoff, 1975 [S8]). The chapter partly guards against this (“It could have been deduced from the outset”, p. 170, is hedged elsewhere), but its gap estimates depend on judging which early signals were “credible”.

Verdict: held up. - As a description of the fourteen cases, the claim stands. For several agents the gap continued well past 2001 (US asbestos 2024; PFAS). - As a generalisation about all hazards, it cannot be tested from a set of confirmed hazards. - The “scarcity” of precautionary action is less true of EU practice after 2001.

Implication for weight. Use the long warning-to-action gap as a well-documented pattern for confirmed hazards, with the EEA’s own year counts clearly attributed. Do not use it as evidence about how often early warnings in general turn out to be right. For that, pair it with the false-positive literature and its critics.

Claim 2. Acceptance of precaution varies between institutions on both sides of the Atlantic, not EU versus North America; the same evidence produced different US and UK decisions on scrapie-affected animals, and DES decisions were timed differently in different countries (pp. 168, 181–182)#

Original claim. - The case histories suggest “different degrees of acceptance of the need for precaution within different institutions in both North America and Europe” (p. 168). - The “same scientific evidence was available to the United States and the United Kingdom in the mid-1970s” on scrapie transmission and a possible scrapie–CJD link. It led the US Department of Agriculture, “but not the UK MAFF”, to decide that scrapie-affected animals should not be used in human or animal foods (p. 181; also p. 174). - DES decisions also varied in timing (p. 181). - Regulatory “sub-cultures” differ within countries too (p. 182).

What happened since. - Systematic comparison supports the core claim. - Hammitt, Wiener, Swedlow, Kall and Zhou (2005 [S13]) built a list of almost 3,000 risks. They randomly sampled 100 and coded relative US/European stringency for every year from 1970 to 2004. They found: “(a) averaging over risks, there is no significant difference in relative precaution over the period”; “(b) … some evidence of a modest shift toward greater relative precaution of European regulation since about 1990”; and “(c) … a diversity of trends across risks”. Overall: “a mixed and diverse pattern”. - Wiener and Rogers (2002 [S14]) found that the EU was more precautionary on hormones in beef, while the US was more precautionary on mad cow disease in blood donations. They conclude that “differences in relative precaution depend more on the context of the particular risk than on broad differences in national regulatory regimes”. - The book-length study (Wiener, Rogers, Hammitt and Sand, 2011 [S15]) describes “the selective application of precaution to particular risks on both sides of the Atlantic”. - A qualification. Vogel (2012 [S16]) argues that between 1960 and 1990 US health, safety and environmental rules “were more stringent”, and that “since around 1990 global regulatory leadership has shifted to Europe”. His cases include ozone depletion, beef hormones and antibiotics in feed, all cases in the 2001 volume. His argument is about the direction of change in prominent cases. Hammitt et al.’s random sample finds that shift only “weakly”. The two are compatible: average parity with a post-1990 tilt in high-salience consumer and environmental issues. - BSE on the US side. The US General Accounting Office (25 January 2002 [S17]) found that the FDA “has not acted promptly to force firms to keep prohibited proteins out of cattle feed” and that its “data on inspections are severely flawed”. It also found that US rules still allowed some cattle central-nervous-system tissue in human food. So US institutions were not uniformly more precautionary on BSE after the 1970s. - The specific mid-1970s US decision. I could not independently confirm a mid-1970s USDA decision barring scrapie-affected animals from human and animal food. The EEA’s 2013 Table A2.15 (p. 716 [S4]) repeats the claim in different words: the US “bans scrapie-infected sheep and goat meats from cattle food chain”. I found no primary source in this pass. - DES timing. Nothing I found contradicts the claim that national DES decisions differed in timing. The 2013 Table A2.1 dates bans to “1971–1985 US, EU, global” (p. 702 [S4]). See “Minor factual checks” on the 1974 US reversal.

Verdict: held up. The best later comparative evidence agrees that relative precaution is risk-specific and institution-specific rather than a transatlantic divide. Vogel’s documented post-1990 European tilt means “EU versus US” is not pure myth for high-profile consumer and environmental issues. The specific mid-1970s US scrapie decision remains unverified here.

Implication for weight. For the lens, the transferable insight is that the same evidence yields different decisions because of institutional mandates, sponsors and cultures, not geography. The US/UK scrapie example should be cited only as the case authors’ account, since it lacks independent verification.

Claim 3. In most cases hazard information existed but was not brought to decision-makers in time, or was discounted (“institutional ignorance”), sometimes because of short-term economic and political interests (pp. 168, 171)#

Original claim. - “Adequate information about potential hazards was available well before decisive regulatory advice was taken, but the information was either not brought to the attention of the appropriate decision-makers early enough, or was discounted” (p. 168). - Some warnings were “effectively ignored … because of short-term economic and political interactions” (asbestos, PCBs, Great Lakes, sulphur dioxide) (p. 168). - “Institutional ignorance”, where information “may be extant in society, but is not available to the decision-makers”, is “exemplified in most of the case studies” (p. 171). - The chapter adds that “Not all of these cases demonstrate the delaying or distorting effect of non-independent sources” (p. 179).

What happened since. - More direct evidence of knowledge withheld or discounted. - For PCBs, discovery in US litigation produced “an enormous trove of previously private” company documents. Markowitz and Rosner (2018 [S11]) use them to argue that the producer knew of toxicity and global environmental spread long before regulators acted. The authors are expert witnesses for plaintiffs. - For PFAS, a persistent-chemical parallel outside the 2001 set, Gaber et al. (2023 [S10]) coded previously secret documents. They found strategies of “suppressing unfavorable research and distorting public discourse”, but no evidence in that archive of funding favourable research. - In the 2013 volume, cases such as tobacco, lead in petrol, vinyl chloride and beryllium add further examples. Its Summary lists among recurring barriers “the power of some stakeholders” and “manufacturing doubt” (pp. 39–41 [S2]). - Evidence that sponsorship shapes conclusions in a sector the chapter covers (pharmaceuticals, via DES). A Cochrane review of 75 methodological studies found that industry-sponsored drug and device studies more often had favourable efficacy results (RR 1.27) and conclusions (RR 1.34). It found “an industry bias that cannot be explained by standard ‘Risk of bias’ assessments” (Lundh et al., 2017 [S12]). - A partial counterweight from the chapter’s flagship case. The Phillips Inquiry examined whether MAFF had leaned toward producers on BSE (Vol. 1, paras 1171–1173, 1189 [S76]). - It concluded: “So far as the policy decisions are concerned we are satisfied that this criticism is without foundation”. It also found that “MAFF never did less, and on occasion did more, than SEAC recommended”. - It found “more force in the argument” only for risk communication. Even there, it noted that Department of Health officials without a producer interest “showed themselves as eager as MAFF” to minimise alarm. - Phillips did confirm a failure to pass information between departments. DH “was not asked to collaborate with MAFF … It should have been”. MAFF advised the Chief Medical Officer only in March 1988 (Exec. Summary §4 [S76]). - The EEA later called Phillips’ conclusions “not sufficiently rigorous” (Table A2.15, p. 716 [S4]). So the BSE case supports “institutional ignorance” (information not reaching the right body) more clearly than “interest-driven discounting”. - Hindsight. The claim that “adequate” information existed is judged after the fact. Fischhoff’s (1975 [S8]) experiments show outcome knowledge inflates judgements of what was foreseeable. This does not refute the claim, but it argues for the chapter’s own caution.

Verdict: held up. Discovered documents have strengthened the “discounted because of economic interests” strand for PCBs, and for the persistent-chemical class more generally. The pure “institutional ignorance” strand (information not reaching decision-makers) is well illustrated by BSE. However, the most thorough official inquiry into BSE rejected producer bias in policy decisions.

Implication for weight. The distinction between information not reaching decision-makers and information deliberately discounted is analytically valuable and holds up. For the lens, keep the two apart. The first calls for information flows and duties to share across departments. The second calls for independence and disclosure (Claim 5). Evidence for the second often comes from litigation and advocates, so attribute it carefully.

Claim 4. Using persistence and bioaccumulation (with novelty, dispersal and irreversibility) as screening proxies will make future “surprises” smaller and less serious (pp. 170–171; Swedish chemicals policy, p. 183)#

Original claim. - For halocarbons, PCBs and MTBE, “their very novelty might be taken as a warning sign”. Their persistence and ready dispersal were further warnings (p. 170). - “If persistence and bioaccumulation are used as screening for eliminating potential hazards, then the size and seriousness of future ‘surprises’ are likely to be smaller” (p. 171). - Global scale, “where there is only one ‘experimental’ model”, also counts (p. 171). - Swedish chemicals policy uses “persistence and bioaccumulation as ‘proxies’ for unknown but possible impacts” (p. 183).

What happened since. - Institutionalisation. - Stockholm Convention. Adopted May 2001, it screens candidates on persistence, bioaccumulation, long-range transport potential and adverse effects. From the original 12 POPs, the Secretariat’s current list names 37 chemicals or groups across Annexes A–C (my count, 25 September 2026 [S19]). - Recent Stockholm listings include PFOA (2019), PFHxS (2022), methoxychlor, Dechlorane Plus and UV-328 (2023). At COP-12 (28 April–9 May 2025), the Parties added chlorpyrifos, long-chain perfluorocarboxylic acids and medium-chain chlorinated paraffins [S20]. - REACH (2006). It adopted PBT/vPvB criteria (Annex XIII) and made such substances candidates for authorisation. It also placed on manufacturers the duty to show their substances do not adversely affect health or the environment [S21] (cited by ELI; not re-read in this pass). - CLP (2023). The EU added new hazard classes: PBT/vPvB, PMT/vPvM (persistent, mobile, toxic; very persistent, very mobile) and endocrine disruption. For endocrine disruptors, classification applies from 1 May 2025 for new substances (Delegated Regulation (EU) 2023/707, 19 December 2022 [S22]). - Persistence vindicated, perhaps more than the chapter claimed. - Cousins et al. (2019 [S23]) use CFCs, PCBs and PFAS as case studies, three of the chapter’s own themes. They show that under continuous release, “continuously increasing contamination” follows “irrespective of the chemical’s physical-chemical properties”. Their conclusion: high persistence alone should suffice for regulation (the “P-sufficient approach”). - Cousins et al. (2020 [S24]) argue that the “high persistence of PFAS is sufficient for their management as a chemical class”. - Five national authorities followed that logic. On 7 February 2023 they proposed restricting the entire PFAS group, about 10,000 substances, “in order to avoid one PFAS be[ing replaced by another]”. As of RIVM’s page (video text published October 2025), ECHA’s committees expected to finish their opinions in spring 2026; I could not verify the outcome [S27]. - At the global scale, Persson et al. (2022 [S25]) and Cousins et al. (2022 [S26]) argue that chemical pollution and PFAS in particular already exceed a “planetary boundary”. This echoes the chapter’s “only one ‘experimental’ model” (p. 171). - The bioaccumulation half proved too narrow. - The PFAS of greatest concern bind to proteins rather than partitioning into fat. Many are highly mobile in water. - Classic lipid-based bioaccumulation criteria therefore did not capture them well. This is why regulators added “mobility” as a hazard property in 2023 [S22]. - The chapter’s MTBE example (persistent and mobile in groundwater) already pointed this way, but its stated screen was persistence and bioaccumulation. - Surprises that these proxies would not catch. A tyre-rubber antioxidant’s quinone transformation product was identified in 2020–21 as the cause of acute coho-salmon deaths in urban streams (Tian et al., 2021 [S28]). It was acutely toxic at environmental concentrations, a hazard that screening on persistence and bioaccumulation was not designed to find. - Where persistent, bioaccumulative chemicals were eliminated, burdens fell, but slowly. WHO/UNEP global human-milk surveys (2000–2010) indicate “a temporal downward trend for PCDDs, PCDFs and PCBs”. Levels nonetheless “are still significantly above those considered toxicologically safe” (van den Berg et al., 2017 [S29]). This supports the chapter’s point about slow reversibility (p. 171). - Speed. Listing under Stockholm remains slow, typically a handful of substances per Conference of the Parties [S20]. That fits the chapter’s warning that regulators react slowly once persistence becomes apparent (p. 171). - Swedish policy. The Swedish Environmental Code, 1998:808, in force from 1999, codified three principles in chapter 2 [S30]: - a general burden of proof on operators (2 kap. 1 §); - precautionary measures “as soon as there is reason to assume” harm (2 kap. 3 §, “så snart det finns skäl att anta”); - a substitution duty (2 kap. 4 §). Sweden’s persistence and bioaccumulation logic reached EU law through REACH Annex XIII.

Verdict: partly held up. - The direction of the lesson is strongly supported, and persistence in particular has become more central to regulation than the chapter proposed. - Bioaccumulation as a paired screen missed the largest persistent-chemical problem of the post-2001 period until criteria were broadened. - Surprises from non-persistent agents continued. - The prediction that surprises would be “smaller” has not been tested.

Implication for weight. For the lens, the robust, technology-neutral version of this lesson is: properties that raise the cost of being wrong (persistence, irreversibility, ubiquity, global scale) justify earlier and stricter action even when specific harms are unknown. The narrower claim that one fixed pair of screening criteria will shrink surprises should not be carried over. Later experience shows that screening criteria themselves need revising as new routes to harm appear, such as mobility and transformation products.

Claim 5. Regulatory appraisal “frequently fails” because risk information is produced and owned by the actors whose products are being assessed; remedies include independent information institutions, independent food agencies, and moving advisory committees out of “producer” directorates (pp. 179–180)#

Original claim. - “Regulatory appraisal frequently fails due to the dependence of risk assessment on information produced and owned by the very actors whose products are being assessed” (p. 179). - “Independent sources of risk information are a necessary, if not sufficient, condition” (p. 179). - The chapter recommends “Independent information institutions” with “rights, resources and responsibilities” (pp. 179–180). - It welcomes the shift of Commission advisory committees from producer directorates to the Health and Consumer Directorate, and “independent food agencies” nationally and at EU level (p. 180).

What happened since. - Institutions created. - The UK’s Food Standards Agency began work in 2000, and MAFF was replaced by DEFRA in 2001. This is background; the Phillips Inquiry noted the FSA’s creation (para. 1261 [S76]). - The European Food Safety Authority was established by Regulation (EC) No 178/2002 [S31]. - Reliance on applicant data persisted by design. - The EU’s 2019 Transparency Regulation ((EU) 2019/1381, applicable from 27 March 2021) was “a direct response to a successful European Citizens’ Initiative ‘Stop Glyphosate’” (European Commission [S32]). - It restates the principle that “it is for the applicant … to prove” compliance. It argues that health “and … the environment are better protected where the burden of proof is on the applicant” (recital 19). - It adds safeguards: applicants must notify EFSA of every study they commission, “to guarantee … that companies … do not hold back unfavourable studies” (Art. 32b). Studies are disclosed. In exceptional “serious controversies”, the Commission can commission “verification studies” (recital 25). - The EU thus accepted the chapter’s diagnosis but chose disclosure and verification rather than independent generation of data. - The courts on applicant data. - In Blaise (C-616/17, Grand Chamber, 1 October 2019 [S34]), a French court asked whether impartiality is maintained when tests “are conducted by the applicants alone, who may be biased”. The Court upheld the Regulation. It held that authorities “are of necessity bound to take into account relevant evidence other than the tests, analyses and studies submitted by the applicant”, and “not to give in all cases preponderant weight to the studies provided by the applicant” (paras 93–94). - In Tweedale v EFSA (T-716/14, 7 March 2019 [S38]), the General Court ordered wider public access to industry toxicity studies on glyphosate. - A concrete example of the failure the chapter describes. Mie, Rudén and Grandjean (2018 [S35]) re-examined the raw data of the industry-funded developmental-neurotoxicity studies for chlorpyrifos. They found “treatment-related changes in a brain dimension measure … at all dose levels tested, although not been reported in the original test summary”. They also found design choices that reduced the chance of detecting effects. - In 2019 EFSA reported concerns about genotoxicity and developmental neurotoxicity [S36]. - The EU declined to renew approval in January 2020 (Implementing Regulation (EU) 2020/18 [S37]). - Chlorpyrifos was listed under the Stockholm Convention in 2025 [S20]. - I do not claim that the re-analysis caused the regulatory outcome. - Independent evidence can prevail. EFSA’s 2023 re-evaluation of BPA drew on the wider literature. It cut the tolerable daily intake from 4 µg/kg bw/day (2015) to 0.2 ng/kg bw/day, a 20,000-fold reduction (EFSA CEP Panel, 2023 [S40]). The EU then banned BPA and similarly classified bisphenols in food-contact materials (Regulation (EU) 2024/3190 [S41]). - Independence of the agencies themselves. The European Court of Auditors (Special Report 15/2012 [S33]) examined EASA, ECHA, EFSA and EMA. It concluded “that none of the selected agencies adequately managed the conflict of interest situations”. - Contested outcomes remain. Glyphosate was re-approved for ten years on 28 November 2023. Neither the Standing Committee nor the appeal committee delivered an opinion, so the Commission decided (Implementing Regulation (EU) 2023/2660 [S39]).

Verdict: held up. The diagnosis is confirmed by later official reforms, court rulings and at least one well-documented re-analysis. The remedies were partly adopted: independent agencies, transparency and verification powers. The central dependency, applicant-generated data, was deliberately retained. EU law defends that choice as putting the burden of proof in the right place.

Implication for weight. This is one of the chapter’s most durable lessons. For the lens, the technology-neutral form is: - when the party seeking approval generates and controls the evidence, independence depends on duties to register all studies in advance; - it also depends on disclosure of raw data, and on a funded capacity for independent verification; - separating the assessor organisationally from the promoter is necessary but not sufficient.

Claim 6. Conventional science’s emphasis on avoiding false positives systematically generates false negatives; low-power monitoring and “negative” epidemiological studies give false reassurance; policy should set explicit, graded and potentially asymmetric levels of proof, as in Sweden’s 1973 chemicals law (Box 16.1, p. 170; Table 16.1 and text, p. 184)#

Original claim. - “Traditionally there has been a strong emphasis on avoiding Type I errors. … In effect not being wrong is more important than being safe” (p. 184). - Small samples and exposure misclassification make “associations … more likely to be missed than falsely implicated”. This leads to “a false sense of security from so-called ‘negative’ studies” (p. 184). - “The bias in science towards avoiding false positives inevitably involves generating false negatives” (p. 184). - Decision-making should be “more explicit and systematic about the level of proof needed”. Table 16.1 ranges from “beyond all reasonable doubt” (the Swedish 1973 law, for manufacturers’ evidence of safety) to “scientific suspicion of risk” (the same law, for regulators to act) (Box 16.1, p. 170; Table 16.1, p. 184).

What happened since. - Mainstream statistics moved toward the chapter’s position on non-significance. - The American Statistical Association’s 2016 statement on p-values warned against equating statistical significance with importance or truth (Wasserstein and Lazar, 2016 [S43]). - In 2019, more than 800 signatories backed a Nature comment calling for an end to treating non-significant results as evidence of “no effect” (Amrhein, Greenland and McShane, 2019 [S44]). - Grandjean (2008 [S47]) made the environmental-epidemiology case for weight-of-evidence judgements that “consider what could be known given the opportunities for research”. - But the bias is not only one way. - The replication crisis showed that low power also inflates false positives. “Low power also reduces the likelihood that a statistically significant result reflects a true effect” (Button et al., 2013 [S45]). - Ioannidis (2005 [S46]) argued that in many fields “it is more likely for a research claim to be false than true”. - The 2013 EEA volume itself became more balanced. Grandjean’s Table 26.4 lists sources of error in both directions, including “publication bias towards positive findings” and “post hoc hypothesis” as sources of false positives (p. 635 [S5]). - The 2013 Summary concedes that over-reliance on point significance and simple models “have sometimes led to the production of false positives” (p. 40 [S2]). - Graded and asymmetric evidentiary standards were institutionalised in several places. - Gee (2013, Table 27.2, p. 658 [S6]) expands Table 16.1 into probability bands. “Very strong (90–99 %)” covers beyond reasonable doubt and the Swedish 1973 safety burden. “Moderate (33–65 %)” covers reasonable grounds for concern. “Weak (10–33 %)” covers scientific suspicion, the Swedish 1973 threshold for regulators. The numeric bands are Gee’s own mapping. - Sweden. The 1999 Environmental Code keeps the asymmetry: operators must show compliance (2 kap. 1 §), and precautions apply “as soon as there is reason to assume” harm (2 kap. 3 §) [S30]. - EU pesticides law. The Court of Justice held that where studies are “inconclusive, but the likelihood of real harm … persists … the precautionary principle justifies the adoption of restrictive measures”. It added that the Commission “is not subject to a higher proof requirement” when reviewing an approved substance than an applicant faces when seeking approval (C-499/18 P, paras 80–81 [S49]). - Uncertainty expression. EFSA’s Scientific Committee adopted guidance requiring uncertainty analysis in all assessments (2018 [S48]). It was then applied in the BPA opinion (“a probability of 57–73%” that another endpoint was more sensitive; genotoxicity “Unlikely to Very Unlikely”) [S40]. - The logical critique. Sunstein (2005 [S7]) argues that asymmetric standards favouring action on one risk can raise others. Hansen and Tickner’s own review classes some alleged false positives as “risk-risk trade-offs” [S3]. The chapter does not address this.

Verdict: held up (with a qualification). The claims that “negative” low-power studies give false reassurance, and that standards of proof should be explicit, are now close to mainstream. The asymmetric, graded model has been codified in Swedish and EU law. The stronger claim that science’s bias “inevitably” runs toward false negatives needs qualifying: the replication crisis shows that the same low-power conditions also generate false positives, and the EEA’s 2013 volume says so.

Implication for weight. For the lens, carry forward three things: - “absence of evidence from an under-powered or unperformed search is not reassurance”; - evidentiary thresholds are policy choices that should be stated and justified by the relative costs of each kind of error; - who carries the burden of proof is a design decision.

Do not carry forward the simpler claim that science is biased only toward false negatives.

Claim 7. Second-generation CFC substitutes were “unduly tolerated”, and more benign alternatives were not properly examined; halocarbon smuggling “is threatening the effectiveness of global controls” (pp. 175, 177)#

Original claim. - “The ozone-depleting properties of second-generation CFC substitutes were perhaps also unduly tolerated, simply because of their relatively low impacts when compared with the original substances, and the existence of more benign substitutes or alternative approaches was not properly looked at” (p. 177). - “The smuggling of halocarbons is threatening the effectiveness of global controls” (p. 175). - “Optimistic assumptions” about containment and decommissioning reduced control effectiveness (pp. 174–175).

What happened since. - HCFCs. In 2007 the Parties agreed Decision XIX/6 to accelerate the HCFC phase-out [S50]: - For developed countries: -75% by 2010, -90% by 2015, completion in 2020 (with 0.5% for servicing to 2030). - For developing countries: a freeze in 2013 at the 2009–2010 baseline, then -10% (2015), -35% (2020) and -67.5% (2025), and completion in 2030 (2.5% servicing to 2040). - The Parties were encouraged to select “alternatives to HCFCs that minimize environmental impacts, in particular impacts on climate”. - The acceleration shows the Parties themselves judged the original timetable too lax. - Radiative forcing and equivalent effective chlorine from HCFCs have fallen since 2021, “5 years before the most recent projected decrease” (Western et al., 2024 [S60]). - The next substitutes carried their own problems, supporting “not properly looked at”. - HFCs, which replaced CFCs and HCFCs, are potent greenhouse gases. The Kigali Amendment (2016; in force 2019; 174 parties as of 25 September 2026 [S52]) phases them down. Compliance “is estimated to avoid 0.3–0.5°C of warming by 2100” (WMO/UNEP 2022 Assessment [S53]). - HFC-23, a by-product of HCFC-22 production, is emitted at levels “as much as eight times larger than expected” [S53]. In 2018 its emissions were “higher than at any point in history” despite reported abatement (Stanley et al., 2020 [S61]). - The HFOs now replacing HFCs break down to trifluoroacetic acid (TFA), “a persistent chemical with potential harmful effects”. Its concentrations are “in general significantly below known toxicity limits at present”, but it “require[s] future evaluation due to its persistence” [S53]. - Each substitution in the fluorochemical family solved one problem and exposed another. That is the pattern the chapter describes. - Illegal production and non-compliance. - Montzka et al. (2018 [S55]) reported that the decline in atmospheric CFC-11 slowed by about 50% after 2012. This implied an emissions increase of 13 ± 5 Gg per year “despite reported production being close to zero”. - Rigby et al. (2019 [S56]) traced 40–60% of the increase to eastern mainland China, around Shandong and Hebei. - By 2019, global emissions had fallen back to about the 2008–2012 level (Montzka et al., 2021 [S57]). Park et al. (2021 [S58]) found eastern China’s emissions “returned to pre-2013 levels”. They conclude “that CFC-11 production occurred in eastern China after the mandated global phase-out”. - The 2022 Assessment estimates that unreported production over 2012–2019 delays polar ozone recovery “by up to 3 years” and global recovery by about 1 year. Gaps in observation are “too large to determine whether all unexpected emissions have ceased”. It also reports “unexplained emissions” of CFC-13, CFC-112a, CFC-113a, CFC-114a, CFC-115 and CCl4, some “likely … leaks of feedstocks or by-products” [S53]; see also Western et al. (2023 [S59], metadata only). - The Parties adopted a further decision on “Preventing illegal trade” in 2007 (Decision XIX/12 [S51]). - The regime as a whole. The Montreal Protocol remains effective. Total column ozone is expected to return to 1980 values around 2066 in the Antarctic, 2045 in the Arctic and 2040 for the near-global average [S53]. The next quadrennial assessment is due at the end of 2026 (NOAA CSL [S54]).

Verdict: held up. - The HCFC judgement was effectively endorsed by the Parties’ 2007 acceleration. - The “alternatives not properly examined” point was borne out by the HFC and TFA sequel. - Illegal production proved a real and recurring threat, as the CFC-11 episode showed. - On “threatening the effectiveness of global controls”: the regime detected and contained the largest known episode. Smuggling was a real threat that the regime proved robust enough to absorb, helped by independent atmospheric monitoring.

Implication for weight. For the lens, two strong transferable points: - regulating the worst member of a class can channel substitution into close relatives that share other hazards; - compliance with controls on persistent, globally dispersed agents cannot rely on self-reporting. Independent monitoring of the shared environment was what caught the breach.

Claim 8. Curtailing a hazardous option tends to channel and intensify innovation in alternatives and may give leading countries a competitive edge; meeting needs with several technologies rather than one global near-monopoly reduces the size of surprises (pp. 182, 187)#

Original claim. - “There is an enormous difference between the discouraging of a particular innovatory pathway, and the channelling of innovation into alternative routes”. For asbestos, halocarbons, PCBs and antimicrobials, “curtailment of a particular option may actually serve to foster and intensify innovation in other areas”. “It may also provide a competitive edge to the economies of the countries leading such innovations” (p. 182). - “The size of any future surprises will be smaller if there are several competing technologies … rather than just one, global, near monopoly, as was the case with asbestos, halocarbons and PCBs” (p. 187). - Unpriced harms gave hazardous products “an unjustifiable advantage” that kept superior substitutes out (p. 177).

What happened since. - Porter-hypothesis evidence. The chapter’s claim parallels Porter and van der Linde (1995 [S62]). - Ambec, Cohen, Elgie and Lanoie (2013 [S63]) reported “conflicting evidence” twenty years on. - A meta-analysis of 103 publications and more than 2,000 estimates found “considerable heterogeneity in both the sign and significance”. A positive effect was “more likely at the state, region, or country level”, “although in both cases the most likely scenario is statistical insignificance” (Cohen and Tubb, 2018 [S64]). - Dechezleprêtre and Sato (2017 [S65]) found that “environmental regulations induce innovation in clean technologies, but the resulting benefits do not appear to be large enough to outweigh the costs of regulations for the regulated entities”. They also found short-run adverse effects concentrated in pollution- and energy-intensive sectors that are “small relative to general trends”. - Induced innovation is well supported. National “competitive edge” is possible but not reliably demonstrated. - The EEA’s own follow-up. Hansen and Tickner (2013 [S3]) report that even the four genuine false positives “sparked innovation within industry and within government”. The 2013 Summary asserts “growing evidence that precautionary measures do not stifle innovation, but instead can encourage it” (p. 38 [S2]). Both come from within the precaution-oriented network. - The substitution histories are double-edged. - Curtailing CFCs did channel innovation. However, the channel was mostly within one family of chemistry: HCFCs, then HFCs, then HFOs. Each generation carried a new problem: residual ozone depletion, then climate forcing, then persistent TFA (see Claim 7 [S53]). - This is consistent with the chapter’s warning about substitutes sharing hazardous properties, and with its case for diversity. It also shows that “channelling” alone does not guarantee better outcomes. - Regulators later moved to class-based rules explicitly to block like-for-like substitution: the group PFAS proposal [S27] and the 2024 bisphenols rule covering “other bisphenols … with harmonised classification” [S41]. - The US National Research Council published A Framework to Guide Selection of Chemical Alternatives (2014 [S67]) to formalise alternatives assessment. - Diversity as insurance. Stirling’s framework for analysing diversity (2007 [S66]) makes the concept more rigorous and measurable. I found no empirical test of the claim that portfolio diversity reduced the size of surprises in the case areas. Stirling was a member of the editorial team, so this is development of the editors’ own framework.

Verdict: partly held up. - Supported: the claim that curtailment channels innovation is well supported, and regulators have adopted the “evaluate alternatives” lesson. - Mixed: the claim of a national competitive edge. - Not tested: diversity as insurance remains a reasoned hypothesis. - The halocarbon history shows channelled innovation can stay trapped in a hazardous family.

Implication for weight. For the lens, cite “restriction redirects rather than stops innovation” as moderately supported. “Diversity reduces the size of surprises” is a reasoned design principle, not an empirical finding. Pair it with the substitution-trap evidence, which shows that which alternatives are channelled into matters as much as whether innovation continues.

Claim 9. Institutional efforts to reassure the public by researching known uncertainties are “futile” and feed mistrust; publics correctly distinguish uncertainty from ignorance and focus on purposes and “who benefits”; stakeholder involvement must begin at the framing stage (pp. 185–186, 188)#

Original claim. - Surveys on GMOs “on both sides of the Atlantic (Levy and Derby, 2000; Wynne et al., 2000), indicate that non-experts do make a basically correct distinction between uncertainty and ignorance”. “The issues of what are the driving purposes and who benefits are foremost in people’s minds” (p. 185). - Policy responses that intensify research on known uncertainties to reassure the public “are futile”. They “only feed public mistrust” (p. 185). - “The involvement of stakeholders in regulatory appraisal needs to begin at the beginning” (p. 186). Participation should be “at an early stage, broadly drawn, and carried down to the appropriate local level” (p. 188). - The evidence cited is qualitative and partly co-authored by an editor (Wynne).

What happened since. - Strong case evidence that reassurance backfired (BSE). The Phillips Inquiry (Vol. 1, paras 1176–1190 [S76]) found the following. - Officials “did not trust the public to adopt as sanguine an attitude”. They “followed an approach whose object was sedation”. - “Dispute displaced debate”. - The approach contributed to “lack of diligence in implementing Regulations”. People responsible for enforcement did not regard BSE as a human risk. - The public “felt that the Government had not been telling the truth” (para. 1188). - It also found that “often media critique was pertinent and well informed” and that public concern “proved well-founded” (para. 1190). - This supports the chapter’s claims about reassurance and mistrust. It concerns communication rather than “intensifying research”, however. - Experimental evidence that publics tolerate uncertainty. Van der Bles et al. (2020 [S73]) ran five experiments (n = 5,780), including a field test on the BBC News website. Communicating epistemic uncertainty produced “only a small decrease in trust in numbers and trustworthiness of the source”. This supports the chapter’s view that people can accept uncertainty. It tested quantified uncertainty, though, not “ignorance”. - The deficit model is not simply wrong. A meta-analysis of 193 surveys in 40 countries found “a small positive correlation between general attitudes towards science and general knowledge of scientific facts” (Allum et al., 2008 [S72]). Siegrist’s review (2021 [S74]) concludes that “the importance of trust varies by hazard and respondent group”, and that most studies cannot establish causality. - Upstream engagement: taken up, then tempered. - The UK’s GM Nation? debate (2003) put the idea into practice. Over 36,000 open responses proved “not fully representative”. Wider opinion was “fragmented, with considerable ambivalence coexisting alongside outright opposition” (Pidgeon et al., 2005 [S68]). - The UK research councils adopted a “responsible innovation” framework built on “anticipation, reflexivity, inclusion and responsiveness” (Stilgoe, Owen and Macnaghten, 2013 [S70]). This is a direct institutional descendant of the framing-stage argument. - The same community later wrote that engagement is “necessary but insufficient”, and that its advocates “could be accused of over-promising” (Stilgoe, Lock and Wilsdon, 2014 [S71]). - Wynne himself (2006 [S69]) argued that “previously discredited and supposedly abandoned public deficit explanations of ‘mistrust’ have actually been continually reinvented”. On that view, the institutional failing the chapter describes persisted despite engagement programmes. - A theoretical counterpoint. Collins and Evans (2002 [S75]) argued that extending technical decision rights to publics creates a “Problem of Extension”. They use Wynne’s own case as an example and conclude that their approach “sometimes … argues for more public involvement, sometimes for less”. - Institutional uptake of uncertainty communication. The 2013 Summary records “increasing attention to communicating scientific uncertainty, especially in … food safety” (p. 38 [S2]). EFSA’s 2018 guidance requires uncertainty analysis in all its assessments [S48]. The EU Transparency Regulation calls for “participatory and open dialogue” in risk communication (recital 4 [S32]), but at the risk-communication stage rather than the framing stage.

Verdict: partly held up. - Supported: the claim that institutional reassurance backfires is strongly supported by the chapter’s own flagship case. Experimental work supports the idea that publics can handle disclosed uncertainty. - Not established: the claim that publics “correctly” separate uncertainty from ignorance, and that this is what drives concern, remains thinly evidenced and partly contested. - Mixed: framing-stage engagement was widely adopted in principle, with mixed practical results and candid second thoughts from its proponents. - Overstated: “futile” is too strong as a general rule.

Implication for weight. For the lens: - Moderately strong: reassurance that outruns the evidence erodes both trust and the diligence of those implementing safeguards. - Moderate: publics weigh purposes and beneficiaries, not just probabilities. - Normative, with limited evidence: stakeholder involvement at the framing stage improves outcomes.

Claim 10. On BSE, UK reassurance cited the absence of evidence “when no evidence was actually being sought”; key research started only in 1989 (maternal transmission) and 1996 (scrapie to cattle); no survey of infectious but symptomless cattle had ever been done; avoiding offal in ruminant feed would “at least significantly” have limited BSE and CJD (pp. 172, 178)#

Original claim. - BSE was identified in 1986, but research on maternal transmission “was not initiated until 1989. Ultimately it showed that maternal transmission did occur”. Scrapie-to-cattle experiments “were not begun until 1996” (p. 172). - “No surveys of the number of infectious but asymptomatic cattle entering the food chain have ever been conducted”. UK reassurance cited “the absence of evidence, when no evidence was actually being sought” (p. 172). - “It seems likely that avoiding offal in ruminant feed would have at least significantly limited the scale of the subsequent BSE and CJD problems” (p. 178). - Related claims: - The Southwood committee self-censored on what was “politically feasible” (p. 179). - MAFF took 17 months to inform the Department of Health (p. 180). - Withdrawn 1979 rendering standards might have mattered, though “It is not clear” (p. 180).

What happened since. - The Phillips Inquiry (published 26 October 2000, cited by the chapter as “Phillips et al., 2000”) [S76]. - Reassurance. “The Government did not lie to the public … It is now clear that this campaign of reassurance was a mistake” (Exec. Summary §1). - Southwood. The committee’s “most unlikely” conclusion did not, “as it should have done, make clear the basis for its assessment of risk”. Its caveat “was lost from sight” (§4). The committee “did not wish to raise needless alarm” and “accommodated” concerns about vaccines (para. 1184). This supports the chapter’s self-censorship point. - Research. “After some initial delay, BSE research was adequately funded”. A research coordinator “might have identified … areas where research could profitably have been started earlier”. The first item listed is “experiments to transmit scrapie to cattle to test the scrapie origin assumption” (§15). This vindicates the chapter’s research-delay point, while softer on funding. - Feed. The 1988 ruminant feed ban “had a dramatic effect in reducing to a fraction what had been an escalating rate of infection”, but did not end it. A “period of grace” and cross-contamination in feed mills let thousands more cattle become infected. “A cow can become infected … [from] an amount of infectious tissue as small as a peppercorn”. The animal ban on specified bovine offal “was unenforceable and widely disregarded” (§6). This supports p. 178 and the chapter’s p. 175 point about unrealistic assumptions about slaughterhouse practice. - Origin. “BSE probably originated from a novel source early in the 1970s … The theory that BSE resulted from changes in rendering methods has no validity” (§3). This undercuts the implication at p. 180, which the chapter itself hedged. - Producer bias. See Claim 3: Phillips rejected it for policy decisions. The EEA later judged Phillips “not sufficiently rigorous” (Table A2.15, p. 716 [S4]). - Active surveillance from 2001 did exactly what the chapter said had never been done. - Commission Decision 2000/764/EC (29 November 2000) required rapid BSE testing from 1 January 2001 of risk-group cattle over 30 months. From 1 July 2001 it extended testing to “all bovine animals over 30 months of age subject to normal slaughter for human consumption” [S77]. - Council Decision 2000/766/EC (4 December 2000) suspended feeding processed animal proteins to farmed animals from 1 January 2001. It noted that the testing programme would “identify Member States where BSE recycling via processed animal protein remains a possibility” [S78]. - The results: “Around 7,000 BSE cases were detected through the screening of about 50 million cattle with rapid tests in Europe. It confirmed that the clinical surveillance had a poor capacity to detect cases”. It also found that feed-borne meat-and-bone meal was “the only substantiated route of infection – even after the feed ban” (Ducrot et al., 2008 [S79]). - The epidemic’s end, and residual cases. - UK cases peaked in 1992 at more than 37,000 (Ritchie et al., 2021 [S82]); Phillips counted “over 170,000” animals dead or destroyed by 2000 [S76]. - In 2024, the EU27 and Northern Ireland tested 980,624 cattle and found 3 atypical cases and no classical BSE. The UK reported one classical case, and EFSA notes that UK classical cases in 2018, 2021 and 2024 “provide evidence that bovine animals have been exposed to the C-BSE agent after the reinforcement of the feed ban” (EFSA, 2025 [S80]). - The human toll. - As of July 2021: “232 clinical cases of definite or probable vCJD … worldwide”, including 178 in the UK and 28 in France. UK deaths peaked in 2000, and “the last known UK case … was reported in 2016” (Ritchie et al., 2021 [S82]). - ECDC’s table shows France reported one case in 2021. No EU/EEA cases were reported in 2022 (ECDC, December 2024 [S81]). - The final toll is far below the worst projections of the late 1990s. But it is not settled. Appendix surveys suggest about 1 in 2,000 people in the UK carry abnormal prion protein (493 per million), “with a wide confidence interval”. A survey of supposedly unexposed cohorts also found positives (7 of 29,516), which complicates the interpretation (Ritchie et al., 2021 [S82]; ECDC [S81]). - The first clinical case in a codon-129 heterozygote (2016) leaves open longer incubation in about half the population [S81]. - Maternal transmission. The cohort study begun in July 1989 found “a statistically significant risk difference … of 9.7 per cent and a relative risk of 3.2” for offspring of affected dams (Wilesmith et al., 1997 [S83]). - Donnelly et al. (1997 [S84]) read this as “direct maternal transmission”. - Ferguson et al. (1997 [S84]) showed the results “could be explained by the hypothesis of genetic predisposition”. - Later reviews say only that maternal transmission cannot be excluded as a source of some infections (Ducrot et al., 2008 [S79]). - The chapter’s “Ultimately it showed that maternal transmission did occur” (p. 172) overstates the finding. - The scrapie-to-cattle experiments begun in 1996, and the origin question. - Intracerebral inoculation of cattle with British scrapie pools produced “two distinct disease phenotypes, which were unlike classical BSE” (Konold et al., 2006; 2015 [S85]). - In 2019, transgenic-mouse work found that classical BSE prions can emerge from atypical scrapie isolates (Huor et al., 2019 [S86]). This revived a version of the scrapie-origin hypothesis. - The origin remains unresolved. Phillips expected it “will probably never be known with certainty” [S76].

Verdict: strengthened. - EU active testing after 2001 showed that clinical reporting had missed large numbers of infected animals. That directly confirms the chapter’s “absence of evidence, when no evidence was actually being sought”. - The Phillips Inquiry independently condemned the reassurance strategy and listed the scrapie-to-cattle experiments among research that should have started earlier. - The feed-ban history confirms that removing ruminant material from feed was the decisive control, though enforcement failures delayed its effect. - Two details are overstated: maternal transmission, and the implied role of rendering changes.

Implication for weight. This is the chapter’s best-documented example. For the lens, three technology-neutral lessons carry high weight: - Reassurance is only as good as the search behind it. Without active surveillance, the absence of reported harm says little. - Controls assume a level of compliance that must itself be checked. The offal ban was widely ignored. - Messaging that plays down risk can weaken the diligence of the people enforcing safeguards.

Minor factual checks#

These are not headline claims, but readers using the chapter should know them.

Implications for the section’s transferable insights#

Stated in technology-neutral terms, with the weight that later evidence supports. Page references are to the 2001 chapter.

Strong (confirmed or strengthened since 2001) - Reassurance is only as good as the search behind it (pp. 172, 184). The strongest example is BSE surveillance after 2001, backed by the statistics profession’s warnings about reading non-significance as “no effect” [S43], [S44], [S79]. - Evidence produced and controlled by the party seeking approval is a structural weakness (p. 179). EU law now acknowledges this through study notification, disclosure and verification powers [S32], [S34], [S35]. The fix was partial by design. - Persistence and irreversibility justify earlier action (pp. 170–171). Post-2001 regulation leans even further on persistence than the chapter proposed [S22], [S23], [S27]. - Substitutes in the same family can share the hazard (pp. 173–174, 177). The fluorochemical refrigerant sequence repeated it [S53], and class-based restrictions are the regulatory response [S27], [S41]. - Assumed compliance is not actual compliance (pp. 174–175). Examples are the specified-offal controls [S76] and unreported CFC-11 production [S55]–[S58]. Independent monitoring of outcomes was what exposed both. - Reassurance that outruns the evidence erodes trust and implementation (p. 185, applied to BSE). The Phillips Inquiry’s “sedation” findings support this [S76].

Moderate (supported with qualifications) - Evidentiary thresholds are policy choices that should be stated explicitly, and may be asymmetric (pp. 170, 184). This has been codified in Swedish and EU law [S30], [S49]. However, the “bias runs one way” framing is too simple [S45], [S46]. - Institutions, not nations, set the level of precaution (pp. 168, 181–182) [S13]–[S16]. - Restriction redirects innovation (p. 182) [S63]–[S65]. National competitive advantage is not reliably demonstrated.

Suggestive or normative (cite with caveats) - Diversity of technologies as insurance against surprise (p. 187) [S66]. - Publics correctly distinguishing uncertainty from ignorance (p. 185) [S72]–[S74]. - Framing-stage participation improving outcomes (pp. 186, 188) [S68]–[S71], [S75]. - Enlarging responsibility to cover foreseeable-in-principle unknowns (p. 189). No later test was found.

New from hindsight (not in the chapter) - Independent monitoring of the shared environment as the enforcement backstop. The CFC-11 episode was caught by atmospheric measurement, not by reporting [S55], [S56]. - Screening criteria need revising as new routes to harm emerge. Examples are mobility (PMT/vPvM) and toxic transformation products [S22], [S28]. - The false-positive record matters for credibility. The EEA itself had to fill this gap in 2013 [S3]. A lens built on these lessons should state its error costs in both directions.

Method and access notes#

Sources#

EEA reports

  1. [S1] EEA. Late lessons from early warnings: the precautionary principle 1896–2000. Environmental issue report No 22, 2001. Ch. 16, pp. 168–191. http://www.eea.europa.eu/publications/late-lessons-2001
  2. [S2] EEA. Late lessons from early warnings: science, precaution, innovation — Summary. EEA Report No 1/2013 (January 2013), pp. 9, 38–41. https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-2/@@download/file
  3. [S3] Hansen, S. F., Tickner, J. A. The precautionary principle and false alarms — lessons learned. In EEA Report No 1/2013, Ch. 2, pp. 17–25 and Table 2.3. https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-chapter-2/@@download/file
  4. [S4] EEA Report No 1/2013, Annexes 1–3: Table A2.1 (p. 702), Table A2.7 (p. 708), Table A2.15 (p. 716), Annex 3 introduction (p. 717) and ozone-layer update (p. 728). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-annex-1/@@download/file
  5. [S5] Grandjean, P. Science for precautionary decision-making. EEA Report No 1/2013, Ch. 26, pp. 623–642 (Table 26.4, p. 635; p. 639). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-chapter-26/@@download/file
  6. [S6] Gee, D. More or less precaution? EEA Report No 1/2013, Ch. 27, pp. 643–669 (Table 27.2, p. 658). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-chapter-27/@@download/file

Critiques, method and document-based histories

  1. [S7] Sunstein, C. R. Laws of Fear: Beyond the Precautionary Principle. Cambridge University Press, 2005. https://doi.org/10.1017/CBO9780511790850
  2. [S8] Fischhoff, B. Hindsight ≠ foresight: the effect of outcome knowledge on judgment under uncertainty. J Exp Psychol Hum Percept Perform 1975;1(3):288–299. https://doi.org/10.1037/0096-1523.1.3.288
  3. [S9] US EPA. Asbestos Part 1; Chrysotile Asbestos; Regulation of Certain Conditions of Use Under TSCA. Final rule, Federal Register, 28 March 2024. https://www.federalregister.gov/documents/2024/03/28/2024-05972/asbestos-part-1-chrysotile-asbestos-regulation-of-certain-conditions-of-use-under-the-toxic ; Asbestos Part 2 final supplemental risk evaluation, notice, 3 December 2024. https://www.federalregister.gov/documents/2024/12/03/2024-28285/asbestos-part-2-supplemental-evaluation-including-legacy-uses-and-associated-disposals-risk
  4. [S10] Gaber, N., Bero, L., Woodruff, T. J. The Devil they Knew: chemical documents analysis of industry influence on PFAS science. Annals of Global Health 2023;89(1):37. https://doi.org/10.5334/aogh.4013
  5. [S11] Markowitz, G., Rosner, D. Monsanto, PCBs, and the creation of a “world-wide ecological problem”. J Public Health Policy 2018;39(4):463–540. https://doi.org/10.1057/s41271-018-0146-8
  6. [S12] Lundh, A., Lexchin, J., Mintzes, B., Schroll, J. B., Bero, L. Industry sponsorship and research outcome. Cochrane Database Syst Rev 2017;MR000033. https://doi.org/10.1002/14651858.MR000033.pub3

Transatlantic comparison

  1. [S13] Hammitt, J. K., Wiener, J. B., Swedlow, B., Kall, D., Zhou, Z. Precautionary regulation in Europe and the United States: a quantitative comparison. Risk Analysis 2005;25(5):1215–1228. https://doi.org/10.1111/j.1539-6924.2005.00662.x
  2. [S14] Wiener, J. B., Rogers, M. D. Comparing precaution in the United States and Europe. J Risk Research 2002;5(4):317–349. https://doi.org/10.1080/13669870210153684
  3. [S15] Wiener, J. B., Rogers, M. D., Hammitt, J. K., Sand, P. H. (eds). The Reality of Precaution: Comparing Risk Regulation in the United States and Europe. RFF Press, 2011 (Routledge edition 2013; publisher description read). https://doi.org/10.4324/9781936331802
  4. [S16] Vogel, D. The Politics of Precaution: Regulating Health, Safety, and Environmental Risks in Europe and the United States. Princeton University Press, 2012 (publisher description read). https://doi.org/10.1515/9781400842568
  5. [S17] US General Accounting Office. Mad Cow Disease: Improvements in the Animal Feed Ban and Other Regulatory Areas Would Strengthen U.S. Prevention Efforts. GAO-02-183, 25 January 2002. https://www.gao.gov/products/gao-02-183
  6. [S18] Chemetron Corp. v. US Department of Health, Education & Welfare, 495 F.2d 995 (D.C. Cir., 24 January 1974), decided with Hess & Clark v. FDA. https://www.courtlistener.com/opinion/7360609/chemetron-corp-v-united-states-department-of-health-education-welfare/

Persistent chemicals and screening criteria

  1. [S19] Stockholm Convention Secretariat. All POPs listed in the Stockholm Convention (accessed 25 September 2026). https://chm.pops.int/TheConvention/ThePOPs/AllPOPs/tabid/2509/Default.aspx
  2. [S20] Stockholm Convention Secretariat. The new POPs: listing decisions of COP-9 (2019), COP-10 (2022), COP-11 (2023) and COP-12 (28 April–9 May 2025). https://chm.pops.int/TheConvention/ThePOPs/TheNewPOPs/tabid/2511/Default.aspx
  3. [S21] Regulation (EC) No 1907/2006 (REACH), Art. 1(3), Annex XIII (cited; not re-read in this pass). https://eur-lex.europa.eu/eli/reg/2006/1907/oj
  4. [S22] Commission Delegated Regulation (EU) 2023/707 of 19 December 2022 amending CLP as regards hazard classes (ED, PBT/vPvB, PMT/vPvM). https://eur-lex.europa.eu/eli/reg_del/2023/707/oj
  5. [S23] Cousins, I. T., Ng, C. A., Wang, Z., Scheringer, M. Why is high persistence alone a major cause of concern? Environ Sci: Processes Impacts 2019;21:781–792. https://doi.org/10.1039/c8em00515j
  6. [S24] Cousins, I. T. et al. The high persistence of PFAS is sufficient for their management as a chemical class. Environ Sci: Processes Impacts 2020;22:2307–2312. https://doi.org/10.1039/d0em00355g
  7. [S25] Persson, L. et al. Outside the safe operating space of the planetary boundary for novel entities. Environ Sci Technol 2022;56:1510–1521. https://doi.org/10.1021/acs.est.1c04158
  8. [S26] Cousins, I. T., Johansson, J. H., Salter, M. E., Sha, B., Scheringer, M. Outside the safe operating space of a new planetary boundary for PFAS. Environ Sci Technol 2022;56:11172–11179. https://doi.org/10.1021/acs.est.2c02765
  9. [S27] RIVM (Netherlands). PFAS restriction proposal (proposal published by ECHA on 7 February 2023; page includes text dated October 2025). https://www.rivm.nl/en/pfas/restriction-proposal
  10. [S28] Tian, Z. et al. A ubiquitous tire rubber-derived chemical induces acute mortality in coho salmon. Science 2021;371:185–189. https://doi.org/10.1126/science.abd6951
  11. [S29] van den Berg, M. et al. WHO/UNEP global surveys of PCDDs, PCDFs, PCBs and DDTs in human milk and benefit–risk evaluation of breastfeeding. Arch Toxicol 2017;91:83–96. https://doi.org/10.1007/s00204-016-1802-z
  12. [S30] Miljöbalk (1998:808) [Swedish Environmental Code], ch. 2 §§ 1–4. Sveriges riksdag. https://www.riksdagen.se/sv/dokument-och-lagar/dokument/svensk-forfattningssamling/miljobalk-1998808_sfs-1998-808/

Food and chemicals regulation, independence and data

  1. [S31] Regulation (EC) No 178/2002 of 28 January 2002 (General Food Law; establishing EFSA) (cited; not re-read). https://eur-lex.europa.eu/eli/reg/2002/178/oj
  2. [S32] Regulation (EU) 2019/1381 of 20 June 2019 on the transparency and sustainability of the EU risk assessment in the food chain (recitals 4, 19, 25; Art. 32b; applicable from 27 March 2021). https://eur-lex.europa.eu/eli/reg/2019/1381/oj ; European Commission, “Transparency and sustainability of the EU risk assessment in the food chain”. https://food.ec.europa.eu/horizontal-topics/general-food-law/transparency-and-sustainability-eu-risk-assessment-food-chain_en
  3. [S33] European Court of Auditors. Management of conflict of interest in selected EU agencies. Special Report No 15/2012. https://www.eca.europa.eu/Lists/ECADocuments/SR12_15/SR12_15_EN.PDF
  4. [S34] Court of Justice (Grand Chamber). Blaise and Others, C-616/17, 1 October 2019, EU:C:2019:800 (paras 43–46, 85–95). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62017CJ0616
  5. [S35] Mie, A., Rudén, C., Grandjean, P. Safety of Safety Evaluation of Pesticides: developmental neurotoxicity of chlorpyrifos and chlorpyrifos-methyl. Environmental Health 2018;17:77. https://doi.org/10.1186/s12940-018-0421-y
  6. [S36] EFSA. Statement on the available outcomes of the human health assessment in the context of the pesticides peer review of the active substance chlorpyrifos. EFSA Journal 2019;17(8):5809. https://doi.org/10.2903/j.efsa.2019.5809
  7. [S37] Commission Implementing Regulation (EU) 2020/18 of 10 January 2020 (non-renewal of chlorpyrifos). https://eur-lex.europa.eu/eli/reg_impl/2020/18/oj
  8. [S38] General Court. Tweedale v EFSA, T-716/14, 7 March 2019. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62014TJ0716
  9. [S39] Commission Implementing Regulation (EU) 2023/2660 of 28 November 2023 renewing the approval of glyphosate. https://eur-lex.europa.eu/eli/reg_impl/2023/2660/oj
  10. [S40] EFSA CEP Panel. Re-evaluation of the risks to public health related to the presence of bisphenol A (BPA) in foodstuffs. EFSA Journal 2023;21(4):6857. https://doi.org/10.2903/j.efsa.2023.6857
  11. [S41] Commission Regulation (EU) 2024/3190 of 19 December 2024 on the use of BPA and other bisphenols … in food contact materials. https://eur-lex.europa.eu/eli/reg/2024/3190/oj
  12. [S42] Directive (EU) 2019/1937 of 23 October 2019 on the protection of persons who report breaches of Union law (cited; not re-read). https://eur-lex.europa.eu/eli/dir/2019/1937/oj

Statistics, proof and uncertainty

  1. [S43] Wasserstein, R. L., Lazar, N. A. The ASA statement on p-values: context, process, and purpose. The American Statistician 2016;70(2):129–133. https://doi.org/10.1080/00031305.2016.1154108
  2. [S44] Amrhein, V., Greenland, S., McShane, B. Scientists rise up against statistical significance. Nature 2019;567:305–307. https://doi.org/10.1038/d41586-019-00857-9
  3. [S45] Button, K. S. et al. Power failure: why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci 2013;14:365–376. https://doi.org/10.1038/nrn3475
  4. [S46] Ioannidis, J. P. A. Why most published research findings are false. PLoS Med 2005;2(8):e124. https://doi.org/10.1371/journal.pmed.0020124
  5. [S47] Grandjean, P. Seven deadly sins of environmental epidemiology and the virtues of precaution. Epidemiology 2008;19(1):158–162. https://doi.org/10.1097/EDE.0b013e31815be031
  6. [S48] EFSA Scientific Committee. Guidance on Uncertainty Analysis in Scientific Assessments. EFSA Journal 2018;16(1):5123. https://doi.org/10.2903/j.efsa.2018.5123
  7. [S49] Court of Justice. Bayer CropScience and Bayer v Commission, C-499/18 P, 6 May 2021 (paras 79–81). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62018CJ0499

Ozone layer and halocarbon substitutes

  1. [S50] Montreal Protocol, Decision XIX/6 (2007): accelerated HCFC phase-out. https://ozone.unep.org/treaties/montreal-protocol/meetings/nineteenth-meeting-parties/decisions/decision-xix6-adjustments-montreal-protocol-regard-annex-c-group-i-substances
  2. [S51] Montreal Protocol, Decision XIX/12 (2007): preventing illegal trade in ozone-depleting substances. https://ozone.unep.org/treaties/montreal-protocol/meetings/nineteenth-meeting-parties/decisions/decision-xix12-preventing-illegal-trade-ozone-depleting-substances
  3. [S52] Ozone Secretariat. Kigali Amendment (2016). https://ozone.unep.org/treaties/montreal-protocol/amendments/kigali-amendment-2016-amendment-montreal-protocol-agreed ; Status of ratification (174 parties to Kigali, accessed 25 September 2026). https://ozone.unep.org/all-ratifications
  4. [S53] WMO/UNEP. Scientific Assessment of Ozone Depletion: 2022, Executive Summary (GAW Report No. 278; released January 2023). https://csl.noaa.gov/assessments/ozone/2022/executivesummary/
  5. [S54] NOAA CSL. Scientific Assessment of Ozone Depletion: 2026 (to be released end of 2026). https://csl.noaa.gov/assessments/ozone/2026/
  6. [S55] Montzka, S. A. et al. An unexpected and persistent increase in global emissions of ozone-depleting CFC-11. Nature 2018;557:413–417. https://doi.org/10.1038/s41586-018-0106-2
  7. [S56] Rigby, M. et al. Increase in CFC-11 emissions from eastern China based on atmospheric observations. Nature 2019;569:546–550. https://doi.org/10.1038/s41586-019-1193-4
  8. [S57] Montzka, S. A. et al. A decline in global CFC-11 emissions during 2018–2019. Nature 2021;590:428–432. https://doi.org/10.1038/s41586-021-03260-5
  9. [S58] Park, S. et al. A decline in emissions of CFC-11 and related chemicals from eastern China. Nature 2021;590:433–437. https://doi.org/10.1038/s41586-021-03277-w
  10. [S59] Western, L. M. et al. Global increase of ozone-depleting chlorofluorocarbons from 2010 to 2020. Nature Geoscience 2023;16:309–313 (metadata only). https://doi.org/10.1038/s41561-023-01147-w
  11. [S60] Western, L. M. et al. A decrease in radiative forcing and equivalent effective chlorine from hydrochlorofluorocarbons. Nature Climate Change 2024 (published 11 June 2024). https://doi.org/10.1038/s41558-024-02038-7
  12. [S61] Stanley, K. M. et al. Increase in global emissions of HFC-23 despite near-total expected reductions. Nature Communications 2020;11:397. https://doi.org/10.1038/s41467-019-13899-4

Innovation, competitiveness and alternatives

  1. [S62] Porter, M. E., van der Linde, C. Toward a new conception of the environment–competitiveness relationship. J Econ Perspect 1995;9(4):97–118. https://doi.org/10.1257/jep.9.4.97
  2. [S63] Ambec, S., Cohen, M. A., Elgie, S., Lanoie, P. The Porter Hypothesis at 20. Rev Environ Econ Policy 2013;7(1):2–22. https://doi.org/10.1093/reep/res016
  3. [S64] Cohen, M. A., Tubb, A. The impact of environmental regulation on firm and country competitiveness: a meta-analysis of the Porter Hypothesis. J Assoc Environ Resour Econ 2018;5(2):371–399. https://doi.org/10.1086/695613
  4. [S65] Dechezleprêtre, A., Sato, M. The impacts of environmental regulations on competitiveness. Rev Environ Econ Policy 2017;11(2):183–206. https://doi.org/10.1093/reep/rex013
  5. [S66] Stirling, A. A general framework for analysing diversity in science, technology and society. J R Soc Interface 2007;4:707–719. https://doi.org/10.1098/rsif.2007.0213
  6. [S67] National Research Council. A Framework to Guide Selection of Chemical Alternatives. National Academies Press, 2014. https://doi.org/10.17226/18872

Publics, uncertainty and engagement

  1. [S68] Pidgeon, N. F. et al. Using surveys in public participation processes for risk decision making: the case of the 2003 British GM Nation? public debate. Risk Analysis 2005;25(2):467–479. https://doi.org/10.1111/j.1539-6924.2005.00603.x
  2. [S69] Wynne, B. Public engagement as a means of restoring public trust in science — hitting the notes, but missing the music? Community Genetics 2006;9(3):211–220. https://doi.org/10.1159/000092659
  3. [S70] Stilgoe, J., Owen, R., Macnaghten, P. Developing a framework for responsible innovation. Research Policy 2013;42(9):1568–1580. https://doi.org/10.1016/j.respol.2013.05.008
  4. [S71] Stilgoe, J., Lock, S. J., Wilsdon, J. Why should we promote public engagement with science? Public Understanding of Science 2014;23(1):4–15. https://doi.org/10.1177/0963662513518154
  5. [S72] Allum, N., Sturgis, P., Tabourazi, D., Brunton-Smith, I. Science knowledge and attitudes across cultures: a meta-analysis. Public Understanding of Science 2008;17(1):35–54. https://doi.org/10.1177/0963662506070159
  6. [S73] van der Bles, A. M., van der Linden, S., Freeman, A. L. J., Spiegelhalter, D. J. The effects of communicating uncertainty on public trust in facts and numbers. PNAS 2020;117(14):7672–7683. https://doi.org/10.1073/pnas.1913678117
  7. [S74] Siegrist, M. Trust and risk perception: a critical review of the literature. Risk Analysis 2021;41(3):480–490. https://doi.org/10.1111/risa.13325
  8. [S75] Collins, H. M., Evans, R. The third wave of science studies: studies of expertise and experience. Social Studies of Science 2002;32(2):235–296. https://doi.org/10.1177/0306312702032002003

BSE and vCJD

  1. [S76] The BSE Inquiry (Lord Phillips). Report, Vol. 1, Findings and Conclusions (26 October 2000): Executive Summary §§1, 3–6, 15; ch. 13 paras 1171–1190; ch. 14 para. 1261. Internet Archive copies: https://web.archive.org/web/2003/http://www.bseinquiry.gov.uk/report/volume1/execsum2.htm (and execsum4–7, execsu16; chapt135.htm; chapter14.htm)
  2. [S77] Commission Decision 2000/764/EC of 29 November 2000 on the testing of bovine animals for BSE. https://eur-lex.europa.eu/eli/dec/2000/764/oj
  3. [S78] Council Decision 2000/766/EC of 4 December 2000 concerning certain protection measures with regard to TSEs and the feeding of animal protein. https://eur-lex.europa.eu/eli/dec/2000/766/oj
  4. [S79] Ducrot, C., Arnold, M., de Koeijer, A., Heim, D., Calavas, D. Review on the epidemiology and dynamics of BSE epidemics. Veterinary Research 2008;39(4):15. https://doi.org/10.1051/vetres:2007053
  5. [S80] EFSA. The European Union summary report on surveillance for the presence of TSE in 2024. EFSA Journal 2025;23:9732. https://doi.org/10.2903/j.efsa.2025.9732
  6. [S81] ECDC. Variant Creutzfeldt-Jakob disease: Annual Epidemiological Report for 2022 (December 2024). https://www.ecdc.europa.eu/en/publications-data/variant-creutzfeldt-jakob-disease-annual-epidemiological-report-2022
  7. [S82] Ritchie, D. L., Peden, A. H., Barria, M. A. Variant CJD: reflections a quarter of a century on. Pathogens 2021;10(11):1413. https://doi.org/10.3390/pathogens10111413
  8. [S83] Wilesmith, J. W. et al. A cohort study to examine maternally-associated risk factors for bovine spongiform encephalopathy. Veterinary Record 1997;141:239–243. https://doi.org/10.1136/vr.141.10.239
  9. [S84] Donnelly, C. A. et al. Analysis of dam–calf pairs of BSE cases: confirmation of a maternal risk enhancement. Proc R Soc B 1997;264:1647–1656. https://doi.org/10.1098/rspb.1997.0229 ; Ferguson, N. M. et al. A genetic interpretation of heightened risk of BSE in offspring of affected dams. Proc R Soc B 1997;264:1445–1455. https://doi.org/10.1098/rspb.1997.0201
  10. [S85] Konold, T. et al. Different prion disease phenotypes result from inoculation of cattle with two temporally separated sources of sheep scrapie from Great Britain. BMC Veterinary Research 2006;2:31. https://doi.org/10.1186/1746-6148-2-31 ; Konold, T. et al. Further characterisation … BMC Research Notes 2015;8:312. https://doi.org/10.1186/s13104-015-1260-3
  11. [S86] Huor, A. et al. The emergence of classical BSE from atypical/Nor98 scrapie. PNAS 2019;116(52):26853–26862. https://doi.org/10.1073/pnas.1915737116

Other

  1. [S87] Farman, J. C., Gardiner, B. G., Shanklin, J. D. Large losses of total ozone in Antarctica reveal seasonal ClOx/NOx interaction. Nature 1985;315:207–210 (metadata only). https://doi.org/10.1038/315207a0