Conflicts of interest

1
Conflicts of interest

• Bernard Lo, M.D.
• September 10, 2009

2
Views of audience

• Student loan officer at college receives annual
  payment from bank. Should this relationship be
  disclosed to college, students and parents?
• Yes
• No
• Not sure

3
Views of audience

• Should such a relationship be banned (even if it
  is disclosed)?
• Yes
• No
• Not sure

4
Views of audience

• Does the amount of payment matter? Should the
  relationship be banned if
• 500 a year
• 5000 a year
• Percentage of loans by bank

5
Views of audience

• Unlike competitor, bank does not defer repayment
  while student works for Teach for America, VISTA,
  etc. Should the relationship be banned
• Yes
• No
• Not sure

6
What is a conflict of interest?

• Special relationships of trust, dependency
• Primary interest in well-being of clients,
  patients, integrity of research
• Secondary interest in personal gain (,
  professional advancement)

7
What is a conflict of interest?

• Unacceptable risk that primary interest unduly
  influenced by secondary interest
• Not proven bias

8
Points to consider

• Arrangement may be problematic even if no bias
• Hard to tell if bias exists
• Disclosure necessary but not sufficient
• May not know what risks are or how to reduce them

9
Points to consider

• Some arrangements should be managed or prohibited
• Permit honorarium for talk to meeting of all banks

10
Points to consider

• Goal of COI policies to prevent bias and loss of
  trust, not respond to misbehavior
• COI policies should consider
• Risk of relationship
• Burdens and unintended adverse effects of policies

11
Spiral CT for lung cancer screening

• Advocates have pushed states to pay for screening
• NY bill would pay only at Cornell

12
Spiral CT for lung cancer screening

• Screened gt31,000 high risk persons
• F/u Stage 1 cancers who had surgical resection
• 10 year estimated survival 92
• Could prevent 80 of deaths from lung cancer

13
Methodologic concerns

• No concurrent control group
• Lead time bias
• Few patients followed for 5 years
• Assume at all persons found to have cancer would
  die without screening
• Ignore false positive scans

14
Spiral CT for lung cancer screening

• 200 million NIH trial comparing CT screening vs.
  CXR
• Allowing hundreds of thousands to die is
  unethical

15
Financial relationships

• 3.6 million from parent company of cigarette
  manufacturer
• Grant to Foundation for Lung Cancer
• PI, co-PI as officers
• Dean, vice-chair of overseers as directors
• Didnt know origin of money
• Gift announced publicly

16
Financial relationships

• Not disclose 10 patents on CT screening in some
  articles and CME lectures
• Claim not relevant
• Cornell receives royalties

17
Financial relationships

• Not disclose funding as required
• To NEJM
• Would not have published
• On American Cancer Society grant
• Prohibit funding if also tobacco support

18
Concerns about COIs

• Are errors due to
• Honest disagreement, lack of skill, lapse, slip?
• Self-interest

19
Concerns about COIs

• Voluntary disclosure ineffective
• Impetus for public disclosure of payments
• Institutions have their own COI
• Bright people find loopholes

20
Biederman case

• Research supporting diagnosis of bipolar disorder
  in children and treatment with antipsychotics

21
Biederman case

• Ties to manufacturers of psychiatric drugs
• Research funding and consulting
• Director of J J Center for Pediatric
  Psychopathology Research
• Move forward the commercial goals of J J

22
Biederman case

• Failed to report 1.4 million in income from drug
  companies, as required by
• Universities
• NIH

23
The New York Times

• Expert or shill?

24
Concerns about COIs

• Disagreements and errors inevitable
• Due to reasonable differences of opinion, slip,
  or financial profit?
• Disclosure not remove risk of bias

25
Interactions between physicians and industry
26
Academia-industry collaboration to develop new
therapies

• Antiretrovirals for HIV infection
• Pulmonary surfactant
• Imatinib, tyrosine kinase inhibitor for CML
• Rituximab, MAB against CD20 on B cells

27
(No Transcript)
28
Concerns about industry-sponsored clinical trials

• Academic investigators lack
• Access to data
• Independent statistical analysis
• Drafting of papers by medical writers

29
Concerns about industry-sponsored clinical trials

• Favorable conclusions
• Not report negative findings

30
Association between funding and conclusions

• Investigator ties to manufacturer
• 3.6 times more likely to find drug effective
• Sponsored by manufacturer
• 4.0 times more likely to find drug effective
• JAMA 2003 289 454
• BMJ 2003 327 1167

31
Explanations for association

• Publication bias
• Investigators
• Journals
• Less rigorously designed
• Evidence not support this

32
Explanations for association

• Manufacturers sponsor trials that are likely to
  succeed

33
Not report negative findings

• Celecoxib
• 6-month data reported, showed benefit.
• 12-month data known, no benefit
• BMJ 2002 324 1287
• JAMA 2001 2886 2398

34
Not report negative findings

• SSRIs in children
• Aprotinin
• Ezetimibe plus statin

35
Mission of drug companies

• Drug companies have to continue to be successful
  businesses But their primary mission is products
  that save lives and improve lives.
• This is an area thats different from ice cream,
  bubble gum, and automobiles.

36
(No Transcript)
37
Responses to conflicts of interest

• Disclosure
• To university
• Annually
• With grant submission (NIH requirement)
• To IRB
• To research subjects

38
Responses to conflicts of interest

• Disclosure
• To public
• Required for some companies by legal settlements
• Voluntary by some companies
• Some state laws
• Proposals in Congress

39
IOM recommendation

• Public website
• All payments from drug, device, biotech companies
• To MD, researcher, medical institution
• Can aggregate all payments

40
Arguments for disclosure

• Sunshine the best of disinfectants
• May deter questionable behavior
• May also deter desirable behaviors

41
Concerns about disclosure

• Not specific or standardized
• Categories like consultant ambiguous
• May be misinterpreted
• Discrepancies due to different requirements
• Not prevent undue influence or bias
• Necessary but limited first step

42
Disclosure of conflicts of interest

• Necessary but limited first step

43
Responses to conflicts of interest

• Disclosure
• Management of relationship
• Prohibition

44
Clinical trial investigators with significant
conflicts of interest

• Default or presumption is that may not
  participate
• Patent, royalties
• Stock options
• Management position in sponsor

45
Clinical trial investigators with significant
conflicts of interest

• Exceptions permitted
• For safety of participants
• Inventor of medical device
• For validity of study
• Only lab that can conduct assay

46
Clinical trial investigators with significant
conflicts of interest

• Limit role of conflicted investigator
• Not tasks that others can carry out
• Recruitment of subjects
• Data analysis
• Drafting of manuscript
• May not be PI
• Alternatively, may end conflicting role

47
(No Transcript)
48
Responses to conflicts of interest

• Disclosure
• Management of relationship
• Prohibition
• Reduce undue influence and bias

49
Reporting on clinicaltrials.gov

• Required
• For FDA consideration
• By leading medical journals
• Before trial begins
• Protocol

50
Reporting on clinicaltrials.gov

• After completion
• Basic results
• Serious adverse events
• Important exceptions

51
(No Transcript)