Translating findings from animal models to humans

1
(No Transcript)
2
AEs
3
Issues re Adverse Events

• Elicited vs. volunteered
• Nuisance AEs
• Attribution of cause

4
AEs for FAT?

• The companys standard approach
• Record any symptoms or conditions the subject
  has experienced
• _________________________________
• _________________________________

5
AEs for FAT?

• The companys standard approach
• Record any symptoms or conditions the subject
  has experienced
• _________________________________
• _________________________________
• Whats wrong with this approach?

6
An alternative?

• Since your last visit, has a doctor told you
  you had (check all that apply)
• ? A blood clot in the leg (venous thrombosis)
• ? A blood clot in the lung (P.E.)
• (for all possible diseases)
• ditto for all possible symptoms

7
An alternative?

• Since your last visit, has a doctor told you
  you had (check all that apply)
• ? A blood clot in the leg (venous thrombosis)
• ? A blood clot in the lung (P.E.)
• (for all possible diseases)
• ditto for all possible symptoms
• Whats wrong with this approach?

8
The Aspirin Myocardial Infarction Study
9
Which approach is most likely to find real AEs?

• Eliciting AEs increases rates in placebo and
  treatment groups
• Not clear whether one approach is more likely to
  detect AEs as statistically significant

10
Approaches to AEsVolunteered vs. elicited

• Pro elicited
• More sensitive?
• Easier to code

• Pro volunteered
• Catch unexpected AEs
• Fewer AEs?

11
Approaches to AEsVolunteered vs. elicited

• Pro elicited
• More AEs
• Easier to code cheaper
• Con
• Miss unexpected AEs

• Pro volunteered
• Fewer AEs
• Finds unexpected AEs
• Con
• Hard to code costly

12
Issues re Adverse Events

• Elicited vs. volunteered
• Nuisance AEs
• Attribution of cause

13
The Bunion Problem

• FIT Trial of alendronate in 6,400 women for 4
  years
• Recorded over 20,000 episodes of URIs (and
  thousands of reports of bunions!)
• Enormous data management effort and cost
• How could this be avoided?

14
How to minimize nuisanceAEs

• Elicit uncommon, plausible and important AEs
• Limit collection of minor AEs to samples of
  subjects

15
Issues re Adverse Events

• Elicited vs. volunteered
• Nuisance AEs
• Attribution of cause

16
FDA AE classifications

• Serious AEs
• Deaths
• Hospitalized
• Cancer (except skin cancer)
• Birth defects
• SAEs definitely or probably due to study drug
  must be reported to company and by the company to
  FDA in 24

17
The problem of attribution

• AEs must be classified as
• Definitely
• Probably
• Possibly, or
• Not...
• ...related to the study drug

18
The problem of attribution

• 67 year old FAT volunteer starts taking the study
  drug.
• At 1 month visit she reports an itchy rash on her
  forearms that started 7 days after starting and
  disappeared 2 days after stopping the drug.
• Your attribution?

19
Attribution

• AEs must be classified as
• Definitely
• Probably
• Possibly, or
• Not...
• ...related to the study drug

What is wrong with this system?
20
Attribution

• Attributions to drug may be as likely with
  placebo as with active drug
• This could be studied using Coordinating Center
  databases
• Volunteer?

21
The FAT AE plan

• Elicit DVT, PE, hot flashes (etc) with
• Has doctor told you that you have?
• Open ended collection for other AEs
• in a sample?
• Elicit serious AEs in all
• Hospitalizations

22
Multicenter trials Industry-sponsored trials
23
Multicenter trials

• 2 to 1,000 centers
• Usually individual practices
• Sometimes professional research centers
• Standard protocol
• Case-report forms (AKA CRFs)
• Usually in addition to records for clinical care
• Data management system
• Paper forms, electronic entry (fax, web)
• Fed to a Coordinating Center

24
Multicenter trialsThe Cast of Characters

• Sponsor
• Provides the
• Industry designs the study and owns the data
• Contract Research Organization (CRO)
• Does the sponsors bidding
• Collects clinical sites
• Develops the CRFs (usually)
• Manages the data and provides to the sponsor or
  FDA
• (Usually) Hires and supervises the site monitors
  (CRAs)

25
Multicenter trialsSite Monitoring

• To make sure every entry into the study data
  system matches the paper CRFs and entries into
  the medical record
• Make sure that the clinical site is following
  instructions
• Do not check or oversee the quality of exams or
  interviews for data collection
• Accounts for about 30 of the total cost of
  multicenter trials!

26
Multicenter trialsScientific Structures

• Steering Committee (or Scientific Advisory Board)
• Investigator Assembly
• Subcommittees
• Publications, Recruitment and Retention
• Data Safety Monitoring Board (DSMB, DMC)
• Universal in NIH-sponsored trials
• Uncommon in industry-initiated trials

27
Multicenter trials

• Why organize them for research?

28
Multicenter trials

• Why organize them for your research?
• When you need the statistical power.
• Later stages of your career.

29
Multicenter trials

• Why participate?

30
Multicenter trials

• Why participate?
• Money
• New treatments for your patients
• Can be fun/interesting
• May be able to analyze data (NIH) or publish.
• Depends on circumstance, sponsor, and your
  initiative

31
Multicenter trials Industry-sponsored trials
32
Industry-sponsored research has become a
commercial enterprise

• 70 to 80 of all industry-sponsored trials are
  done in private practices or commercial research
  clinics and run by sponsors or CROs
• lt30 of industry-sponsored research is done by
  academic centers
• About 2-3 dozen small university-based non-profit
  coordinating centers

33
Reasons to avoid industry-sponsored trials

• Bad reputation for biased results
• They control the money and data
• Little or no value for academic promotion
• Can lose money
• Can get into trouble at UCSF

34
Reasons for working with industry

• They have the drugs and resources
• NIH usually will not fund trials involving one
  drug from one company
• Example Can PTH cure osteoporosis?

35
Reasons for working with industry

• They have the drugs and resources
• The results of industry trials influence practice
• Money

36
Reasons for working with industry

• They have the drugs and resources
• The results of industry trials influence practice
• Money
• 3 to 20,000 per patient per year
• These are discretionary funds
• Opens the door to other types of funding
• Available in months (vs 1-2 years for NIH)

37
Reasons for working with industry

• They have the drugs and resources
• The results of industry trials influence practice
• Money
• The experience can be impactful and educational,
  if managed well

38
Individual Cases
39
Dr. S

• Junior faculty, investigator on mentors grant
  from Pfizer to test a drug for incontinence
• Invited to attend meeting at Miami resort to
  attend an advisory board on a new selective
  estrogen receptor modulator that might influence
  incontinence.
• Trip cost 3,500 consultation fee 1,500
• Should you attend?

40
Vote on a UCSF Policy

• Academic Senate Committee is split on a UCSF-wide
  policy governing allowable payments from industry
  to faculty for consulting and honoraria (for
  speaking) and related expenses.
• 1. Faculty should be allowed to receive
  10,000/year from sponsors of their research
• 2. Faculty should not be allowed to receive any
  payments from sponsors of their research

41
Vote on a UCSF Policy

• Academic Senate Committee is split on a UCSF-wide
  policy governing allowable payments from industry
  to faculty for consulting and honoraria (for
  speaking) and related expenses.
• 1. Faculty should be allowed to receive
  10,000/year from sponsors of their research
• 2. Faculty should not be allowed to receive any
  payments from sponsors of their research
• 3. This policy should be applied only to research
  involving human subjects

42
Is this the real issue?
43
A Scandal

• Results of a trial of a new AIDS drug are
  negative (except in a small subgroup)
• A lead investigator on a trial of a new drug for
  AIDS writes a paper emphasizing the negative
  results. The company blocks publication
  insisting on including the results from the
  subgroup.
• The investigator publishes the paper over the
  objections of the sponsor. The sponsor sues.
• What is wrong with this approach?

44
MORE Trial Results

• 7,705 women treated with raloxifene. Primary
  outcome-reduced vertebral fractures 40.
• Main paper in JAMA on 3-year results
• Graphs and tables regarding vertebral fracture
• Mention in text no effect on other fractures
• Paper on 4-year results
• Continued reduction in vertebral fracture
• No mention of other fractures

45
Bias

• Financial incentives bias the research
• Biased analysis and reporting of results
• (Bekelman industry sponsorship likelihood of
  pro-industry result about 3.6-fold.)
• Biased trial designs
• Biased selection of research questions

46
Fundamental problems

• The sponsor has
• data
• data analysts
• medical writers

47
Fundamental problems

• The sponsor has
• data
• data analysts
• medical writers
• Usually, sponsors have no checks on what they can
  publish, except the first author
• First authors often too busy or inexperienced

48
A set of principles

• Always
• Get the data (or unfettered access)
• Get a data analyst (and medical writer?)
• Have a publications committee with a voting
  majority from outside
• Make your own slides
• Report your sponsors and support in presentations
  and papers

49
What is the responsibility of a publicly
sponsored university?

• Work with industry to influence the rigor and
  objectivity of research
• Be a model of ethical principles, uninfluenced by
  ties to industry?