LL2-02 hindsight check: The precautionary principle and false alarms — lessons learned (Hansen and Tickner), Late lessons from early warnings: science, precaution, innovation (EEA Report No 1/2013), Ch. 2, pp. 17–45#

Checked 25 September 2026. The check covers what happened between publication (January 2013) and September 2026 that bears on the chapter’s claims, evidence, predictions and recommendations. Page numbers are report pages (PDF page minus 2). The chapter summarises and updates Hansen, Krayer von Krauss and Tickner (2007, Risk Analysis), and most of its case analysis dates from 2004–07. Both authors are long-standing advocates of precaution and alternatives assessment; that context matters for how some claims are framed.

Annex 3 note. Not applicable. The chapter is new in the 2013 report and is not one of the 2001 cases updated in Annex 3. Its mobile phone discussion cross-refers to Chapter 21 of the same report (p. 22), which is a separate section.

Access note. The web-search quota for this session was exhausted before this check began. Everything below was therefore verified by fetching primary sources directly (EUR-Lex/EU Publications Office, EFSA, FDA, CDC, GOV.UK/UKHSA, IARC, the US Federal Register) and by checking bibliographic details and abstracts through Crossref and OpenAlex. Items not re-verified are flagged. Some of the 32 “jury still out” cases were not checked at all; they are marked as such in the Claim 1 table.

Overview#

The chapter’s core move is to test critics’ lists of “false positives” (precautionary regulation of risks that turned out not to be real) against a strict definition. Of 88 cases, only four met the definition. The chapter concluded that “fear of false positives is misplaced” (p. 17) and that false positives are “few and far between as compared to false negatives” (pp. 17, 35). Thirteen years on, the record splits as follows.

Vindicated or strengthened - The “jury still out” cases mostly did not turn into false alarms (claim 1). Of the roughly 18 of 32 cases checked, about 12 moved towards harm or towards tighter regulation. Examples: - BPA: EFSA cut the tolerable daily intake 20,000-fold in 2023, and the EU banned BPA in food-contact materials in 2024. - Phthalates, including DEHP: an EU restriction applies from 2020, and DEHP in medical devices has a 2030 sunset. - Endocrine disruptors: the EU adopted criteria in 2017–18 and new hazard classes in 2023. - PFOA (“Teflon”): IARC Group 1 in 2023. - Perchloroethylene: IARC 2A, and a US ban on most uses in 2024. - PBBs: IARC 2A. - Amitrole (“cranberry scare”): EU non-renewal in 2016. - Breast implants: a new lymphoma risk was established. - Acrylamide: EFSA reported a margin-of-exposure concern and the EU set benchmark levels.

Only about three moved clearly towards reassurance: GM food safety, mobile phones (human studies) and the Bt-pollen threat to monarch butterflies. - Organised doubt-manufacturing (claim 6) is now far better documented from internal records than it was in 2013. Examples include sugar-industry research sponsorship, ExxonMobil’s climate communications and PFAS manufacturers’ knowledge of toxicity. - Nitrites were not a false alarm (claim 8). IARC classified processed meat as carcinogenic (Group 1) in 2015. EFSA raised concerns about nitrosamine exposure in 2017 and 2023. The EU cut permitted nitrite additions again in 2023, applying from October 2025. This vindicates the chapter against critics who called nitrites an “unfounded” scare.

Weakened, contested or wrong - The mobile-phone evidence the chapter leaned on (claim 3) has weakened. Large prospective studies and a WHO-commissioned systematic review (2024) found moderate-certainty evidence of no increased brain tumour risk. Animal evidence keeps the question formally open: a 2025 WHO-commissioned review rated as “high certainty” the evidence for gliomas and heart schwannomas in male rats. - False positives as short-lived and narrow (claim 7) is contradicted by the chapter’s own cases: - Saccharin labelling lasted 23 years (1977–2000), and the EPA hazardous-waste listing lasted until December 2010. - Irradiation approvals stalled for about 15 years after 1968, and the EU list has not been extended since 1999. - Cyclamate is still banned in the United States more than 55 years on. - “Concerns about over-regulation are not justified, based on empirical evidence” (claim 2) has not been confirmed by any later independent systematic study. The sample was drawn from critics’ showcase lists. The classification rules were contested (Cox, 2007). Nobody has since estimated both error rates on a common base. - Later peer-reviewed economics shows that some precautionary-type decisions carry large, diffuse, long-lived costs. The clearest example is nuclear phase-outs after Fukushima: about €3–8 billion a year in Germany, mostly from air-pollution mortality. - The chapter’s typology would file such cases as “risk-risk trade-offs” or “too narrow a definition of risk”, not as false positives. That keeps them out of the count by construction. - MMR–autism as a contained, “unregulated alarm” (claim 9) aged badly: - UK MMR coverage recovered until 2016, then fell again. The UK lost measles elimination status twice, most recently on 2024 data (3,681 cases). - Since late 2025 the US CDC’s own web page has said that “vaccines do not cause autism” is “not an evidence-based claim”. - The chapter’s statement that thimerosal was the suspected culprit in MMR is wrong: MMR “do[es] not and never did contain thimerosal” (CDC).

What the chapter could not see - Several regulatory systems institutionalised pieces of its agenda: - alternatives assessment (US National Academies framework, 2014; EU “safe and sustainable by design” Recommendation, 2022); - built-in re-evaluation (EFSA’s re-evaluations of legacy additives; the US TSCA reform of 2016 and the chemical bans that followed in 2024); - court rulings that precaution justifies restriction when studies are inconclusive (CJEU, Bayer CropScience v Commission, 2021). - The opposite current also strengthened. The EU wrote an “innovation principle” into Horizon Europe (2021). In 2025 the US adopted a 10-for-1 deregulation order and an order directing agencies to use “overly precautionary assumptions” only where the law requires it.

Weight for the lens. Treat the chapter as a strong critique of critics’ lists and a strong conceptual lesson that definitions and evidential thresholds decide how many errors of each kind get found. Do not cite it as an empirical estimate of how often precaution errs, or as evidence that false positives are cheap and brief. Its most durable lessons are: - scrutinise claimed false alarms as hard as claimed harms; - absence of evidence of harm is not evidence of safety; - alternatives and graduated measures reduce the cost of being wrong; - build in re-evaluation.

Its weakest are: - the claim that false positives are short-lived and narrow; - the claim that precaution, even when mistaken, reliably sparks innovation; - the generalisation from 88 cases to “over-regulation concerns are not justified”.

Claim 1. Of 88 alleged regulatory false positives, only four meet the definition; about a third are real risks and about a third are “jury still out” (pp. 17, 20–21, 25; Table 2.3 pp. 35–36)#

Original claim. Hansen and Tickner define a regulatory false positive as a case where government regulation was undertaken on a suspected risk and later evidence gives at least “high confidence” (67–95 %) of no harm (p. 18). Applying it to 88 cases taken mostly from critics’ literature, they found four false positives: - US swine flu immunisation (1976); - saccharin labelling (1977); - Southern corn leaf blight planting (1971); - FDA reluctance to approve food irradiation (p. 25).

In the “jury still out” third, “lack of evidence of harm has been misinterpreted as evidence of safety” (p. 21). Table 2.3 lists 32 such cases (pp. 35–36).

What happened since: the 32 “jury still out” cases. I checked about 18. Direction of travel since 2013, most-studied first:

Tally among cases checked. About 12 moved towards harm or regulation, about 3 towards reassurance, and about 3 are unchanged.

This tally has two biases: - I checked cases where I knew of developments, and harm findings generate more regulatory records than reassurance does. - “Towards regulation” is not the same as “harm proven”. BPA shows a live disagreement among competent agencies.

Has any “jury still out” case become a false positive in the chapter’s own terms? GM food is the strongest candidate: it was regulated in the EU, and a national-academy review now finds no substantiated health difference. Whether “no substantiated evidence” reaches the chapter’s “high confidence of no harm” is a judgement call.

Mobile phones would not qualify even if fully exonerated. Most responses were advisory, and the chapter excludes those from the false-positive definition (pp. 18–19).

The four false positives themselves. - Saccharin. FDA states that “more than 30 human studies demonstrated that the results found in rats were irrelevant to humans”. NTP delisted saccharin in 2000. - Food irradiation. EFSA reaffirmed safety in 2011, with residual caveats (see Claim 4). - Swine flu. The 1976 vaccine–Guillain-Barré link was accepted as causal by the Institute of Medicine (2004). The broader hypothesis of a swine-origin pandemic was borne out in 2009 by a different H1N1 virus, which caused an estimated 151,700–575,400 deaths in its first year (Dawood et al., 2012). That does not make 1976 a correct call, but it complicates the “false alarm” story. - Southern corn leaf blight. Nothing new found.

Reception of the method. Cox (2007, Risk Analysis), responding to the 2007 paper, argued that the criteria treat “highly uncertain risks as ‘real’ even when no real or potential harm has actually been demonstrated”. Examples include upper-bound regulatory assumptions and “claimed (sometimes ambiguous) epidemiological associations”. He argued these criteria give “an alternative possible explanation for why the authors failed to find more false positives”. The chapter cites Cox (p. 33) but does not engage with the substance.

Citation tracing (OpenAlex) finds the 2007 paper cited about 23 times and the 2013 chapter rarely cited as a separate item. I found no later independent replication of the 88-case classification.

Verdict: held up (directionally). Since 2013 the “jury still out” cases have mostly moved towards harm or towards more regulation, not towards vindicating the critics. The headline count (“four”) still rests on a contested, asymmetric definition, and at least one case (GM food safety) arguably now meets the false-positive bar.

Weight. Strong support for the insight that claimed false alarms deserve the same scrutiny as claimed harms, and that “no evidence of harm” is often premature (digest insights 1, 3). Moderate support for “few false positives” as a statement about the critics’ lists. Weak support for it as a statement about precautionary regulation in general (see Claim 2).

Claim 2. “False positives are few and far between as compared to false negatives”; “fear of false positives should not be a rationale for avoiding precautionary actions where warranted” (pp. 17, 35); “common concerns about over-regulation are not justified, based on empirical evidence” (p. 33)#

Original claim. Having found four false positives among 88 alleged ones, the authors conclude that over-regulation concerns are not empirically justified (p. 33). They add that decision-making is structurally biased towards false negatives (p. 34). They concede that classification involves “some level of subjectivity” and that “different researchers might come up with slightly different categorisations” (p. 33).

What happened since. - No systematic, independent re-estimate of error rates exists that I could find. The comparison with false negatives is not a measured ratio. The chapter counts no false negatives, and the Late Lessons volumes are a curated set of false negatives, not a sample. No later study has put both error types on a common denominator. - Comparative studies suggest precaution is selective rather than uniformly excessive or deficient. - Wiener, Rogers, Hammitt and Sand’s The Reality of Precaution (RFF Press, 2011; reviewed in Journal of Risk Research, Nov 2013) compared US and European regulation across many risks. It concluded that neither side is uniformly more precautionary; precaution varies risk by risk [characterised from the book’s standard summaries; the text was not re-fetched]. The chapter does not cite it. - Tosun (2013, Journal of European Public Policy) finds that “political considerations ultimately determine whether and how” the EU invokes precaution. - Garnett and Parsons (2017, Risk Analysis; they cite this chapter) reviewed 15 EU legislative and judicial decisions. Invoking precaution “appears to be poorly defined”, Commission guidance “was not followed consistently in forming legislation”, but courts “tended to be more consistent” in “requiring plausible evidence of potential hazard”. - Löfstedt (2014, Risk Management) called for a formal review of the EU’s precautionary principle Communication.

None of these studies settles whether errors run more one way than the other. - Regulatory cost overestimation, a related over-regulation argument, remains unsettled. Simpson (2014, Journal of Benefit-Cost Analysis) found that most studies report ex ante cost estimates higher than ex post. He argued this does not by itself show bias and “cannot reject the hypothesis that estimates are unbiased”. - Peer-reviewed evidence on the costs of some precautionary-type decisions. These cases fall outside the chapter’s false-positive category, but they matter for its broader claim. - Jarvis, Deschenes and Jha (2022, Journal of the European Economic Association) estimated the social cost of Germany’s post-Fukushima nuclear shutdowns at €3–8 billion a year, “the majority” from increased mortality due to local air pollution from replacement coal generation. “Policymakers would have to significantly overestimate the risk or cost of a nuclear accident to conclude that the benefits of the phase-out exceed its social costs.” - Neidell, Uchida and Veronesi (2021, Journal of Health Economics) found that halting Japanese nuclear generation raised electricity prices, reduced heating and increased cold-weather mortality. An earlier working-paper version was titled “Be Cautious with the Precautionary Principle”.

The chapter treats nuclear power as “too narrow a definition of risk” and argues the US de facto moratorium was not over-regulation (pp. 23–24). Its typology files such cases outside the false-positive count by design. That is defensible, since accident risk is real, but it means the count cannot show the costs of precaution that arise through risk-risk trade-offs. - Official US policy now asserts the opposite empirical premise. Executive Order 14303, “Restoring Gold Standard Science” (23 May 2025), cites examples of alleged worst-case misuse and directs that “highly unlikely and overly precautionary assumptions and scenarios should only be relied upon in agency decision-making where required by law or otherwise pertinent to the agency’s action”. This is an assertion, not a study. It shows the empirical dispute the chapter tried to close is still live in policy.

Verdict: contested. The claim holds as a rebuttal of the critics’ case lists. As a general empirical finding (“concerns about over-regulation are not justified”), it was not warranted by the sample, has not been confirmed by later independent work, and is contradicted in part by later evidence on large, diffuse costs of some precautionary-type decisions. Those decisions mostly sit in categories the chapter excluded from its count. The normative conclusion (fear of false positives alone should not block warranted action) is not refuted, but it rests on argument more than on the data offered.

Weight. Use the chapter to show that evidential thresholds and category definitions decide how many errors of each kind are found (digest insight 2, strong). Do not use it as a quantitative estimate of precaution’s error rate. When applying the lens, look explicitly for risk-risk trade-off costs, which this typology keeps out of view.

Claim 3. Mobile phones are “jury still out”; pooled Nordic and UK Interphone data reported a significant increase in glioma on the side of the head where the phone was used after 10+ years (pp. 21–22)#

Original claim. Most reviews found no hazard after less than 10 years of use, but several said long-term risk could not be excluded. Pooled analyses (Lahkola et al., 2007; Schoemaker et al., 2005) reported increased ipsilateral glioma after 10 or more years. Other reasons for the classification were: no completed studies in children, cellular effects of unknown significance, and “a vast amount of research currently under way” (pp. 21–22). The chapter does not mention IARC’s May 2011 classification of radiofrequency electromagnetic fields as “possibly carcinogenic” (Group 2B; Baan et al., Lancet Oncology, July 2011). That classification was based on “limited” evidence for glioma and acoustic neuroma.

What happened since. - Prospective human evidence has been reassuring. Cohorts avoid the recall bias that affected ipsilateral case-control findings. - UK Million Women Study update (Schüz et al., JNCI, 2022; 776,156 women, 14 years): relative risk for glioma, ever vs never use, 0.89 (0.80–0.99). No association for daily use or 10+ years of use. For temporal and parietal gliomas, the lobes most exposed, relative risks were “slightly below 1.0”. - COSMOS (Feychting et al., Environment International, March 2024; 264,574 participants in five countries): hazard ratio per 100 cumulative call-hours 1.00 (0.98–1.02) for glioma. For 15+ years of use, HR 0.97 (0.62–1.52). - Children and adolescents. The chapter noted no completed studies (p. 21). MOBI-Kids (Castaño-Vinyals et al., Environment International, 2022; 14 countries, 899 cases aged 10–24) found odds ratios that “appeared to decrease” with increasing use. The authors attributed this pattern to bias rather than protection [attribution from the paper’s discussion; abstract truncated]. Hardell and colleagues published a critique (2022). - WHO-commissioned systematic review of human studies (Karipidis et al., Environment International, Sept 2024; GRADE): “moderate certainty evidence” that mobile phone use “likely does not increase the risk of glioma, meningioma, acoustic neuroma, pituitary tumours, and salivary gland tumours in adults, or of paediatric brain tumours”. Evidence for cordless phones was low certainty. - WHO-commissioned systematic review of animal studies (Mevissen et al., Environment International, April 2025; corrigendum 2026): “high” certainty of evidence for increased glioma and heart schwannomas in male rats, based on two chronic bioassays. The US NTP and Ramazzini studies are the likely pair, though I did not confirm which. The review notes that “none of these findings were dose-dependent when compared to the sham controls”. It also notes that the two tumour types match those IARC found “limited evidence” for in humans, and that extrapolation to humans “is particularly complex”. - IARC. No re-evaluation has been held. Radiofrequency radiation was again listed among agents recommended for evaluation in 2025–2029 (Advisory Group, Lancet Oncology, May 2024; priorities poster April 2024). The IARC meetings page showed no scheduled RF meeting in September 2026. - Minority view. Hardell and Carlberg (2021) continue to argue that the evidence shows increased human cancer risk and calls for stronger action. This is not the mainstream assessment.

Verdict: weakened. In 2013, “jury still out” was a reasonable label. The specific evidence the chapter highlighted, the pooled ipsilateral glioma increase, has not been borne out by later prospective and pooled evidence. The mainstream human-evidence assessment now leans to no effect at moderate certainty. The question stays formally open because of the male-rat bioassays and IARC’s pending re-evaluation.

Weight. The case supports the chapter’s methodological point that evidence lags a changing exposure (digest insight 5). It also shows why case-control signals subject to recall bias should not carry much weight. Do not cite it as an example of an emerging harm. It is better used as an example of a long-running signal that weakened under better study designs, with residual uncertainty handled by continued research and monitoring rather than regulation.

Claim 4. Food irradiation is safe for consumers on international consensus (WHO 1981, 1994, 1999; SCF 2003), so FDA reluctance counts as a false positive; consumer resistance is the main barrier; the cyclobutanone genotoxicity concern is doubtful (pp. 29–32)#

Original claim. All international and national evaluations since the early 1980s found irradiated food safe for consumers (p. 29). Only 5 of more than 400 animal studies met 1980 standards (p. 30). Cyclobutanone genotoxicity findings were doubtful because compound identity and purity had not been verified (p. 30). Consumer resistance limited use (p. 29). The authors also argue irradiation “provides an obvious opportunity to cover bad practices” in hygiene (p. 32).

What happened since. - EFSA (2011) reaffirmed safety, with caveats. The chemical safety opinion (EFSA Journal 9(4):1930, April 2011) found: - radiolytic products are mostly “also formed in food that has been subjected to other processing treatments”; - “at least some 2-alkylcyclobutanones may induce DNA damage in vitro”. With no in vivo data, “a genotoxic hazard in humans is considered unlikely by the Panel in view of the plausible indirect mechanism”; - “the only new contrary evidence” was leukoencephalomyelopathy in cats fed highly irradiated feed, whose relevance “should be clarified”; - because “only a very limited quantity of food is irradiated in Europe”, there is “not an immediate cause for concern”.

The microbiological opinion (EFSA Journal 9(4):2103) found “no microbiological risks for the consumer”. It recommended that irradiation “only be used in conjunction with an integrated food safety management program” (good hygiene and manufacturing practice, HACCP). This echoes the chapter’s point that irradiation should not substitute for hygiene. - FDA states it has “evaluated the safety of irradiated food for more than 30 years and has found the process to be safe” (page dated 5 March 2024). Approved uses now include beef and pork, poultry, crustaceans, molluscan shellfish, fresh produce, lettuce and spinach, seeds for sprouting, shell eggs and spices. The Radura label is still required. - EU uptake stayed very low and is falling. The EU-wide positive list under Directive 1999/3/EC “currently includes only one category of food: dried aromatic herbs, spices and vegetable seasonings”, unchanged since 1999; some member states keep national authorisations. Volumes irradiated in the EU were 7,832 tonnes in 2018–19 (down 23.3 % on 2016–17) and 5,029 tonnes in 2020–21 (down 35.8 %). Most was frog legs (76 %), and 83 % was done in Belgium (Commission report COM(2023) 676, 27 Oct 2023). - Consumer resistance. I found no new primary survey data within this check. The falling volumes and unextended EU list are consistent with, but do not prove, continued weak demand.

Verdict: held up. The safety consensus was reaffirmed. The cyclobutanone concern has not disappeared: EFSA accepted in vitro DNA damage but judged human hazard unlikely, so “doubtful” is a fair summary, with a residual caveat. Uptake remains limited, especially in Europe. EFSA’s residual caveats slightly soften “high confidence of no harm” into “high confidence at current, low exposure levels”.

Weight. Useful as the chapter’s clearest false positive and for its secondary lesson: the availability of alternatives (hygiene, sanitation) limited the cost of the error (p. 32; digest insight 8). It is also a case where a precautionary delay, once set, persisted for decades without new adverse evidence (see Claim 7).

Claim 5. Precautionary actions, even unnecessary ones, spark innovation, research and institutional capacity; decision-makers should choose measures that do so (pp. 32–33; Lesson 6, p. 35)#

Original claim. “It appears … that the four false positives identified actually sparked innovation within industry and within government” (p. 32). The chapter cites Ashford and Porter on regulation-induced innovation, research spurred by saccharin and irradiation, US surveillance built after swine flu, and gene-diversity research after corn blight (pp. 32–33). Lesson 6 urges choosing “regulatory measures that can spark innovation even if the precautionary action proves unnecessary” (p. 35).

What happened since. - Evidence on the Porter hypothesis is mixed. - Ambec, Cohen, Elgie and Lanoie (2013, Review of Environmental Economics and Policy) “continue to find conflicting evidence”. - Dechezleprêtre and Sato (2017, same journal) find “evidence that environmental regulations induce innovation in clean technologies”. However, “the resulting benefits do not appear to be large enough to outweigh the costs of regulations for the regulated entities”. Short-run adverse effects are concentrated in pollution- and energy-intensive sectors. - Cohen and Tubb’s (2018) meta-analysis of 103 publications and over 2,000 estimates finds “considerable heterogeneity”. Positive effects are more likely at country level, “although in both cases the most likely scenario is statistical insignificance”. - Rubashkina, Galeotti and Verdolini (2015, Energy Policy) support the “weak” (innovation) version but not the “strong” (productivity) version for European manufacturing [characterised from the paper; abstract not retrieved]. - Evidence on innovation lost to precaution. - Smart, Blum and Wesseler (2016, Journal of Agricultural Economics): mean GE crop approval times of 1,763 days in the EU and 2,467 days in the US (1998–2015). - Wesseler and Zilberman (2014) used a real-options model to infer that delay in approving vitamin-A-enriched rice in India implied perceived costs of at least USD 199 million a year and “about 1.4 million life years lost”. This is a model-based estimate by advocates of the technology and should be treated as indicative. - The chapter’s own mechanism is untested. No study estimates what innovation would have occurred without the four false positives. The saccharin argument (its existence “was a major deterrent” to other sweeteners, pp. 32–33) is a plausible narrative, not a measured effect.

Verdict: partly held up. There is reasonable support for the weaker proposition that regulation can induce innovation in targeted technologies, and that mistaken precaution can leave behind useful research and surveillance capacity. There is no support for a general expectation that precaution, including mistaken precaution, is on balance innovation-positive. Some later work documents real innovation costs from prolonged precautionary gatekeeping.

Weight. Treat Lesson 6 as a design heuristic: prefer measures that open alternatives over bare prohibitions. Do not treat it as an empirical regularity. Digest insight 11 (“suggestive: anecdotal, with no counterfactual”) stands.

Claim 6. Manufacturing doubt, disregarding evidence of risk and claiming over-regulation “appear to be a deliberate strategy for some industry groups and think tanks”; this partly explains why many false positives are claimed but few are genuine (pp. 17, 33–34)#

Original claim. “Several references and leaked documents” show regulated parties recruiting scientists, media experts and politicians “to denounce any risk or to manufacture uncertainty about the risk, regardless of whether the risk is real or not” (pp. 33–34; citing Barnes and Bero, 1998; Rampton and Stauber, 2001; Michaels, 2005).

What happened since: document-based evidence has grown substantially. - Sugar. The Sugar Research Foundation set the objective of a 1967 NEJM review that “singled out fat and cholesterol” and “downplayed evidence” on sucrose, and did not disclose its role (Kearns, Schmidt and Glantz, JAMA Internal Medicine, Sept 2016). A follow-up covered a terminated industry-funded animal study (Kearns et al., PLOS Biology, 2017). - Climate. In ExxonMobil documents 1977–2014, “83% of peer-reviewed papers and 80% of internal documents acknowledge that climate change is real and human-caused, yet only 12% of advertorials do so” (Supran and Oreskes, ERL, 2017). The company’s own projections “accurately forecast warming”, while “public statements about climate science contradicted its own scientific data” (Supran, Rahmstorf and Oreskes, Science, Jan 2023). - PFAS. Companies knew of toxicity “by 1970, forty years before the public health community”, and used “suppressing unfavorable research and distorting public discourse” (Gaber, Bero and Woodruff, Annals of Global Health, 2023). PFAS appear in the chapter’s “Teflon” case. - Synthesis. Goldberg and Vandenberg (2021, Environmental Health) identified 28 doubt-manufacturing tactics across tobacco, coal, sugar, an atrazine manufacturer and the Marshall Institute; five were used by all five. Legg, Hatchard and Gilmore (2021, PLOS ONE) proposed a “Science for Profit Model”. Schäffer et al. (2023, Environmental Science & Technology) summarise established tactics and propose conflict-of-interest safeguards for the new intergovernmental science-policy panel on chemicals, waste and pollution. Björnberg et al. (2017) reviewed 161 articles on environmental science denial. - The explanatory claim remains untested. No study links doubt-manufacturing to the composition of the critics’ false-positive lists. The chapter’s statistical point (few genuine false positives) is equally consistent with Cox’s (2007) alternative: the classification criteria make “real risk” easy to assign. - Organised misinformation is not one-sided. The MMR–autism case (Claim 9) shows sustained promotion of a false alarm, which since 2025 has entered official US health communication.

Verdict: strengthened as a claim that organised doubt-manufacturing exists and has delayed action on real hazards. The explanatory claim, that it partly accounts for the gap between claimed and genuine false positives, is plausible but unproven.

Weight. Strong support for digest insight 10 as a general mechanism: interested parties can shape the evidentiary record and public perception of it. Weight the explanatory use lower. The lens should also watch for organised amplification of unfounded alarms.

Claim 7. Compared with false negatives, false positives’ impacts “may be more short term (over-regulation can be quickly caught) and affect a relatively small number of actors” (p. 34)#

Original claim. Stated without supporting data. The chapter also notes that swine flu was recognised as a false positive “almost immediately after the flu season” (p. 31).

What happened since and what the chapter’s own cases show. - Saccharin. Labelling was imposed in 1977. FDA confirms that the NTP concluded in 2000 that saccharin “should be removed from the list of potential carcinogens”, and warning labels are no longer required. The US EPA removed saccharin from its hazardous-waste and hazardous-substance lists only on 17 December 2010 (75 FR, final rule). That is 23 years to lift the label and 33 years for the last federal listing. - Food irradiation. - FDA: after the 1968 withdrawal, “no additional approvals were given for the next twenty years” (p. 29). Spices came in 1983, poultry in 1990 and red meat in 1997 [the last two dates are standard FDA history, not re-fetched]. Lettuce and spinach, crustaceans, molluscan shellfish and seeds for sprouting are now on FDA’s list. - EU: the harmonised list has stayed at herbs and spices since 1999, and volumes are falling (Claim 4). - A false positive that the chapter itself classes as settled by a 1981 international consensus was still shaping practice 45 years later. - Cyclamate (classed “jury still out”). Still prohibited by FDA more than 55 years after the 1969–70 ban, although permitted in the EU and other jurisdictions. - Diffuse impacts. Jarvis et al. (2022) put most of the €3–8 billion annual social cost of Germany’s nuclear phase-out on population-wide air-pollution mortality, not on a small number of actors. Neidell et al. (2021) found cold-related mortality from higher electricity prices in Japan. The chapter would class these as risk-risk trade-offs, not false positives, but they contradict the general proposition about the scope of precautionary error. - Swine flu. The chapter itself reports deaths, Guillain-Barré cases and more than 4,100 lawsuits (pp. 28–29).

Verdict: weakened. The chapter’s own false positives include some of the longest-lived regulatory errors in the record. Later evidence on precautionary-type decisions shows costs spread across whole populations. Swine flu fits the claim; saccharin and irradiation do not.

Weight. Do not use this proposition in the lens. The opposite mechanism is better supported: once a precautionary measure or label is set, reversal can take decades. Institutional inertia works on both error types. This strengthens the chapter’s own Lesson 7 (build in re-evaluation).

Claim 8. The nitrite response (1978 USDA reduction plus ascorbate/erythorbate, with monitoring) was “smart management of risks, rather than over-regulation”; residual nitrite was reduced five-fold “without compromising antibotulinal effects” (p. 25)#

Original claim. Lieberman and Kwon had called nitrites one of the “greatest unfounded health scares of recent times”. The chapter responds that the fact that a ban could have created botulism risk “is not evidence that concerns about nitrites were unfounded”. It argues the graduated response (reduction, blocking agents, monitoring, technical help) made bacon nearly free of confirmable nitrosamines within a year (p. 25). The five-fold figure comes from the Institute of Food Technologists (1998).

What happened since. - IARC (October 2015) classified processed meat as carcinogenic to humans (Group 1, colorectal cancer; Bouvard et al., Lancet Oncology, Dec 2015). This rests on processed meat as a whole, not on nitrite alone. IARC had earlier (2010) classified ingested nitrate or nitrite under conditions leading to endogenous nitrosation as Group 2A [earlier classification not re-fetched]. - EFSA re-evaluation (June 2017), as summarised in the recitals of Regulation (EU) 2023/2108: - nitrite ADI 0.07 mg/kg bw/day, exceeded by some groups when all dietary sources are counted; - “some concern for the overall exposure to exogenous nitrosamines at high levels”; - “evidence to link preformed N-nitrosodimethylamine and colorectal cancers and some evidence to link dietary nitrite with gastric cancers and to link the combination of nitrite plus nitrate from processed meat with colorectal cancers”, “in line with” IARC 2015. - EFSA nitrosamines opinion (March 2023): the margin of exposure for carcinogenic nitrosamines at high exposure is “highly likely (98–100% certain) to be less than 10,000 for all age groups, which raises a health concern”. “Meat and meat products” is the main contributor. - EU tightening (Commission Regulation (EU) 2023/2108, 6 October 2023). - Added nitrite in most non-heat-treated and heat-treated meat products falls from 150 mg/kg (expressed as sodium nitrite, about 100 mg/kg nitrite ion) to 80 mg/kg nitrite ion from 9 October 2025. Sterilised meat products drop to 55 mg/kg. New maximum residual limits apply (45 mg/kg nitrite ion for most products). - The recitals frame this as keeping nitrosamines “as low as possible while ensuring microbiological safety”, the same balance the chapter describes. - The Commission’s 2014 and 2016 studies found industry typically added less than the permitted maximum. Denmark had kept stricter national limits.

Verdict: partly held up. - The chapter’s rebuttal of the critics is vindicated: nitrite concerns were not unfounded, and regulators still treat botulism protection as a binding constraint. - The “smart management” framing now looks incomplete. The 1978 measures dealt with preformed nitrosamines in bacon. Processed meat remains a Group 1 carcinogen, and European regulators judged in 2023 that further cuts were warranted. - The “without compromising antibotulinal effects” claim remains the regulatory premise, though it was not independently re-tested in this check.

Weight. A good example of graduated, adaptive risk management under a risk-risk trade-off (digest insight 8), and of why such measures need periodic re-evaluation. It should not be cited as a case where the risk was resolved. The episode also shows that “neither over- nor under-regulation” can be a moving target over four decades.

Claim 9. MMR–autism is an “unregulated alarm”; UK vaccination numbers “are increasing again”; the mumps outbreak was not caused by the MMR scare (pp. 22–23)#

Original claim. No regulator restricted MMR, so the case is an unregulated alarm, not a false positive (p. 22). Citing the UK Department of Health (2003), uptake had fallen but “the numbers are increasing again”. The England and Wales mumps outbreak was largely among young adults who missed the MMR programme begun in 1988 (pp. 22–23). The chapter also says the suspected culprit was the preservative thimerosal (p. 22), and that “10 of the 12 authors” retracted support (p. 22).

What happened since. - The underlying science. - The Lancet formally retracted Wakefield et al. (1998) in February 2010. The BMJ judged the article “fraudulent” (Godlee et al., Jan 2011). - A Danish cohort of 657,461 children found a fully adjusted autism hazard ratio of 0.93 (0.85–1.02) for MMR-vaccinated vs unvaccinated children (Hviid et al., Annals of Internal Medicine, 2019). - The 2004 “Retraction of an interpretation” (Murch et al.) was signed by 10 of the paper’s 13 authors, not “10 of the 12”. - Thimerosal error confirmed. CDC: “Measles, mumps, and rubella (MMR) vaccines do not and never did contain thimerosal” (page dated 28 Oct 2025). The chapter conflates the MMR scare with the separate US thimerosal controversy. - UK uptake and measles (UKHSA, updated 26 January 2026): - First-dose MMR coverage rose from 92.5 % (2010) to a peak of 95.3 % (2016). So “increasing again” was right for roughly a decade after 2003. - It then fell to 92.3 % (2023–24), with second-dose coverage at 84.4 % in 2024. - The UK “achieved measles elimination in 2016, but this was short-lived”. It lost elimination status in 2018, regained it in 2021, and on 2024 data (3,681 cases) WHO “confirmed that endemic measles transmission was re-established”. - The South Wales outbreak of 2012–13 was widely linked to local MMR hesitancy dating from the late-1990s scare (McCartney, BMJ, April 2013) [outbreak case and death totals not re-fetched]. - The alarm entered US government communication. The CDC page “Vaccines and Autism” (last updated 22 July 2026) now states: “The claim ‘vaccines do not cause autism’ is not an evidence-based claim because studies have not ruled out the possibility that infant vaccines cause autism”. It still acknowledges reviews finding “with a high strength of evidence that there is no association” for MMR. The change first appeared in November 2025 [date from contemporaneous reporting]. US measles cases were 2,289 in 2025 and 3,659 in 2026 to 24 September (CDC). - Mumps. Nothing found that contradicts the chapter’s attribution of the 2004–05 outbreak to cohorts who missed MMR.

Verdict: partly held up. - The classification is technically correct: no regulator acted against MMR. - The mumps point and the short-term uptake recovery were right. - The thimerosal statement is wrong. - The implied reassurance has aged badly. The regulation-only scope left out the largest real-world costs of this false alarm: lower coverage, outbreaks and lost elimination status. Thirteen years on, the alarm has reached official health communication in the United States.

Weight. This case is the strongest evidence against the chapter’s scope choice, and relevant to Claim 7. Unfounded alarms can do lasting harm without any regulatory action, through markets, rhetoric and individual choice. They can also migrate into institutions over time. The lens should count such “unregulated” alarms as consequential errors even though the chapter’s typology does not.

Claim 10. Recommendations: give “equal weight to avoiding both false negatives and false positives”; assess alternatives, including no action; do not treat research as a substitute for regulation; build re-evaluation into decisions (pp. 34–35)#

Original claim. “There is a need to develop more nuanced policy analysis methods that give equal weight to avoiding both false negatives and false positives” (p. 34). The chapter’s seven lessons (pp. 34–35) include: - being open about disagreement and uncertainty; - resourcing alternatives assessment “(including no action)”; - taking care with large-scale introductions; - treating research as a supplement to regulation, not a substitute (“should supplement other risk-reducing regulatory measures and not be seen as a regulatory measure in itself”); - designing decisions “with re-evaluation as a key component”.

What happened since. - Alternatives assessment was institutionalised in several places. - The US National Academies published A Framework to Guide Selection of Chemical Alternatives (Oct 2014). - The European Commission’s Chemicals Strategy for Sustainability (2020) led to Recommendation (EU) 2022/2510 (8 Dec 2022) on “safe and sustainable by design” criteria, which aims to “prioritise innovation” for safer chemicals. - REACH restrictions such as the 2018 phthalates rule rest on formal analysis of alternatives. The DEHP medical-device sunset was postponed to 2030 to allow alternatives to be qualified, which shows the constraint the chapter identified (Lesson 3). - Research as a substitute for action was challenged in court. - In Case T-521/14, Sweden v Commission (16 December 2015), the General Court held that the Commission had failed to act by not adopting endocrine-disruptor criteria by the December 2013 deadline, and that the pending impact assessment did not excuse the delay [judgment text not re-fetched; characterised from its well-known holding]. Criteria followed in 2017–18. - In Bayer CropScience v Commission (C-499/18 P, 6 May 2021), the Court of Justice upheld neonicotinoid restrictions. Where “it is not possible to determine with certainty the existence or extent of the alleged risk, because the results of the studies carried out are inconclusive”, but likely harm persists, “the precautionary principle justifies the adoption of restrictive measures”. The burden then falls on the party concerned to show safety. - Re-evaluation was built in, with mixed pace. - EFSA’s systematic re-evaluation of pre-2009 additives (nitrites 2017) and its BPA re-evaluation (2023) both led to tighter rules. - The EU phthalate decision was revisited within five years once new biomonitoring data arrived. - In the US, TSCA was amended in 2016 (Frank R. Lautenberg Chemical Safety for the 21st Century Act, Pub. L. 114-182 [not re-fetched]). The amendments require risk evaluations of existing chemicals and led to 2024 rules on perchloroethylene and trichloroethylene. Compliance dates for these were extended in 2025–26. - The “equal weight” principle was taken up in opposite directions. - In the EU, Horizon Europe (Regulation (EU) 2021/695, recital 6) directs support to “innovation-friendly regulation, in line with the innovation principle”. The principle was promoted as a counterweight to precaution. - In the US, Executive Order 14192 (31 Jan 2025) requires ten regulations to be identified for repeal for each new one. Executive Order 14303 (23 May 2025) restricts “overly precautionary assumptions”. Both push towards weighting false positives more heavily. - EO 14303 also requires agencies to “transparently acknowledge and document uncertainties”, which matches the chapter’s Lesson 2.

Verdict: partly held up. Alternatives assessment, built-in re-evaluation, and the refusal to let further study stand in for decision have each gained institutional footholds, especially in EU chemicals and food law and EU case law, with visible results (BPA, phthalates, nitrites, endocrine disruptors). “Equal weight” has not been operationalised as a method anywhere I found. In US federal policy since 2025, the balance has moved deliberately the other way.

Weight. These recommendations are the chapter’s most transferable content, and the later record supports them as practical design principles. How they fare depends heavily on political context and on the availability of alternatives.

Minor factual checks#

• Thimerosal and MMR (p. 22): wrong. MMR never contained thimerosal (CDC, 2025).
• “10 of the 12 authors” (p. 22): the original paper had 13 authors; 10 signed the 2004 retraction of the interpretation (Murch et al., Lancet, March 2004).
• IARC and mobile phones: the chapter does not mention IARC’s May 2011 Group 2B classification, which predates publication. Chapter 21 of the same report covers it.
• Swine flu hindsight: the 1976 vaccine–GBS association was later judged causal (IOM, 2004). A swine-origin H1N1 pandemic did occur in 2009 (Dawood et al., 2012). Neither changes the chapter’s classification, but both qualify the “false alarm” narrative.
• BSE/vCJD as “jury still out” (Table 2.3): hard to defend in 2013, given established vCJD deaths. The October 2013 appendix survey reinforced the real-risk reading.
• Saccharin: the EPA hazardous-waste delisting (Dec 2010) was already in the past when the chapter was published but is not mentioned. It lengthens the reversal time the chapter implies.

Implications for the section’s transferable insights#

Grades refer to the digest’s list (LL2-02 digest, “Transferable insights”).

1. Claims of false alarm deserve the same scrutiny as claims of harm (pp. 19–25). Strengthened. Most checked “jury still out” cases moved towards harm or regulation.
2. Definitions and evidential thresholds determine how many errors of each kind are found (pp. 18–19, 33). Strong, and reinforced by hindsight. The chapter’s own definition (Cox, 2007) and its exclusion of risk-risk trade-offs and “unregulated” alarms shape its results as much as the cases do.
3. Absence of evidence of harm is not evidence of safety (p. 21). Strong. The logic is now written into EU case law (CJEU, 2021). Several cases later shifted, including PFOA and BPA.
4. Judge decisions ex ante, not by outcome (pp. 31–32). Moderate, unchanged. The 2009 H1N1 pandemic shows why 1976 should not be judged solely by its outcome.
5. Evidence lags a changing technology (p. 21). Moderate. Mobile phones show the lag working in both directions: early signals weakened with better designs.
6. Suppressed dissent and pre-commitment degrade decisions (pp. 27–28, 34). Moderate, unchanged. Not re-tested here.
7. Scale amplifies unknowns (pp. 28, 35). Moderate, unchanged.
8. Alternatives and graduated, adaptive measures reduce the cost of error (pp. 25, 32, 35). Moderate to strong. Supported by irradiation (hygiene alternatives), nitrites (graduated tightening) and DEHP in medical devices (sunset delayed pending alternatives).
9. Agencies face structural incentives towards inaction (p. 34). Moderate, but not one-directional. Reversal of precautionary measures (saccharin, irradiation, cyclamate) is also slow. Inertia affects both error types.
10. Organised doubt-manufacturing exists (pp. 33–34). Strengthened as a mechanism by extensive document analyses since 2016. Still unproven as the explanation for the scarcity of false positives. Organised promotion of unfounded alarms is also documented (MMR).
11. Mistaken precaution spurs innovation (pp. 32–33). Weak to moderate. The weak Porter hypothesis has some support, the strong version does not, and innovation costs of prolonged gatekeeping are documented.

Two lessons the chapter points towards but does not state: - Error typologies are themselves governance choices. Where the costs of precaution fall (risk-risk trade-offs, diffuse population effects, alarms acting outside regulation) determines whether they are counted at all. - Reversal is slow in both directions. Once set, precautionary measures and public alarms both persist for decades, so the design of re-evaluation matters more than the initial call.

Method and access notes#

• The web-search quota was exhausted before this check started. Sources were verified by direct retrieval:
• EU legal texts from the EU Publications Office (cellar) or EUR-Lex ELI pages;
• GOV.UK content via the GOV.UK content API;
• EFSA, FDA, CDC and FWS pages via direct fetch;
• the Federal Register via its public API;
• scholarly records and abstracts via Crossref and OpenAlex.
• Items marked “[not re-fetched]” rest on well-established facts I could not retrieve in this session. Examples: the IARC 2010 nitrate/nitrite classification, the T-521/14 judgment text, and the Lautenberg Act citation.
• Some 14 of the 32 “jury still out” cases were not checked. The Claim 1 tally should be read as indicative, not complete.
• Search for later critiques of the chapter relied on citation tracing (OpenAlex “cites” queries for Hansen et al., 2007, and the 2013 chapter). Critiques published in books or grey literature may be missed.
• A few items could not be retrieved: the ANSES (France) 2022 nitrite opinion, the PAHO statement on Canada’s measles elimination status, and news reports on think-tank funding. They are therefore not used.

Sources#

Chapter and direct responses - Hansen, S.F. and Tickner, J.A. (2013). The precautionary principle and false alarms — lessons learned. In EEA, Late lessons from early warnings: science, precaution, innovation, EEA Report No 1/2013, pp. 17–45. https://www.eea.europa.eu/publications/late-lessons-2 - Hansen, S.F., Krayer von Krauss, M.P. and Tickner, J.A. (2007, Feb). Categorizing mistaken false positives in regulation of human and environmental health. Risk Analysis 27:255–269. https://doi.org/10.1111/j.1539-6924.2006.00874.x - Cox, L.A. Jr. (2007, Oct). Regulatory false positives: true, false, or uncertain? Risk Analysis 27:1083–1086. https://doi.org/10.1111/j.1539-6924.2007.00975.x - Hansen, S.F., von Krauss, M.K. and Tickner, J.A. (2007, Oct). Response to “Regulatory false positives: true, false, or uncertain?” Risk Analysis 27:1087–1089. https://doi.org/10.1111/j.1539-6924.2007.00970.x

Claim 1: “jury still out” cases - EFSA (19 April 2023). Bisphenol A in food is a health risk (news). https://www.efsa.europa.eu/en/news/bisphenol-food-health-risk ; Scientific opinion, EFSA Journal 21(4):6857. https://doi.org/10.2903/j.efsa.2023.6857 ; EFSA BPA topic page. https://www.efsa.europa.eu/en/topics/topic/bisphenol - Commission Regulation (EU) 2024/3190 of 19 December 2024 (BPA and other bisphenols in food contact materials). https://eur-lex.europa.eu/eli/reg/2024/3190/oj - US FDA. Bisphenol A (BPA): use in food contact application (content current as of 20 April 2023; accessed Sept 2026). https://www.fda.gov/food/food-packaging-other-substances-come-contact-food-information-consumers/bisphenol-bpa-use-food-contact-application - Commission Regulation (EU) 2018/2005 of 17 December 2018 (DEHP, DBP, BBP, DIBP restriction). https://eur-lex.europa.eu/eli/reg/2018/2005/oj - Commission Regulation (EU) 2023/2482 of 13 November 2023 (DEHP in medical devices). https://eur-lex.europa.eu/eli/reg/2023/2482/oj - Commission Delegated Regulation (EU) 2017/2100 of 4 September 2017 (endocrine disruptor criteria, biocides). https://eur-lex.europa.eu/eli/reg_del/2017/2100/oj - Commission Delegated Regulation (EU) 2023/707 of 19 December 2022 (new CLP hazard classes). https://eur-lex.europa.eu/eli/reg_del/2023/707/oj - Zahm, S. et al. (2024, Jan). Carcinogenicity of perfluorooctanoic acid and perfluorooctanesulfonic acid. Lancet Oncology 25:16–17. https://doi.org/10.1016/S1470-2045(23)00622-8 - Guha, N. et al. (2012, Dec). Carcinogenicity of trichloroethylene, tetrachloroethylene, some other chlorinated solvents, and their metabolites. Lancet Oncology 13:1192–1193. https://doi.org/10.1016/S1470-2045(12)70485-0 - US EPA (18 Dec 2024). Perchloroethylene (PCE); Regulation under TSCA, final rule. https://www.federalregister.gov/documents/2024/12/18/2024-30117/perchloroethylene-pce-regulation-under-the-toxic-substances-control-act-tsca ; compliance date extensions (28 July 2026). https://www.federalregister.gov/documents/2026/07/28/2026-15192/perchloroethylene-pce-and-carbon-tetrachloride-ctc-regulation-under-the-toxic-substances-control-act - Lauby-Secretan, B. et al. (2013, April). Carcinogenicity of polychlorinated biphenyls and polybrominated biphenyls. Lancet Oncology 14:287–288. https://doi.org/10.1016/S1470-2045(13)70104-9 - Commission Implementing Regulation (EU) 2016/871 of 1 June 2016 (non-renewal of amitrole). https://eur-lex.europa.eu/eli/reg_impl/2016/871/oj - US FDA (23 Oct 2019). Questions and answers about breast implant-associated anaplastic large cell lymphoma. https://www.fda.gov/medical-devices/breast-implants/questions-and-answers-about-breast-implant-associated-anaplastic-large-cell-lymphoma-bia-alcl - EFSA CONTAM Panel (June 2015). Scientific opinion on acrylamide in food. EFSA Journal 13(6):4104. https://doi.org/10.2903/j.efsa.2015.4104 - Commission Regulation (EU) 2017/2158 of 20 November 2017 (acrylamide mitigation and benchmark levels). https://eur-lex.europa.eu/eli/reg/2017/2158/oj - Gill, O.N. et al. (15 Oct 2013). Prevalent abnormal prion protein in human appendixes after BSE epizootic. BMJ 347:f5675. https://doi.org/10.1136/bmj.f5675 - Filippini, T. et al. (2019). Association between outdoor air pollution and childhood leukemia: systematic review and dose–response meta-analysis. Environmental Health Perspectives 127:046002. https://doi.org/10.1289/EHP4381 - National Academies of Sciences, Engineering, and Medicine (2016). Genetically Engineered Crops: Experiences and Prospects. https://doi.org/10.17226/23395 ; https://www.nationalacademies.org/publications/23395 - Pleasants, J.M. and Oberhauser, K.S. (2013). Milkweed loss in agricultural fields because of herbicide use. Insect Conservation and Diversity 6:135–144. https://doi.org/10.1111/j.1752-4598.2012.00196.x - US Fish and Wildlife Service. Monarch butterfly (proposed threatened listing, Dec 2024; accessed Sept 2026). https://www.fws.gov/initiative/pollinators/monarchs - US FDA (27 Feb 2025). Aspartame and other sweeteners in food (saccharin history; cyclamate prohibition). https://www.fda.gov/food/food-additives-petitions/aspartame-and-other-sweeteners-food - Dawood, F.S. et al. (2012, Sept). Estimated global mortality associated with the first 12 months of 2009 pandemic influenza A H1N1. Lancet Infectious Diseases 12:687–695. https://doi.org/10.1016/S1473-3099(12)70121-4 - Institute of Medicine (2004). Immunization Safety Review: Influenza Vaccines and Neurological Complications. https://doi.org/10.17226/10822

Claim 2: error rates and over-regulation - Wiener, J.B., Rogers, M.D., Hammitt, J.K. and Sand, P.H. (eds) (2011). The Reality of Precaution: Comparing Risk Regulation in the United States and Europe. RFF Press. Review: Nappert (2013, Nov), Journal of Risk Research 16:1317–1320. https://doi.org/10.1080/13669877.2012.756610 - Tosun, J. (2013, Oct). How the EU handles uncertain risks: understanding the role of the precautionary principle. Journal of European Public Policy 20:1517–1528. https://doi.org/10.1080/13501763.2013.834549 - Löfstedt, R. (2014, Aug). The precautionary principle in the EU: why a formal review is long overdue. Risk Management 16:137–163. https://doi.org/10.1057/rm.2014.7 - Garnett, K. and Parsons, D.J. (2017). Multi-case review of the application of the precautionary principle in European Union law and case law. Risk Analysis 37:502–516 (online May 2016). https://doi.org/10.1111/risa.12633 - Simpson, R.D. (2014, June). Do regulators overestimate the costs of regulation? Journal of Benefit-Cost Analysis 5:315–332. https://doi.org/10.1515/jbca-2014-0027 - Jarvis, S., Deschenes, O. and Jha, A. (2022). The private and external costs of Germany’s nuclear phase-out. Journal of the European Economic Association 20:1311–1346. https://doi.org/10.1093/jeea/jvac007 - Neidell, M., Uchida, S. and Veronesi, M. (2021). The unintended effects from halting nuclear power production: evidence from Fukushima Daiichi accident. Journal of Health Economics 79:102507. https://doi.org/10.1016/j.jhealeco.2021.102507 - Executive Order 14303, Restoring Gold Standard Science (signed 23 May 2025; FR 29 May 2025). https://www.federalregister.gov/documents/2025/05/29/2025-09802/restoring-gold-standard-science

Claim 3: mobile phones - Baan, R. et al. (2011, July). Carcinogenicity of radiofrequency electromagnetic fields. Lancet Oncology 12:624–626. https://doi.org/10.1016/S1470-2045(11)70147-4 - Schüz, J. et al. (2022). Cellular telephone use and the risk of brain tumors: update of the UK Million Women Study. JNCI 114:704–711. https://doi.org/10.1093/jnci/djac042 - Castaño-Vinyals, G. et al. (2022). Wireless phone use in childhood and adolescence and neuroepithelial brain tumours: MOBI-Kids. Environment International 160:107069. https://doi.org/10.1016/j.envint.2021.107069 - Feychting, M. et al. (2024, March). Mobile phone use and brain tumour risk – COSMOS. Environment International 185:108552. https://doi.org/10.1016/j.envint.2024.108552 - Karipidis, K. et al. (2024, Sept). The effect of exposure to radiofrequency fields on cancer risk in the general and working population: systematic review of human observational studies – Part I. Environment International 191:108983. https://doi.org/10.1016/j.envint.2024.108983 - Mevissen, M. et al. (2025, May). Effects of radiofrequency electromagnetic field exposure on cancer in laboratory animal studies, a systematic review. Environment International 199:109482. https://doi.org/10.1016/j.envint.2025.109482 ; Corrigendum (2026, Aug), 214:110368. https://doi.org/10.1016/j.envint.2026.110368 - Berrington de González, A. et al. (2024, May). Advisory Group recommendations on priorities for the IARC Monographs. Lancet Oncology 25:546–548. https://doi.org/10.1016/S1470-2045(24)00208-0 ; IARC priorities poster (April 2024). https://monographs.iarc.who.int/wp-content/uploads/2024/04/Priorities-poster-streams-April2024.v5.pdf ; IARC news (4 Nov 2024). https://monographs.iarc.who.int/news-events/advisory-group-recommendations-on-priorities-for-the-iarc-monographs-during-2025-2029/ - Hardell, L. and Carlberg, M. (2021). Lost opportunities for cancer prevention: historical evidence on early warnings with emphasis on radiofrequency radiation. Reviews on Environmental Health. https://doi.org/10.1515/reveh-2020-0168

Claim 4: food irradiation - EFSA CEF Panel (April 2011). Scientific opinion on the chemical safety of irradiation of food. EFSA Journal 9(4):1930. https://doi.org/10.2903/j.efsa.2011.1930 - EFSA BIOHAZ Panel (April 2011). Scientific opinion on the efficacy and microbiological safety of irradiation of food. EFSA Journal 9(4):2103. https://doi.org/10.2903/j.efsa.2011.2103 - European Commission (27 Oct 2023). Report on food and food ingredients treated with ionising radiation for 2020–2021, COM(2023) 676. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52023DC0676 ; report for 2018–2019, COM(2021) 79. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52021DC0079 ; EU food irradiation pages. https://food.ec.europa.eu/food-safety/biological-safety/food-irradiation_en - US FDA (5 March 2024). Food irradiation: what you need to know. https://www.fda.gov/food/buy-store-serve-safe-food/food-irradiation-what-you-need-know

Claim 5: innovation - Ambec, S., Cohen, M.A., Elgie, S. and Lanoie, P. (2013). The Porter hypothesis at 20. Review of Environmental Economics and Policy 7:2–22. https://doi.org/10.1093/reep/res016 - Dechezleprêtre, A. and Sato, M. (2017). The impacts of environmental regulations on competitiveness. Review of Environmental Economics and Policy 11:183–206. https://doi.org/10.1093/reep/rex013 - Cohen, M.A. and Tubb, A. (2018). The impact of environmental regulation on firm and country competitiveness: a meta-analysis of the Porter hypothesis. JAERE 5:371–399. https://doi.org/10.1086/695613 - Rubashkina, Y., Galeotti, M. and Verdolini, E. (2015). Environmental regulation and competitiveness: empirical evidence on the Porter hypothesis from European manufacturing sectors. Energy Policy 83:288–300. https://doi.org/10.1016/j.enpol.2015.02.014 - Smart, R.D., Blum, M. and Wesseler, J. (2016). Trends in approval times for genetically engineered crops in the US and the EU. Journal of Agricultural Economics 68:182–198. https://doi.org/10.1111/1477-9552.12171 - Wesseler, J. and Zilberman, D. (2014). The economic power of the Golden Rice opposition. Environment and Development Economics 19:724–742. https://doi.org/10.1017/S1355770X1300065X ; follow-up (2016): https://doi.org/10.1017/S1355770X16000292

Claim 6: doubt-manufacturing - Kearns, C.E., Schmidt, L.A. and Glantz, S.A. (2016, Sept). Sugar industry and coronary heart disease research. JAMA Internal Medicine 176:1680. https://doi.org/10.1001/jamainternmed.2016.5394 - Kearns, C.E., Apollonio, D. and Glantz, S.A. (2017). Sugar industry sponsorship of germ-free rodent studies. PLOS Biology 15:e2003460. https://doi.org/10.1371/journal.pbio.2003460 - Supran, G. and Oreskes, N. (2017, Aug). Assessing ExxonMobil’s climate change communications (1977–2014). Environmental Research Letters 12:084019. https://doi.org/10.1088/1748-9326/aa815f - Supran, G., Rahmstorf, S. and Oreskes, N. (2023, Jan). Assessing ExxonMobil’s global warming projections. Science 379:eabk0063. https://doi.org/10.1126/science.abk0063 - Gaber, N., Bero, L. and Woodruff, T.J. (2023). The Devil they knew: chemical documents analysis of industry influence on PFAS science. Annals of Global Health 89:37. https://doi.org/10.5334/aogh.4013 - Goldberg, R.F. and Vandenberg, L.N. (2021, March). The science of spin. Environmental Health 20:33. https://doi.org/10.1186/s12940-021-00723-0 - Legg, T., Hatchard, J. and Gilmore, A.B. (2021, June). The Science for Profit Model. PLOS ONE 16:e0253272. https://doi.org/10.1371/journal.pone.0253272 - Schäffer, A. et al. (2023, Nov). Conflicts of interest in the assessment of chemicals, waste, and pollution. Environmental Science & Technology. https://doi.org/10.1021/acs.est.3c04213 - Björnberg, K.E. et al. (2017). Climate and environmental science denial: a review of the scientific literature published in 1990–2015. Journal of Cleaner Production. https://doi.org/10.1016/j.jclepro.2017.08.066

Claim 7: reversal times - US EPA (17 Dec 2010). Removal of saccharin and its salts from the lists of hazardous constituents, hazardous wastes and hazardous substances, final rule. https://www.federalregister.gov/documents/2010/12/17/2010-31773/hazardous-waste-management-system-identification-and-listing-of-hazardous-waste-removal-of-saccharin - (Also FDA sweeteners and irradiation pages, and Jarvis et al. 2022 and Neidell et al. 2021, above.)

Claim 8: nitrites - Bouvard, V. et al. (2015, Dec). Carcinogenicity of consumption of red and processed meat. Lancet Oncology 16:1599–1600. https://doi.org/10.1016/S1470-2045(15)00444-1 - EFSA ANS Panel (June 2017). Re-evaluation of potassium nitrite (E 249) and sodium nitrite (E 250) as food additives. EFSA Journal 15(6):4786. https://doi.org/10.2903/j.efsa.2017.4786 - EFSA CONTAM Panel (March 2023). Risk assessment of N-nitrosamines in food. EFSA Journal 21(3):7884. https://doi.org/10.2903/j.efsa.2023.7884 - Commission Regulation (EU) 2023/2108 of 6 October 2023 (nitrites and nitrates as food additives). https://eur-lex.europa.eu/eli/reg/2023/2108/oj (text retrieved via https://publications.europa.eu/resource/celex/32023R2108)

Claim 9: MMR - The Editors of The Lancet (Feb 2010). Retraction — Ileal-lymphoid-nodular hyperplasia… Lancet 375:445. https://doi.org/10.1016/S0140-6736(10)60175-4 - Murch, S.H. et al. (March 2004). Retraction of an interpretation. Lancet 363:750. https://doi.org/10.1016/S0140-6736(04)15715-2 - Godlee, F., Smith, J. and Marcovitch, H. (5 Jan 2011). Wakefield’s article linking MMR vaccine and autism was fraudulent. BMJ 342:c7452. https://doi.org/10.1136/bmj.c7452 - Hviid, A. et al. (2019). Measles, mumps, rubella vaccination and autism: a nationwide cohort study. Annals of Internal Medicine 170:513–520. https://doi.org/10.7326/M18-2101 - UKHSA (updated 26 Jan 2026). UK measles and rubella elimination indicators and status. https://www.gov.uk/government/publications/measles-and-rubella-elimination-uk/uk-measles-and-rubella-elimination - UKHSA (data as of 23 March 2026). Laboratory confirmed cases of measles, rubella and mumps in England: October to December 2025. https://www.gov.uk/government/publications/measles-mumps-and-rubella-lab-confirmed-cases-in-england-2025/laboratory-confirmed-cases-of-measles-rubella-and-mumps-in-england-october-to-december-2025 - McCartney, M. (22 April 2013). MMR, measles, and the South Wales Evening Post. BMJ 346:f2598. https://doi.org/10.1136/bmj.f2598 - US CDC. Vaccines and Autism (last updated 22 July 2026). https://www.cdc.gov/vaccine-safety/about/autism.html - US CDC. Thimerosal and vaccines (28 Oct 2025). https://www.cdc.gov/vaccine-safety/about/thimerosal.html - US CDC. Measles cases and outbreaks (data as of 24 Sept 2026). https://www.cdc.gov/measles/data-research/index.html

Claim 10: recommendations and frameworks - National Research Council (Oct 2014). A Framework to Guide Selection of Chemical Alternatives. https://doi.org/10.17226/18872 - Commission Recommendation (EU) 2022/2510 of 8 December 2022 (safe and sustainable by design framework). https://eur-lex.europa.eu/eli/reco/2022/2510/oj - General Court, Case T-521/14, Sweden v Commission, judgment of 16 December 2015 [not re-fetched]. https://curia.europa.eu/juris/liste.jsf?num=T-521/14 - Court of Justice, Case C-499/18 P, Bayer CropScience and Bayer v Commission, judgment of 6 May 2021. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62018CJ0499 - Frank R. Lautenberg Chemical Safety for the 21st Century Act, Pub. L. 114-182 (22 June 2016) [not re-fetched]. https://www.congress.gov/bill/114th-congress/house-bill/2576 - Regulation (EU) 2021/695 of 28 April 2021 establishing Horizon Europe (recital 6, innovation principle). https://eur-lex.europa.eu/eli/reg/2021/695/oj - Executive Order 14192, Unleashing Prosperity Through Deregulation (signed 31 Jan 2025; FR 6 Feb 2025). https://www.federalregister.gov/documents/2025/02/06/2025-02345/unleashing-prosperity-through-deregulation - Executive Order 14303, Restoring Gold Standard Science (above).