SF Coordinating Center Multicenter Trials Pilot Studies Steve Cummings, MD

SF Coordinating Center

Multicenter TrialsPilot Studies

Steve Cummings, MD


Outline

Multicenter studiesIndustry-sponsored trialsPilot studies for trials


Dr. D

Large dermatology practice Interest in psoriasis Wants to participate in multicenter trials

of new treatments


Multicenter studies

AnatomyPhysiologyHow to work with themExample: MrOS


Anatomy: NIH-style studies

Data Coordinating Center - Makes forms & manuals - Compiles data for analysis

Clinical sites - Recruit subjects - Transmits data to CC


Anatomy: NIH-style studies



Other functions depend on the studyDesign, analysis


Physiology



Other functions depend on the studyDesign, analysis


Physiology: NIH-style studies

Overall PI Got the grant; leads the study

Clinical site PIs Run the sites; variable say re: the study

Steering Committee Represents all sites and CC Sets the policies; ultimate arbiter


Physiology: NIH-style studies

Publications and Analysis CommitteeUsually a subset of the SCommDevelops guidelines for approvalReviews and approves proposals for

analyses, papers (and ancillary studies)

Democratic


Anatomy: Industry sponsored trials

Sponsor - Designs the trial - Chooses sites - Analyzes data

Clinical sites - Recruit subjects - Transmits data to CRO

CROCRO (hired by sponsor) - Creates the forms - Collects/collates data - Monitor sites

Sponsor


Physiology: Most industry trials

Sponsor does it all Designs study, develops forms and

protocols Analyzes data; selects authors; write the

articles Investigators

Paid to recruit patients Investigator meetings No representation in decisions


Physiology: A few industry trials

More collaborative with investigators ‘Lead’ investigators involved in design Steering Committee (a few investigators) Publications Committee

Investigators Most still just paid to recruit patients Investigator meetings

Example: AMG 216 (FREEDOM Trial)


A few (ideal) industry trials

Steering Committee Minority representation from sponsor Real authority (official charter)

Changes in design. Stopping the trial.

Publications and Analysis Committee Minority representation from sponsor Publication guidelines (a contract) Develop and review analysis plans Review and approve papers


Reasons to be a siteNIH multicenter study

Participate in scientific investigation Use data; publish Support staff


Reasons to be a site in the typical industry trial

Funding for you and staff Profit (industry funds are worth more than

NIH funds) New treatment alternatives to patients A chance for correlative science? A step toward larger roles with the sponsor Appear on publications Value for promotion?


What we (SFCC) look for in a site

Past performance Rapid recruitment of large numbers Responsive and involved colleagues

(Avoid complainers and nonresponders) Scientific expertise High quality work (clean data)


What Industry looks for in sites

Past performance Number recruited Number recruited Not a ‘problem site’

Current capability to recruit Reputation of the investigator

Thought leader Active contributor: expertise, consulting,

productive writer.


Would you like to be a site?

Get a reputation Become a local or national leader Recruit successfully

Borrow a reputation Have a mentor or experienced colleague

make contacts (Industry: Contact your local ‘rep’)


Competitive recruitment

Trials have set recruitment goal Industry pays per subject Those who recruit most make most

Possible to make $$$ Possible to lose $$$ The case of the “Lone Academic”

Top recruiter is often 1st author


Disadvantages of being a site in industry trials

Can lose money Distraction from other types of work or

research Hassles

Site monitoring FDA inspections


How to be a great clinical site

An excellent study coordinator A registry of patients (subjects)

Characteristics Consent to be contacted for research

Excellent responsive support Contracts office IRB


Publications and industry


Publications

Sponsors own the data Data analyzed and drafted by

The sponsor statistician and medical writer A hired “CME” firm

External authors invited to author a paper Provided analyses or draft May write a draft or ‘edit’ the sponsors draft


Universities: investigators must have ‘access to the data’

“Access to data” is meaningless Sponsors employ the data analysts Sponsor’s team directs and selects

analyses Authors are rarely involved (or skilled) in

data analysis, rarely look at data output


Universities: investigators must have ‘the right to publish’

“Right to publish” is usually meaningless Your data is one small part of the larger

trial; would be wrong to publish a part Hard to analyze and publish without

support for analysis and writing Investigators should not have a ‘right to

publish’ results that are wrong


An ideal approach to publications from from industry

Insist on a Publications Committee With guidelines as a contract Reviews plans and papers before submission Minority voting representation from the sponsor

Data available for analysis by independent statistician-analyst Some journals are insisting

Writing the paper No medical writers? Or writers assist investigators


Pilot studies


Why pilots

Feasibility To determine if you can do it. To convince funders you can do it

To improve the study More efficient Prevent screw-ups


Example: T trial

Green tea has been associated with lower risk

of breast cancer Inexpensive and safe Proposal

Large randomized trial to test green tea extract vs. placebo X 3 years

Randomize women who attend mammography units and have high breast density to capsules of placebo or green tea


T trial

Easy to recruit from the Center Breast density can be assessed on the

mammogram No need for exclusions 45,000 women get mammography / year

Periodically mail study supplements to home Follow-up by internet and future

mammography visits Ascertain cancer by emailed or phoned self-

report or periodic search of the SEER registry


T trial

Idea has been enthusiastically received A group is meeting to develop a

proposal to NCI


Pilot studies

To test feasibility of recruitment Review records Survey participants Conduct a ‘pilot trial’

To test the feasibility of interventions, procedures and measurements How much green tea to have an effect? Side effects of large doses? Will women adhere?


Pilot studies

To determine power Variability of the main measurement Rates of outcome in real participants

To estimate costs and need for staffing And others…


How big should pilots be?

“It depends” Sample size estimates sometimes

useful (Hologic wanted ‘100’ when 30 will do)

Traditions and guesses NCI phase I trials start with 3 Dry runs: 3 to 10 of same age, health


Do pilots need control groups?

It depends New treatments

Determining the highest tolerable dose of green tea?

Determining the effect of green tea on breast density?

Feasibility? Generally irrelevant (e.g., potential participants’

eligibility and willingness to enroll).


Funding for pilots

Local institution Academic Senate…

Disease foundations Industry reps Mentors…

Can often be done with volunteer help


IRB approval for pilots

Pilots involving people (or medical records) generally need IRB approval

Testing questionnaire or strength testing devise with staff?

Documents

SF Coordinating Center Multicenter Trials Pilot Studies Steve Cummings, MD