Wetenschap is het beste hulpmiddel dat we hebben om te onderscheiden wat waar is van wat we zouden willen dat waar was. Een groot deel ervan wordt betaald door partijen die belang hebben bij het antwoord. Die combinatie is goed gedocumenteerd. Interne memo's en rechtsdossiers tonen vier manoeuvres.
Een suikerbedrijfsvereniging betaalde voor een review uit 1967 die vet de schuld gaf. De tabaksindustrie financierde twijfel. Farmaceutische bedrijven lieten gunstige artikelen door ghostwriters schrijven, en negatieve trials werden niet gepubliceerd. Dezelfde zaken wijzen op een methode voor elk onderzoek: zoek de financiering op, controleer de voorafregistratie, geef de voorkeur aan harde eindpunten en weeg replicaties af.
What This Is
Industry influence on science is a documented record of four repeated moves: paying for a convenient review, funding doubt, ghostwriting favorable papers, and burying negative trials. Each move has left a paper trail since the 1960s.
The Documented Cases
These cases span more than 40 years, from a 1967 journal review to a 2007 guilty plea.
Sugar, fat, and a paid-for review
In 1967 the New England Journal of Medicine ran a review that discounted sugar's link to heart disease and pointed instead at fat and cholesterol. A sugar-industry trade group, the Sugar Research Foundation, had paid Harvard nutrition researchers to write it. The group set the review's objective, supplied material for it, and commented on the drafts. None of that was disclosed, because disclosure was not required at the time. That framing helped shape the low-fat dietary advice that followed for decades (Kearns 2016). Those same decades saw obesity and type-2 diabetes climb, both now driven largely by industry-built ultra-processed food heavy in sugar, fat, and salt.
The tobacco playbook
In 1969 a tobacco executive put the strategy in writing:
Doubt is our product, since it is the best means of competing with the body of fact that exists in the mind of the general public.
Facing mounting evidence that smoking caused disease, the industry mostly left individual findings alone. It funded research bodies and public relations to keep the overall picture looking unsettled, turning a largely closed question into a public controversy. Manufacturing doubt costs less than manufacturing evidence. It works because real science always has some genuine uncertainty, so manufactured doubt is hard to tell apart from it. Other industries facing inconvenient findings took up the same approach (Brandt 2012).
What the funding data show
One review pooled 75 studies that compared industry-funded research with independent research on the same products. The sponsored work reported favorable results about 27% more often and favorable conclusions about 34% more often. Standard quality checks did not explain the gap (Lundh 2017). Across the literature, who funded a study predicts which way it comes out, even after you account for how well it was designed.
Ghostwriting
In litigation files for a hormone-therapy product, ghostwritten reviews promoted its benefits and played down the breast-cancer risk that a large independent trial later confirmed (Fugh-Berman 2010). The pattern repeated for the painkiller rofecoxib, sold as Vioxx: company-drafted manuscripts had outside academics added as first authors after the writing was largely finished (Ross 2008).
The missing trials, and the fix
Of 74 antidepressant trials a regulator reviewed, nearly every one it judged positive reached print. Most of the unfavorable ones did not, or were written up to read positive. The journal record showed 94% of trials succeeding; the regulator had counted 51% (Turner 2008). Preregistration closes this gap: a trial posts its main outcome to a public registry before it starts, so no one can move the target after the numbers land.
OxyContin shows the marketing side of the same problem. Its maker sold it from 1996 with a claim that the formulation carried a low risk of addiction. Prescribing and misuse climbed together. In 2007 the company and three executives pleaded guilty to understating that risk.
The Research & Studies
Everything here is based on the research we have collected and checked, sorted into groups and ordered with the strongest evidence first. Click any claim to open the studies behind it.
Evidence And Methods
Gesponsorde studies rapporteerden gunstige resultaten ongeveer 27% vaker, over 75 studies
Over tientallen vergelijkingen kwamen door de industrie gefinancierde studies ongeveer 27% vaker gunstig uit dan onafhankelijke, en de gebruikelijke kwaliteitscontroles verklaarden het verschil niet.
Lundh and colleagues, in a Cochrane methodology review, examined studies that compared industry-sponsored with non-industry-sponsored drug and device research. Industry sponsorship was associated with more favorable efficacy results (risk ratio 1.27, 95% confidence interval 1.17 to 1.37) and more favorable conclusions (risk ratio 1.34, 95% confidence interval 1.19 to 1.51). Sponsored studies also showed less agreement between their results and their stated conclusions. The association held after accounting for methodological quality such as randomization and blinding, which means the standard risk-of-bias tools do not capture it.
Treat funding source as a piece of evidence in itself. Weight an independent replication above a single sponsor-funded trial, whatever the trial's size.
The study · 1
Lundh et al., industry sponsorship and research outcome (Cochrane methodology review) · Cochrane Database Syst Rev 2017;2:MR000033
Journals showed 94% of antidepressant trials positive; the regulator's files showed 51%
Regulators held the results of 74 antidepressant trials. Almost every trial with a good result got published, while most trials with a poor result never did, or were written up to look better than they were. Someone reading only the journals would think nearly every trial succeeded; the full record showed about half did.
Turner and colleagues compared 74 antidepressant trials registered with the US Food and Drug Administration between 1987 and 2004 against what reached the published literature. Of 38 studies the FDA judged positive, 37 were published. Of 36 studies with negative or questionable results, 22 were not published at all and 11 were published in a form that conveyed a positive outcome. The published record therefore showed 94% of trials as positive against the FDA's 51%, and pooling published and unpublished data together reduced the pooled effect size by about 32%, with individual drugs inflated between 11% and 69%.
One published positive trial can mislead on its own, because the trials that failed may simply be missing. Prefer a systematic review that searched trial registries, not journals alone, and check whether the outcome was preregistered.
The study · 1
Turner et al., selective publication of antidepressant trials and its influence on apparent efficacy · N Engl J Med 2008;358(3):252-260
In 1967 betaalde de suikerindustrie in het geheim Harvard-onderzoekers om vet, niet suiker, de schuld te geven
In the 1960s a sugar-industry group quietly paid Harvard researchers to write an influential review that pointed the finger at dietary fat, not sugar, for heart disease. The payment and the industry's hand in shaping it were not disclosed at the time.
Kearns, Schmidt en Glantz analyseerden bewaard gebleven correspondentie tussen de Sugar Research Foundation (SRF) en Harvard-voedingsonderzoekers. De SRF betaalde het equivalent van ruwweg 50,000 US dollar in 2016-termen voor een tweedelige literatuurreview, gepubliceerd in het New England Journal of Medicine in 1967, die het bewijs dat sucrose samenhing met coronaire hartziekte negeerde en de aandacht richtte op vet en cholesterol. De documenten tonen aan dat de SRF het doel van de review bepaalde, artikelen ervoor aanleverde, en concepten ontving en becommentarieerde. De financiering en de industrierol werden niet bekendgemaakt, en bekendmaking werd destijds niet vereist door het tijdschrift.
Als een review tot een nette conclusie komt, kijk dan wie ervoor heeft betaald en of de financier heeft geholpen de vraag te formuleren. Hier is de niet-bekendgemaakte sponsoring precies wat de conclusie decennialang onbetwist liet staan.
The study · 1
Kearns, Schmidt and Glantz, sugar industry and coronary heart disease research: a historical analysis of internal industry documents · JAMA Intern Med 2016;176(11):1680-1685
Industry funding tracked with pro-industry conclusions at 3.6 times the odds, across 1,140 studies
Reviewing more than a thousand studies, researchers found industry money was common in medical research and was linked to conclusions that favored the funder, and sometimes to results being held back or delayed.
Bekelman, Li and Gross systematically reviewed the evidence on financial conflicts of interest in biomedical research, drawing on 1,140 original articles. About a quarter of investigators had industry affiliations and about two thirds of academic institutions held equity in startups that sponsored research within those institutions. Pooling eight studies that examined the link between industry sponsorship and outcomes gave an aggregate odds ratio of 3.60 (95% confidence interval 2.63 to 4.91) for industry-favorable conclusions. The review also documented industry practices of restricting publication and withholding data.
Read the disclosure statement and the data-availability statement together. A funder able to delay or withhold results shapes the published record before any analysis is even reported.
The study · 1
Bekelman, Li and Gross, scope and impact of financial conflicts of interest in biomedical research: a systematic review · JAMA 2003;289(4):454-465
A 1969 tobacco memo called doubt its product to keep the science looking unsettled
Faced with mounting evidence that smoking caused disease, the tobacco industry did not deny it outright. It paid to keep the science looking unsettled, summed up in an internal line: doubt is our product. The same approach was later reused on other health questions.
Brandt's history, built on tobacco-industry documents released through litigation, describes a strategy that began with a 1953 public-relations plan and ran through industry-funded research bodies such as the Council for Tobacco Research. A 1969 Brown and Williamson memo stated that doubt is our product, since it is the best means of competing with the body of fact that exists in the mind of the general public. The strategy did not dispute individual findings; it framed a settled question as an open scientific controversy, funding work and commentary that emphasized uncertainty. Brandt argues this template of manufactured doubt was later borrowed by other industries facing inconvenient science.
Manufactured doubt can look like ordinary scientific debate. When an industry funds the voices calling a well-settled question open, treat that funding as part of the evidence you are weighing.
The study · 1
Brandt, inventing conflicts of interest: a history of tobacco industry tactics · Am J Public Health 2012;102(1):63-71
A hormone-therapy maker ghostwrote journal reviews that played down breast-cancer risk
A drug company hired writers to produce medical-journal articles favorable to its hormone therapy, then had respected academics put their names on them. Readers saw independent experts; the words came from the company's writers.
Fugh-Berman analyzed documents released in litigation over hormone therapy. The manufacturer engaged a medical-communications company to draft review articles and commentaries that emphasized the benefits of hormone therapy and countered emerging concerns, including the link to breast cancer, and then recruited academically affiliated authors to appear as the named authors. The scheme came to light in documents disclosed after the Women's Health Initiative trial reported that combined hormone therapy carried serious harms, contradicting much of the promotional literature.
A named academic author is not a guarantee of independent authorship. Read the acknowledgements and funding lines for editorial assistance or a medical-communications firm, which is where ghostwriting is disclosed when it is disclosed at all.
The study · 1
Fugh-Berman, the haunting of medical journals: how ghostwriting sold hormone therapy · PLoS Med 2010;7(9):e1000335
The maker of Vioxx drafted trial papers, then recruited academics to be the named authors
In lawsuits over the painkiller rofecoxib, marketed as Vioxx, documents showed the company had drafted journal articles internally and then found university researchers to appear as the lead authors.
Ross en collega's beoordeelden documenten die in rechtszaken over rofecoxib waren geproduceerd. Voor verschillende klinische-trialrapporten werd een manuscript opgesteld door een bedrijfsmedewerker of een door het bedrijf betaalde schrijver voordat een extern gelieerde academische auteur werd geworven om als eerste auteur te verschijnen. De documenten toonden ook gastauteurschap in overzichtsartikelen die werden gebruikt om het geneesmiddel te ondersteunen. Het artikel kwantificeerde het patroon over een gedefinieerde reeks publicaties, waarbij een onderscheid werd gemaakt tussen degenen die daadwerkelijk een artikel schreven en bylines die werden toegevoegd nadat het schrijven grotendeels was voltooid.
Vermeld auteurschap en daadwerkelijk auteurschap kunnen verschillen. Wanneer meerdere artikelen over één product een huisstijl en één sponsor delen, vraag dan wie de eerste versie schreef.
The study · 1
Ross et al., guest authorship and ghostwriting in publications related to rofecoxib · JAMA 2008;299(15):1800-1812
In 2007 pleitte de fabrikant van OxyContin schuldig aan het onderschatten van het verslavingsrisico van het geneesmiddel
De fabrikant van OxyContin voerde een agressieve campagne die artsen vertelde dat het geneesmiddel zelden verslaving veroorzaakte. Voorschriften en misbruik stegen samen, en in 2007 gaf het bedrijf voor de rechtbank toe dat het had misleid over het verslavingsrisico.
Van Zee beschrijft de marketing van OxyContin vanaf de lancering in 1996: een op bonussen gerichte verkoopmacht, uitbreiding van het gebruik ver voorbij kankerpijn, gebrandmerkte promotieartikelen, gesponsorde educatie, en een centrale claim dat de formulering met verlengde afgifte een laag verslavingsrisico droeg. De verkoop steeg van ongeveer 48 miljoen Amerikaanse dollar in 1996 tot ongeveer 1.1 miljard tegen 2000. In 2007 pleitten de fabrikant en drie leidinggevenden schuldig aan federale aanklachten wegens onjuiste etikettering van het geneesmiddel met betrekking tot het misbruikrisico en stemden ze in met boetes van ongeveer 600 miljoen Amerikaanse dollar.
De beweringen van een fabrikant over de veiligheid van een product zijn niet hetzelfde als bewijs ervoor. Weeg onafhankelijke bewakingsgegevens zwaarder dan de geruststelling van een bedrijf, vooral wanneer een product wordt gepromoot voor breder gebruik.
The study · 1
Van Zee, the promotion and marketing of OxyContin: commercial triumph, public health tragedy · Am J Public Health 2009;99(2):221-227
Nadat vooraf registreren standaard werd in 2000, daalde het succes van hartstudies van 57% naar 8%
Voordat onderzoekers hun hoofdvraag van tevoren moesten opgeven, rapporteerden de meeste grote hartstudies succes. Nadat vooraf registreren de norm werd, deden bijna geen enkele dat meer, wat suggereert dat het eerdere succespercentage voor een groot deel te danken was aan uitkomsten die werden gekozen nadat de resultaten al binnen waren.
Kaplan en Irvin onderzochten 55 grote trials gefinancierd door het Amerikaanse National Heart, Lung, and Blood Institute tussen 1970 en 2012, die geneesmiddelen of voedingssupplementen tegen harde uitkomsten testten. Het aandeel dat een significant voordeel rapporteerde op de vooraf gespecificeerde primaire uitkomst daalde van 17 van de 30 (57%) in trials vóór 2000 tot 2 van de 25 (8%) in trials vanaf 2000. De auteurs koppelen de daling aan de vereiste, die rond 2000 vaste vorm kreeg, om het primaire uitkomst- en analyseplan te registreren op ClinicalTrials.gov voordat de trial begon, wat de latere selectie van een gunstig eindpunt beperkt.
Vooraf registreren is een van de sterkste signalen dat een resultaat vertrouwd kan worden. Controleer of de primaire uitkomst werd geplaatst voordat de inschrijving begon, en wees op je hoede voor een opvallende bevinding die niet de eerst genoemde uitkomst was.
The study · 1
Kaplan and Irvin, likelihood of null effects of large NHLBI clinical trials has increased over time · PLoS One 2015;10(8):e0132382
How To Read A Study With The Money In View
Run these six checks on any study, in order.
- Find the funding and the conflicts. Scroll to the disclosure and funding statements at the end. A sponsor with a stake in the outcome does not disqualify the work; it tells you to read the rest slowly and to seek an outside replication.
- Check whether it was preregistered. Look for a public registration of its main outcome and analysis plan, filed before the trial began. Its absence is a reason to hold the finding loosely.
- Ask what was actually measured. Prefer a hard outcome that matters to a person, heart attacks or deaths, over a surrogate marker only assumed to track it. A drug can move the marker and do nothing for the outcome.
- Read the number, then the conclusion. Look at the primary result before the abstract's summary sentence. If the main outcome did not move but the abstract sounds positive, the positive tone usually comes from a secondary result. Check whether a small effect is being reported as a big one.
- Weight replication over any single study. One striking paper is a starting point. A finding that independent groups reproduce, especially groups with no stake in it, is the one to build on.
- Ask who benefits. Use it as the final filter. When a result is convenient for whoever paid for it, that is the result that most deserves an independent check.
The Limits of This Record
The named record here is small: 6 cases: a sugar trade group, tobacco, one hormone therapy, one painkiller, one antidepressant class, and OxyContin. That is enough to prove the practices are real and to teach the tells. It is not proof that every study is bought. Industry money pays for a great deal of essential research, and a sponsored study can be well designed and correct.
Scrutiny runs both ways. When a safety concern is waved off as settled, the people saying so are sometimes citing research funded by whoever profits from the product staying on the market. The tobacco industry ran exactly this play (Brandt 2012). These cases show how to read a study; they do not justify distrusting all research. Ask who paid whether a study raises an alarm or calms one.
Go Deeper
- How we grade evidence, the strength tiers behind the ladder on this topic.
- How to read a study, a single paper walked through line by line.
- Cholesterol and lipids, the field the 1967 review helped steer.
- Whole foods, the free, self-directed side of the same argument.
- The biology of aging: surrogate markers and hard outcomes, worked through.
Common Questions
Does industry funding mean a study is wrong?
No. The skew shows up across the literature as a whole; it says nothing about any single paper. One earlier review of more than 1,000 studies found industry-funded work about 3.6 times as likely to reach conclusions favorable to the sponsor as independently funded work (Bekelman 2003). The same review documented funders restricting publication and withholding data. Treat funding as a reason to read closely and to want an independent replication. It is not a reason to dismiss a single result on its own.
Was the sugar-versus-fat story really deliberate?
Yes. The internal correspondence documents the deliberate part. It does not settle the nutrition itself. That science has changed a lot in the nearly 60 years since. The lesson is narrow: hidden sponsorship made the review persuasive. It does not show that fat is harmless or that sugar by itself causes heart disease.
What is ghostwriting, exactly?
A company commissions professional writers to draft a journal article that favors its product, then recruits academic researchers to appear as the named authors. The paper reads as independent expert opinion while the content is the sponsor's. You can sometimes catch the trace in the acknowledgements, where a medical-communications firm is thanked for 'editorial assistance.'
If a treatment works, why would unpublished trials matter?
You cannot judge a treatment from the published trials alone. When the failures never publish, the trials that remain are the successes, so the average benefit reads higher than it is. For the antidepressants, the true average benefit was about a third smaller than the published trials suggested (Turner 2008). A review that also searches the trial registries beats a handful of standout papers.
What is the single most useful habit for reading past all this?
Check whether the study preregistered its main outcome, it is the fastest single tell. When advance registration became standard for large heart trials, the share reporting a benefit fell from 57% to 8% (Kaplan 2015).
Explore Related
Other pages this one connects to, by the evidence they share, the outcomes they touch, and the ground they cover.
All 9 sources on this page independently checked and cross-referenced.
Thomas Dehli, Founder & Editor, Sacred Lotus
Sacred Lotus has published Chinese medicine reference material since 2001. Integrative pages are held to the same standard as the herb and formula library: cite the source, grade the claim at its real strength, and say where the research has not looked. This page is educational and it is not medical advice. Last reviewed and updated August 10, 2026.