Film
An Inconvenient Study
The film follows ICAN's effort, beginning in 2017 according to Aaron Siri's Senate testimony, to get a Henry Ford Health infectious disease chief to run a vaccinated-versus-unvaccinated study. It presents the resulting 2020 manuscript, a retrospective cohort of 18,468 children born 2000 to 2016 with 1,957 unvaccinated, which reports a 2.5-fold higher hazard of a chronic condition in vaccinated children. It says the authors would not publish it and that Henry Ford shelved it, and it uses hidden-camera footage of a dinner with the lead physician. Its response page answers critics from Henry Ford, The Conversation and STAT. Its content is described here from ICAN's own pages and press material.
Who made it
- Produced by ICAN; the film site's footer carries the ICAN copyright and names no director or crew. ICAN press releases credit Del Bigtree as executive producer.
- Aaron Siri, who testified to the Senate on the study on 9 September 2025 (exhibit 12), writes that ICAN is a non-profit his law firm regularly represents (Siri testimony, section III).
- Siri's testimony says ICAN's CEO had previously met Dr. Marcus Zervos and that ICAN and Siri met Zervos in his office in early 2017 to urge a vaccinated-versus-unvaccinated study (Siri testimony, section III).
- The manuscript lists four authors: Lois Lamerato, Abigail Chatfield, Amy Tang (Henry Ford Health System, Public Health Sciences) and Marcus Zervos (Infectious Diseases, Henry Ford; Wayne State University). Its cover page states 'This study had no external funding.' (exhibit 10, cover page).
- ICAN's film site claims 'over 100 million views worldwide and growing' with no source (aisstudy.com, 'Spread the word' section).
- The film site sells or solicits merchandise, donations and screening partnerships on the same page as the film (aisstudy.com, 'Spread the word' section).
Its claims checked against the record
13 claims checked:7 supported · 4 partly supported · 0 contradicted · 2 not in the record
A verdict rests on the documents named in its row.
| The work says | Verdict | What the record shows |
|---|---|---|
| The study followed 18,468 children born 2000 to 2016 in the Henry Ford health plan, 1,957 with no vaccine exposure and 16,511 with at least one vaccine. | Supported by the record | Abstract: 'A total of 18,468 consecutive subjects met eligibility criteria for the study, of which 1,957 had no exposure to vaccination and 16,511 had received at least one vaccine'. Table 1 gives the same counts. |
| Vaccinated children had about 2.5 times the hazard of a chronic health condition (the film's headline figure). | Supported by the record | The manuscript reports an adjusted HR of 2.53 (CI 2.16-2.96) in the abstract and 2.54 (CI 2.16-2.97) in the results text and Table 3. The figure is the manuscript's; the design limits listed below apply to it. |
| The condition-specific hazard ratios (asthma, autoimmune, atopic, neurodevelopmental) are as the abstract and Siri's testimony give them. | Partly supported | Abstract versus Table 3 (adjusted): any chronic condition 2.53 vs 2.54; asthma 4.25 (3.23-5.59) vs 4.29 (3.26-5.65); autoimmune 4.79 (1.36-16.94) vs 5.96 (1.48-24.11); atopic 3.03 and neurodevelopmental 5.53 match. The abstract's eczema HR of 1.31 has no row in Table 3. Siri's testimony quotes the Table 3 values (4.29, 5.96); the discussion text says 'a 6-fold increased risk of autoimmune disease'. The manuscript carries two sets of numbers. |
| The unvaccinated group's shorter follow-up was tested and the result held: restricting to children enrolled at least 1, 3 and 5 years raised the hazard ratios (2.84, 3.48, 4.05), so follow-up length does not explain the finding. | Partly supported | The three hazard ratios appear in the manuscript's sensitivity analyses. The manuscript reports median follow-up of 461 days for unexposed and 970 days for exposed children and gives no group counts for the 1-, 3- and 5-year subsets, so the reader cannot see how many unvaccinated children remain in each. |
| The finding does not reflect unequal health care use: unvaccinated children had about 2 encounters a year against about 7 for vaccinated children, but a sensitivity analysis limited to children with at least one encounter gave the same direction. | Partly supported | The manuscript reports 7 versus 2 annual encounters ('almost 5' for unvaccinated children with a chronic condition) and an encounter-restricted HR of 1.87 (CI 1.60-2.19), down from 2.54. It states that 'Well visits coincide with the vaccination schedule' and that this 'could introduce an ascertainment bias'. The restriction removes children with zero encounters and leaves the 7-versus-2 gap in place; no analysis matches visit rates. |
| ADHD, learning disability, tics, diabetes and other conditions had cases only among vaccinated children, 'statistically impossible' to compare. | Supported by the record | Table 2: ADHD 262 cases in vaccinated, 0 in unvaccinated; behavioral disability 165 vs 0; learning disability 65 vs 0; tics 46 vs 0; diabetes 42 vs 0. The manuscript states hazard ratios 'could not be calculated' for these. The manuscript does not report how many unvaccinated children were followed to the ages at which these diagnoses are typically made; only children aged 2 and over were evaluated for neurodevelopmental and mental health disorders. |
| Cancer served as a control outcome and showed no association with vaccination, which rules out uncontrolled confounding. | Partly supported | The manuscript uses cancer as its control: IRR 0.79 (0.45-1.39), adjusted HR 0.90 (0.51-1.59), with 13 cases in the unvaccinated group (Table 2). It names 'injuries or cancer' as possible controls and analyzes only cancer. A single rare outcome that needs no routine visits to detect is a control for confounding of that kind and does not test visit-dependent diagnoses. Ear infection, a visit-dependent diagnosis, shows HR 7.00 (6.05-8.10) in the results. |
| The study's authors say it does not prove causation (the film: 'A retrospective study does not prove causation'), and the film calls it a signal for further research. | Supported by the record | Manuscript conclusion: 'Our preliminary findings cannot prove causality and warrant further investigation.' The same conclusion says the results suggest 'in certain susceptible children, exposure to vaccination may increase the likelihood of developing a chronic health condition'. Siri's testimony states the study was withheld because 'its results showed that the vaccinated children in the study suffered from numerous chronic health issues that did not plague the unvaccinated children', a causal framing the manuscript does not adopt. |
| The unvaccinated and vaccinated groups differed at birth, and the unvaccinated group had shorter observation (Henry Ford's stated flaws). | Supported by the record | Table 1: male 55% vs 50%; white 46% vs 36%; prematurity 2% vs 6%; respiratory distress at birth 1% vs 4%. Follow-up IQR for unexposed children is 196-1,081 days, which matches Henry Ford's 'a quarter observed to six months, 75% to age 3'; the vaccinated median is 970 days against 461. |
| The study did not account for the number or type of vaccines, the timing of vaccination relative to diagnosis, or changes in vaccine guidance over the 2000-2016 birth years. | Supported by the record | Manuscript: 'We did not evaluate the influence of temporal relationships, individual vaccines, or the number of vaccines'. The Methods give no index date and do not say whether exposure was entered as a time-varying covariate; they say subjects were 'classified by exposure to immunizations prior to onset of each condition'. Table 1 defines exposure as 'receipt of any vaccine during enrollment'. The manuscript also states no data on socioeconomic status. |
| The manuscript was not submitted for publication for one reason: its results showed harm in vaccinated children. | Not in the record | Both sides state it was not submitted. Siri's single reason rests on his account of conversations. Henry Ford gives a different reason (below). No internal review record, authors' statement or journal correspondence is public in the sources searched. Record not produced. The first internal review of the draft, the reviewers' critiques and the initial study plan have not been released: Lead Stories (updated 31 October 2025) reports it asked Henry Ford Health for the reviewers' critiques, the study plan and an interview with Zervos, and records no response, while Henry Ford's statements of 26 September and 13 October 2025 describe the review without naming reviewers or publishing it. |
| Henry Ford shelved the draft at the first internal peer review because of data and methodology problems, and never considered it for submission; the circulated copy may have been altered to look finished. | Not in the record | Henry Ford's statement says the draft was 'immediately shelved upon the first internal peer review' and that 'there's reason to believe it's been altered from its original form'. No review document is public, the statement does not name the reviewers, and no document comparing the circulated manuscript with an earlier version was located. The flaws it lists in the data are present in the manuscript (previous two claims). Record not produced. The internal peer-review record that Henry Ford says shelved the draft, and any earlier version against which the 'altered' copy could be compared, are not public: Henry Ford's 13 October 2025 fact-check refers only to 'The first internal reviewers', and Lead Stories (31 October 2025) lists its request for the reviewers' critiques and the study plan as outstanding. |
| Henry Ford sent the filmmakers a cease-and-desist letter, which the film presents as suppression. | Supported by the record | Henry Ford states it 'has issued a cease-and-desist letter to the makers of an upcoming "documentary"' and calls its claims 'patently false, defamatory and troubling'. A reviewer quotes the letter as alleging 'false, misleading, and defamatory statements' that violate Michigan and federal law. The letter itself was not located. No suit is recorded in the sources searched, and the manuscript has been public since 9 September 2025 in the Senate hearing record (exhibit 10). |
What the record shows that it leaves out
- Design limits the manuscript itself states: 'Unvaccinated children have less healthcare utilization overall', 7 versus 2 annual encounters, 'could introduce an ascertainment bias' (exhibit 10, Limitations). The film's answer, that sick unvaccinated children averaged almost 5 visits, concerns children already diagnosed.
- Follow-up length. Median follow-up was 461 days for unvaccinated and 970 for vaccinated children, and the Kaplan-Meier figure carries no numbers-at-risk table. The 83% unvaccinated 10-year figure sits on a curve that is flat after about year 6 (exhibit 10, Figure 1).
- The encounter-restricted hazard ratio, 1.87 against 2.54 overall, is a lower figure than the film's headline (exhibit 10, results).
- The abstract, results text, discussion and Table 3 give different values for the any-condition, asthma and autoimmune hazard ratios, and Table 3 has no eczema row (exhibit 10).
- No index date or time-varying exposure definition appears in the Methods, and the authors state they 'did not evaluate the influence of temporal relationships' (exhibit 10).
- No vaccine-specific, dose or schedule analysis, and no socioeconomic data (exhibit 10, Limitations).
- Only one negative-control outcome, cancer, with 13 unvaccinated cases. The authors name injuries as another suggested control and did not use it. Ear infection, a diagnosis made at visits, shows HR 7.00 among the results (exhibit 10).
- The manuscript says the unvaccinated cohort's diagnoses come from records of the plan and HFHS; the plan has about 570,000 members, 'approximately one-third of whom receive care within HFHS', and the inclusion rule requires HFHS as primary care system. Care obtained elsewhere is not in the data (exhibit 10, Methods).
- The manuscript found no association with autism (HR 0.62, CI 0.10-3.69; one unvaccinated case) or cancer, and reports an IOM-consistent anaphylaxis association.
- Henry Ford's account that the paper was never considered for submission, set against the authors' silence in press: no author has made a public statement in the sources searched (Michigan Public, exhibit 11).
Documented efforts and withheld records
Each item is a documented act, with its source: an act against the work, its author or its sources, or the withholding of records on its subject. The list states what was done and does not state why.
- Statement 26 September 2025; letter date not located. Henry Ford Health states it has issued a cease-and-desist letter to the makers of an upcoming 'documentary', saying the film 'falsely and dangerously claims the health system suppressed a draft of a research paper'. The statement gives no date for the letter and does not describe its demands. Henry Ford Health, news statement
- Reported 6 November 2025. Michigan Public reports that Henry Ford Health sent the cease-and-desist letter to the makers of 'An Inconvenient Study' and quotes it as condemning 'the purposeful twisting of information and the spread of disinformation'. The report gives no date or demands and records no reply from ICAN. Michigan Public
- Review 2025 (undated on the page fetched). A Substack reviewer of the film quotes the Henry Ford letter as alleging 'false, misleading, and defamatory statements' that 'violate Michigan and federal law'. The reviewer gives no date or demands. madhavasetty.substack.com, 'A Review of An Inconvenient Study'
- Page fetched November 2026. ICAN's film response page, which answers criticisms from Henry Ford, The Conversation and STAT, does not mention the cease-and-desist letter; no ICAN reply to the letter was found in any document. ICAN, 'Criticisms and Responses'
- 26 June to 29 July 2020. ICAN says it submitted a FOIA request on 26 June 2020 for 'All documents in the CDC's possession which compare the health outcomes of children that have received vaccines with children that have never received any vaccines'. CDC's final response letter, dated 29 July 2020 and answering a request dated 13 July 2020, states: 'A search of our records failed to reveal any documents pertaining to your request. The CDC has not conducted a study of health outcomes in vaccinated vs unvaccinated populations.' ICAN reports that CDC had objected on 6 July that the request was 'Broad' and 'Vague' and on 9 July that the narrowed request was 'not sufficiently quantified'. CDC final response letter (Exhibit 938); ICAN article 'CDC Concedes Never Conducted Vaccinated v. Unvaccinated Study'
- Request 9 April 2025; complaint filed 12 May 2025. ICAN and Mississippi Medical Professionals for Informed Consent sued the CDC and HHS in the Southern District of Mississippi (No. 1:25-cv-135) under FOIA. The complaint states that their 9 April 2025 request for 'All data in the Vaccine Safety Datalink database. Alternatively, provide access to the database' was acknowledged the same day as 25-01124-FOIA and that, as of filing, the agencies had made no determination, given no completion date and produced nothing. Complaint, Document 1, Case 1:25-cv-00135-HSO-BWR (hosted by ICAN)
- ICAN release 14 May 2026. ICAN says Verizon customers who tried to reach AnInconvenientStudy.com and related pages received a 'server cannot be found' error, and that ICAN is collecting screenshots and account details from affected users. The release carries no Verizon statement; it is labeled a paid press release. ICAN press release via GlobeNewswire (read through Yahoo Finance copy)
Published responses
- Henry Ford Health (statement, 26 September 2025): States it issued a cease-and-desist letter, that the draft was shelved at first internal review for data and methodology problems and was never considered for submission, and that the circulated copy may have been altered. Lists group differences, shorter unvaccinated follow-up and missing vaccine-number, timing and guidance controls.
- Henry Ford Health (fact-check page, 13 October 2025): Says the draft proved nothing, that internal review caught the flaws and that the study was not hidden; quotes Jake Scott.
- Michigan Public (6 November 2025): Reports Adnan Munkarah: 'They said there are fatal flaws in that study', 'The data that has been used is not complete data', and that unpublished papers sit in drawers by the thousands. Reports Siri's claim that the system 'buried' the study, and that Lamerato and Zervos did not reply.
- Jake Scott, Stanford infectious diseases physician (STAT First Opinion, 9 September 2025): Argues detection bias from 7 versus 2 visits a year, no adequate correction, and that dropping children with zero encounters 'simply bakes the bias into the results'. Only the opening of the piece was readable (paywall).
- Jeffrey Morris, University of Pennsylvania biostatistician (The Conversation, republished by WLRN 26 September 2025): Lays out detection bias, follow-up (vaccinated median about 2.7 years, 75% followed past 15 months; unvaccinated 25% under 6 months, 50% under 15 months), unadjusted residence, income and insurance, and says the differences 'do not show that vaccines cause chronic disease'.
- Science Feedback (22 September 2025): Rates the claims based on the study 'Misleading'; notes zero ADHD cases among about 1,900 unvaccinated children and ear-infection hazard ratios of 6 to 8.
- ICAN (response page, 'Criticisms and Responses'): Answers seven criticisms: group size (rates, not counts), surveillance bias (1-, 3-, 5-year restrictions), ascertainment bias (almost 5 visits for diagnosed unvaccinated children; encounter-restricted analysis), baseline differences (adjusted HRs, cancer as control), no vaccine-specific analysis, causation ('A retrospective study does not prove causation'), autism counts (DSM-IV era diagnostic substitution). The DSM-IV explanation is ICAN's and does not appear in the manuscript. The page names no authors and does not mention the cease-and-desist letter.
- Aaron Siri, Senate Permanent Subcommittee on Investigations (9 September 2025): Testifies that the authors affirmed the study and declined to submit it for personal reasons, and that it was not published 'for one reason'. Quotes the manuscript's Table 3 values.
Where it connects
- The Studies Not Run: The film's subject is the vaccinated-versus-unvaccinated comparison; the CDC's 3 August 2020 FOIA answer states it has conducted no such study, and the manuscript is a retrospective one-system version with stated design limits.
- The Filtered Record: An unpublished manuscript reached the public through a Senate hearing record, and the documents give two incompatible reasons for its non-submission, neither supported by a released review record.