Scientific Integriy in Medical ResearchPartnerships and Ethical Implications

FEAM Conference – December 15th Lisbon

J. Lobo Antunes MD, Ph.D.President Portuguese Academy of Medicine

“Scientific” has become an all purpose term of epistemic praise meaning “strong, reliable, good”

and yet...

like all human enterprises it is thoroughly fallible, imperfect, uneven in its achievements, often fumbling, sometimes corrupt, and of course incomplete

Edward Gibbon(1737-1794)

“The History of the Decline and Fall of the Roman Empire”

“Consider me not as a

contemptible thief but as an

honest and industrious

manufacturer”

Gregor Johan Mendel(1822-1884)

Louis Pasteur(1822-1895)

Robert A. Millikan(1868-1953)

Some key ideas

• Legally scientific fraud is a deliberate misrepresentation of truth (misconduct may be a better term)

• “Sloppy science”• Contradicted or misguided

interpretations• Mistakes• Poor scientific and

unprofessional practices• Negligence

• It is different from

Wrong observations Wrong analysis Undeclared conflict of interest

Publication bias Undeserved authorship

Supressing data Plagiarism

Falsification

Fabricacion

Non-intentional

Intentional

Error Misconduct Fraud

M. Nylenna, S. SimonsenLancet 367:1882, 2006

Science does not exist until it is published.

Drummond Rennie. Lancet 1998;352:SII18

Publications are fundamental units of information exchange, proof of productivity and creativity, and bases for future research and development

Academic promotion

Productivity (quantity)

Independence (first or senior authorship)

Significance (impact factors)

“The Audit Society”

• 27% of the scientific papers are never cited• Papers published

• Papers published in Nature 1999

citations in 2001 – 10 % (80 papers) = half of citations

1955 – 198755.7% 79,9%

A few interesting numbers…

30 million1 citationno more than 4

If 2/3 of accepted papers were replaced by 2/3 of the rejected, the quality of the journal would not alter(Adair et al. Phys Rev Letters 43:1969, 1979)

There are more >16000 medical journals

Authors/article and Editors do NEJMManuscripts submitted to NEJM

Drummond Rennie. Lancet 1998;352:SII18

972 authors

2 words/author

Committee on Publication Ethics (COPE)

• Falsification• Fabrication• Plagiarism • Failure to get ethical approval• Not admitting that some data are missing • Ignoring outliers without declaring it • Not including data on side effects on a clinical trial• Conducting research without informed consent• Publication of post-hoc analysis without declaring it• Gift /honorary authorship• Not attributing other authors• Redundant publication• Not disclosing conflicts interest• Not attempting to publish completed research• Failure to do an adequate search of existing research before beginning new

research• “Shotgunning” - simultaneous submission of a manuscript to more than one journal.

TAXONOMY OF MISCONDUCT

Fraud in Publishing

• Major research institutions and high impact journals

• Biological sciences

• Clinical research

More common

You catch them in the “NET”

What happens after

• Retraction – ignore it • Expression of concern – we are looking into it• Correction substitute information

papers continue to be quoted after retraction

retracted 2003Example:Jam Hendrik Schön, Nature published 2000

cited 17X after that!

but

The Peer-review system

JAMA 9%

Academic Medicine 15%

Nature 5%

but indispensable

Remote

Mysterious

Crude

Understudied

– Confirmatory bias

Bias against negative results

Give disproportionate credit to the already famous

Orientation and theoretical persuasion

Conflicts of interest [competitors / antagonists]

Agreement between referees 10-15%

86% of unpublished trials have negative results

45% of published trials have negative results

The politically correct

Blinding is not the solution. The authors can be guessed in 46% of manuscripts!

(JAMA 272: 143, 1994)

Gate-Keepers

Rate of acceptance

The pitfalls

The Malefices of Covert Duplicate Publication

Ondasetron on post-operative emesis

9 trials published in 14 further reports duplicating data from 3325 patients

Inclusion of duplicate data in meta-analysis led to a 23% overestimation of the drugs antiemetic efficacy

Tramer et al. Brit Med J 315:635, 1997

Example

Pressure to publishUnhealthy competition?

• “They chose reviewers who they knew to be positive (...) They did not allow their experiments to be reproduced” Robert Laughlin

(Nobel Prize physics)

• “Given the exciting claims made by the papers, we were certainly hoping that the outcomes would be positive”Karl Ziemeli(Chief physical sciences editor, Nature)

The Schön Scandal

The Editors’ Pressure

Manipulation of the impact factor of the journal, encouraging the citation of other papers published in the journal (*)

and yet

“Impact factors tell you more about sociology of science than about science itself”

S. Brenner

(*) (M. Farthing, Science and Engineering Ethics 12:45-52, 2006)

Date withholding

Protect priority [“races”]Strictures of commercial fundingMaterial and financial costs of responding to requests for

biomaterials

Scientists in trainning are discouraged to show data42% genetic38% of OLS

Blumenthal et al Academ Med 81: 137, 2006

Industry support of biomedical research

USA1980 32%2000 62%

- Lead authors 1 every 3 articles hold relevant financial

interests.*

- In biomedicine, with rare exceptions, is the private sector,

not academics that develops diagnostic, therapeutic and

preventive products and brings them to market.

- 2/3 of academic institutions hold equity in “start-up”

businesses that sponsor research by their faculty

* Quoted in Bekelman et al. JAMA 289:454, 2003

- Industry supported faculty is as productive as those who do not receive support

- more productive commercially - 2 x trade secrecy or withhold results from

colleagues -encourage research with commercial

applicability and may reduce fundamental research.

Blumenthal et al. N Engl J Med 335:1734, 1996

Industrial support and academic productivity

Academic investigators –

Industry –

Competing goals in medical researchPublication in peer-reviewed journals

Approval and marketing of drug.Without approval, publication is notworth a cent. Publication in prestigious journalsimportant for the marketing

No drug company gives away its stockholders’ money in an act of desinterested generosity

Journal of Commercial Molecular Biology

Journal of Commercial NeurobiologySidney Brenner “My life in Science”

- Death of volunteer in phase I gene therapy trial: doctor and institution had financial interest in therapy

- Publication biases

- Authors whose work support safety of calcium – channel antagonists had more frequently financial ties with industries.*

- Results favoring new therapy over traditional one are more likely if study is funded by therapy manufacturer.**

- 5% of industry supported pharmoeconomic studies of cancer drugs reached unfavourable conclusions; non funded studies reached the same conclusion in 38% of the studies.***

* Stelfox et al. – N Engl J Med 338:101, 1998** Davidson – J Gen Int Med 1: 155, 1986*** Friedberg et al. – JAMA 282: 1453, 1998

Concerns about industrial funding of medical research

+++Teirstein

++Williams

++Kereiakes

+++Fitzgerald

+++Popma

+++Leon

+++Moses

StockholderFinancingSpeakerConsultant

0

10

20

30

40

50

60

April 18 April 25 May 2 May 9

Am

ount

(do

llars

)

Therapeutic effect. A news report on angiostatin and endostatin’s promise did wonders for WEntreMed’s stock

- “Does declaration of competing interests affect

reader’s perceptions? A randomized trial”*

Results of study on impact of pain in herpes were

found less interesting , important, relevant, valid and

believable when the authors were employees of

fictitious pharmaceutical company than with

ambulatory care centers.

* Chaudhry et al. B M J 325:1391, 2002

Conflict of Interest

Biomedical Research, what is the public interest?

1. The research that it supports is for the search of truth,

uncontaminated by any bias

2. Discoveries with potential therapeutic benefit are rapidly

translated into practice by clinical trials.

3. Participation in development of new therapies will be safe, with

full informed consent, and access to outcome and follow-up.

4. Right to know about potential side effects that might influence

decision to participate

5. Must be assumed that decision to ask patients to participate or

the assessment of risks will not be determined by pressure on the

investigator.

J B Martin et al. New England J Med 343:1646, 2000

A convenient omission

A 4x increase in heart atacks was ommitted

The journal sold 929.000 offprints(Revenue $ 679.000 to $ 836,000)

What does academy have to do?(little scholarship on this topic!)

- Protection of human participants safety and welfare - Academic freedom

- Objectivity - Data integrity- Right to publish- Financial and non financial incentives should

address institutional, senior and junior investigator needs

- Separate human research responsibilities from investment management and technology transfer*

* Task force Am Ass Med Colleges 2003

(The Editors of Ann Int Med, JAMA, New England J Med, Canad MAJ, J Danish M A, Lancet, Medline, etc, Sep 2001)

- When authors submit manuscript they are responsible for disclosing all financial and personal relationships that might bias their work

- Researchers should not enter in agreements that interfere- Their access to the data- Ability to analyze data independently- Prepare manuscripts- Publish them

Sponsorship, authorship, and accountability (1)

- Should describe the role of the study sponsor- Collection, analysis and interpretation of data- Writing the report: “The non-author writer syndrome”,

the guest author.- Avoid selecting external peer reviewers with C.I. (e.g.

same department)- Reviewers must disclosed C.I.

(Drug therapy reviews)- Editors most have no personal, professional or financial

involvement in any issues they might judge.

Sponsorship, authorship, and accountability (2)

How to improve

• Research – Funding agencies establish research grant programs to identify, measure, and assess those factors that influence integrity in research.

• Institutional Commitment – Institutions to develop and implement comprehensive programs

• Education – Effective educational programs • Self-assessment – Implement self-assessment and

external review process. If possible this should be part of existing processes accreditation

[Adapted from “Integrity in Scientific Research. Institute of Medicine. National Research Council, 2002]

“Many people say that is

the intellect which makes

a great scientist.

They are wrong: it is

character”.