Medical journals, drug policies and the pharmaceutical industry

1
Medical journals, drug policies and the
pharmaceutical industry

• Fiona Godlee
• Editor, BMJ
• London School of Economics
• January 22 2007

2
What Im going to talk about

• The structure of the relationship between
  journals and the pharmaceutical industry
• Hows it going some examples of the current
  problems we face
• The BMJs position
• Some lessons and solutions

3
Whats the role of the drug industry?

• Journals have devolved into information
  laundering operations for the pharmaceutical
  industry
• Richard Horton, New York Review of Books, March
  11, 2004

4
Structure of the relationship - 1

• A source of revenue
• Journal business model
• Subscriptions
• Classified advertising
• Pharmaceutical advertising
• Reprints
• Sponsorship

5
Structure of the relationship - 2

• A source of content
• Original research
• Review articles
• Supplements

6
Two thirds of trials in major journals are funded
by the drug industry (Egger M, et al, BMJ 2001
323 773)
7
Fraught with potential conflicts

• Safeguards
• Chinese wall
• Peer review
• Conflict of interest statements
• Transparency policies
• Editorial oversight of advertisements

8
(No Transcript)
9
The methods of HARLOT Richard SmithHow to get
the results you want

• Trial against placebo
• Trial against treatment known to be inferior
• Trial against low dose of competitor
• Trial against high dose of competitor (to give
  advantages on toxicity)
• Equivalence or non-inferiority trial that is too
  small to show a difference from competitor drug
• Use multiple endpoints and select the ones that
  give positive results
• Subgroup analyses and select positive results

10
The methods of HARLOT Richard SmithPublishing
strategies

• Suppress (dont write up) negative studies
• Publish positive studies more than onceperhaps
  in supplements
• Do multicentre trial and publish results of
  individual centresperhaps selectively
• Publish different outcome measures at different
  times
• Publish different follow periods at different
  times3 month results, one year, two year
• Publish positive results in major journals and
  negative or neutral results in minor journals
• Combine results of trials in ways that are
  favourable

11
Some examples

• Data manipulation - Vioxx
• Ghost writing/undeclared competing interests
• Duplicate and redundant publication
• Publication bias

12
(No Transcript)
13
Vioxx (rofecoxib)

• Introduced by Merck in 1999 as an effective safer
  alternative to non-steroidal anti-inflammatory
  drugs for pain in osteoarthritis
• Merck now faces claims from 30 000 people who
  suffered cardiovascular events while taking Vioxx

14
Vioxx the story

• Early concerns that rofecoxib increased thrombus
  formation
• 1996-7 a study sponsored by Merck found signs
  that the drug altered the balance between
  thromboxane and prostacyclin
• Merck officials persuaded the academic authors to
  soften their interpretation

15
Vioxx licensing and promotion

• FDA application for Vioxx (1998) was not designed
  to pick up cardiovascular risk
• 9 intervention studies
• small, short treatment periods
• enrolled patients at low risk of CV disease
• no standard procedures for collecting
  cardiovascular outcomes
• Merck pooled data from these studies and used the
  results to promote rofecoxibs cardiovascular
  safety

16
Vioxx the VIGOR study

• Started in 1999
• Intended to show that Vioxx had fewer GI side
  effects than naproxen for treatment of rheumatoid
  arthritis would mean a new indication
• Over 8000 patients
• No standard operating procedure for collecting
  information on CV events
• No cardiologist on the safety monitoring board

17
Vioxx the VIGOR study

• Interim analyses showed higher CV risk in one
  group, but decision to continue
• Undisclosed conflicts of interest among board
  members head of VIGOR board awarded a two year
  consulting contract two weeks before the trial
  ended, and as the trial was concluding disclosed
  family ownership of Merck shares worth 70 000

18
(No Transcript)
19
GI outcomes VIGOR
17/1000
NEJM, 11/00
20
Vioxx VIGOR study

• CV events obscured
• Report was of interim analysis
• Different end points for GI and CV events (GI
  events counted for one month longer than CV
  events) not described in the publication in
  NEJM favoured GI benefits and understated CV
  risks
• 3 additional myocardial infarctions in treatment
  group occurred in the missing month
• Authors dismissed any additional CV risk by
  suggesting that the difference between the two
  drugs was because naproxen had a
  cardio-protective effect

21
Promotion

• Bought nearly 1 million reprints of the NEJM paper

22
(No Transcript)
23
Comparison of MI Rates Among Subjects Receiving
Placebo vs Rofecoxib or Celecoxib
2001
24
Thromboembolic events
Approve study NEJM 2004
25
Vioxx withdrawn
26
(No Transcript)
27
Lessons from VioxxHarlan Krumholz et al, BMJ 20
January 2007

• Bad news for industry, academics, journals, and
  the public
• Merck has lost vast sums of money, shareholder
  value, and reputation
• Merck conducted the trials, stored and analysed
  the data internally, paid academic researchers as
  consultants to the investigative teams and safety
  boards, and maintained heavy involvement in the
  writing and presentation of the findings

28
Undeclared competing interest/ghosts and guests

• RCT of Vioxx versus Naproxen Lisse et al. Ann
  Intern Med 2003 139 539-46
• New York Times,24 April 2005
• Merck designed the trial, paid for the trial,
  ran the trialMerck came to me after the study
  was completed and said, We want your help to
  work on the paper. The initial paper was written
  at Merck, and then sent to me for editing.

29
Lessons from Vioxx

• The journals presented the findings and doctors
  prescribed the drugs
• Nearly 107 million prescriptions in the US
  1999-2004
• Doctors and patients had no way opportunity to
  balance risks and benefits

30
Lessons from Vioxx

• Define a set of principles code of conduct
• Insist data are stored in academic sites,
  analysed by non-company investigators, and
  eventually made accessible to the public for
  scrutiny
• Independent audits
• Independent data and safety monitoring boards,
  governance not in control of the company
• Industry should not be allowed to select who
  serves on these boards or allowed to compensate
  members after their service

31
Lessons from Vioxx

• More intense scrutiny of such studies by journals
• Ghost writing must be exposed and the academic
  authors who take part must be penalised

32
How frequent is guest authorship and ghost
writing?

• Survey of officials responsible for negotiating
  research contracts at 122 medical schools in the
  US
• Half of 107 respondents reported that their
  institution would allow agreements that allowed
  sponsors to draft manuscripts while limiting the
  academic investigators ability to revise them
• Mello et al. NEJM 2005 352 2202-10

33
How frequent is guest authorship and ghost
writing?

• Survey of NIH funded investigators
• 15.5 admitted tinkering with the design or
  interpretation of research to suit a sponsor
• 10 admitted inappropriate assignment of
  authorship credits
• Martinson et al. Nature 2005 435 737-8

34
Duplicate publication and publication bias
35
Duplicate publication and salami slicing - the
Odansetron story

• 84 trials that included information on 11 980
  patients
• In reality only 70 trials and 8645 patients (17
  of the studies had been published more than once
  and the number of patients had been inflated by
  28)
• Impossible to tell from published studies
• Four pairs of identical trials were published by
  completely different authors without any common
  authorship (Misconduct)

36
Effectiveness of odansetron
37
Publication bias

• Positive studies are more likely to be
• Published Stern and Simes, BMJ 1997
• Published faster Ioannides, JAMA 1998
• Published in higher impact journals Easterbrook
  et al, Lancet 1991 Tierney and Stewart, 1997
• Cited old style narrative review articles

38
Publication bias and the drug industry

• Rochon et al, Arch Intern Med 1994
• Study of 56 industry funded studies of
  non-steroidal anti-inflammatory drugs
• Not one was unfavourable to the company
• All showed the manufacturers drug to be as good
  as the comparison or better in terms of efficacy
  or toxicity
• Rochon PA, Gurwitz JH, Simms RW, Fortin PR,
  Felson DT, Minaker KL, et al. A study of
  manufacturer supported trials of non-steroidal
  anti-inflammatory drugs in the treatment of
  arthritis. Arch Intern Med 1994154 157-63.

39
Publication bias and the drug industry

• Lexchin, Bero et al. BMJ 2003
• Found 30 meta-analyses comparing industry and
  non-industry funded studies
• Overall, studies funded by companies were four
  times more likely to have results favourable to
  the sponsor than studies funded by others
• No difference in the quality of the research
• Lexchin J, Bero LA, Djulbegovic B, Clark O.
  Pharmaceutical industry sponsorship and research
  outcome and quality systematic review. BMJ 2003
  326 1167-70.

40
Publication bias and the drug industry

• Gotzche et al, BMJ 2006
• Compared pairs of meta-analyses looking at same
  drug in same population published within 2 years
  of each other, one Cochrane, one non-Cochrane
• Found 24 pairs (8 industry sponsored, 9 unknown,
  7 no support or non-industry)
• Median quality scores (from 0-7) 7 for Cochrane,
  2 for industry, 2 for unknown
• All industry sponsored reviews recommended the
  experimental drug without reservation vs none of
  the Cochrane reviews

41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
Some solutions?
46
Science is intensely sceptical about the
possibility of error but totally trusting about
the possibility of fraud. Arnold Relman Former
editor, NEJM
47
What are journals doing? Improving and extending
peer review

• Open peer review
• Transparency
• John Bailor, Disclosure is almost a panacea
• conflict of interest statements
• statements of duplicate publication
• contributorship statements
• quick and prominent corrections
• good post-publication peer review

48
What are journals doing?Improving and extending
peer review

• Additional statistical review for drug industry
  trials (JAMA)
• Space on the web (ELPS)
• Emphasis on harms as well as benefits of
  treatments
• Education and training, authors and peer
  reviewers
• Penalties for misconduct
• Trial registration
• Protocols

49
Effect of ICMJE deadline on weekly trial
registrations at clinicaltrials.gov 2005
What are journals doing 1Improving and
extending peer review
(Source Zarin et al. N Engl J Med.
20053532779-87)
50
Whats the future?

• Further attempts to understand and improve peer
  review
• Training for authors and peer reviewers
• Professional peer reviewers?
• Further extensions of trial registration
• Commitment to publish based on the research
  question and the methods, whatever the results
• Better education of junior researchers
• Greater penalties for people or companies who
  commit fraud

51
But will any of this really help?

• A more radical future
• Stop using opinion leaders with any financial
  conflicts of interest?
• Publication of entire data sets why not?
• Stop pharmaceutical companies directly evaluating
  their own products

52
Proposal Drug companies should not be allowed to
directly evaluate their own products

• Marcia Angel (among others)
• Oversight of clinical trials should be
  transferred to a public agency
• Researchers with whom this agency contracts
  should not be allowed to have financial ties to
  industry
• The researchers would retain the data and take
  full responsibility for the design, analysis, and
  publication of the results
• Then there would no need to audit drug company
  data

53
Conclusions

• We all want and need effective safe drug
  treatments
• Pharmaceutical industry is a global business with
  vast sums of money at stake and huge lobbying
  power internationally
• Medical journals are inextricably linked to
  pharmaceutical companies in ways that present
  enormous potential for conflict of interest,
  error and fraud
• There is lots more we could do to keep industry
  on its toes
• The aim should be for collaboration for the good
  of patients

54

• Thank you
• fgodlee_at_bmj.com