Practical Biostatistics for Clinical Trials, or How to Find and Use Your Biostatistician Scott Brown, Ph.D. NAMSA
Dec 22, 2014
Practical Biostatistics for Clinical Trials,or How to Find and Use Your Biostatistician
Scott Brown, Ph.D.
NAMSA
2
Overview Biostatistics Process Finding Your Biostatistician How Does the FDA View and Use Biostatistics? Statistical vs. Clinical Significance What Does FDA Expect?
3
Biostatistical thinking is an integral part of the design – not to mention analysis and presentation – of nearly every clinical trial
4
Biostatistics Process
Clinical Study Planning
Development of Protocol / SAP
“Operational” statistics: tables, listings, figures
Output: clinical study reports,
scientific communications
Interpret: internal /
external consult, regulatory,
reimbursement, customers
5
Finding Your Biostatistician
6
Optimal situation: biostatistician with clinical and regulatory experience including
previous FDA submissions, collaboration with non-statisticians including physicians, and working knowledge of disease state and therapeutic area
7
Reasonable approach: consultant biostatistician with similar levels of experience
8
My take: you want a biostatistician who is willing to take ownership of your
project, to treat it as his or her own
9
For larger studies or a series of studies, a team of biostatisticians may be employed, with tasks suited to skillsets and interests
10
How Does the FDA View and Use Biostatistics?
11
FDA’s Division of Biostatistics within the Center for Devices and Radiological
Health (CDRH) is a section of the Office of Surveillance and Biometrics
12
Statisticians tend to work across different therapeutic areas and may or may not have knowledge of the specific disease state or treatment under consideration
13
Biostatistical analysis is a tool to help interpret data, not a sole means of
reaching conclusions about approvability The FDA statistician will not be the one signing the approval
letter
14
This point cannot be overstated: FDA evaluates “the totality of the data”
That is, hitting your primary endpoint does NOT imply FDA approval of the application – for that matter, failing to hit your primary endpoint does not imply disapproval
15
Statistical vs. Clinical Significance
16
P-values tell us nothing about the size or relevance of an observed benefit – only whether the benefit can be explained by
chance For example, a large study can detect small, clinically irrelevant
differences as being statistically significant
17
Conversely, a large clinical benefit can fail to be statistically significant if the trial is
small or poorly designed Clinical significance depends on the magnitude of the effect and
the importance of the outcome
18
Clinical significance is not a statistical issue, but rather an matter of judgment
19
What Does FDA Expect?
20
Statutory directive: for PMA submissions, rely upon valid scientific evidence to
determine whether there is reasonable assurance that the device is safe and
effective
21
For a PMA (or 510(k)) submission, certain endpoints will be prespecified in the protocol and possibly the statistical
analysis plan (if one exists)
22
To view the complete series about Practical Biostatistics for Clinical Trials, or How to Find and Use Your Biostatistician Visit NAMSA Seminars
For information about NAMSA’s clinical services Check out our Clinical Research page
For additional information on this topic Download our brochure on Clinical Research
If you have any questions or would like to request a quote Contact us at [email protected].