Transcript DSMBs are needed

Data Safety Monitoring Boards /
Data Monitoring Committees
in paediatric studies
Paolo Tomasi, M.D. Ph.D.
Head of Paediatric Medicines
European Medicines Agency
Presented by: Paolo Tomasi
An agency of the European Union
Current EMA/PDCO position for PIPs
Opinion template (published on website) contains
limited guidance:
(independent) DSMB:
Should be foreseen in all studies in neonates, in
safety/efficacy studies in other age groups.
If inserted in PIP opinion, this becomes a key element =
obligation for the applicant
(if not created: non-compliance  non-validation)
Regulatory Requirements and
Guidelines for DSMBs
All sorts of trials: academic, product development,
industry, governments…
All sorts of regulators: FDA, EMA, MHRA, MRC…
All sorts of groups suggesting guidelines:
TDR/WHO, European Commission, universities,
funding agencies, NHS…
Why are DSMBs needed?
WHO: Operational Guidelines for Establishment and Function of Data
Safety and Monitoring Boards.
DSMBs are needed:
• because trials should not continue
if the design is no longer
appropriate
• as a trial might reach its objective
earlier than predicted, or the
primary objective may never be
achievable
• a positive trend to do harm within
trials might become apparent
• trials may need modifying if
accumulated data is not in-line with
the original trial design
assumptions
DSMBs might also need to
recommend protocol
amendments such as:
- Change in dosage of trial or
concomitant medications
- Treatment duration of trial of
concomitant medications
- Entry or exclusion criteria
- Sample size
- Recruitment rate
In red: amendments potentially
requiring a modification of the
agreed PIP
When are DSMBs needed?
WHO: Operational Guidelines for Establishment and Function of Data
Safety and Monitoring Boards.
All trials need safety monitoring and many BUT NOT ALL need
DSMBs. Definitely needed if:
- Controlled CT with mortality or severe morbidity as primary or
secondary endpoint (life-threatening condition).
- RCT evaluating clinical efficacy and safety of an investigational
new product
- Early studies of a high risk new intervention
- CT design or data accrual is complex
- If there is uncertainty about whether the data will impact the trial
design
- Vulnerable populations (Paediatrics, Neonates…)
- Entry or exclusion criteria
- Sample size
- Recruitment rate
When are DSMBs needed?
NIH
• Need to submit a data safety and monitoring plan
for all phase I – III trials.
• This may or may not include a DSMB. Usually
require a DSMB if:
multi-centered
blinded
high risk
vulnerable population
• Need to provide justification if no DSMB
proposed/required.
When are DSMBs needed?
EMA/CHMP DMC guideline
• Usually required if:
Life-threatening disease (any case)
Potentially harmful treatment (even if highly
effective)
Preplanned interim analysis/complex designs
vulnerable population such as children
• Responsibility for trial lies with
sponsor/investigators, not DSMB.
When are DSMBs NOT needed?
Criteria for NOT having a DSMB:
EMA/CHMP guideline:
• Not practical if CT is too
short (may apply even if
life-threatening)
• Known drugs used within
the license
• Non-critical conditions
(mild pain trials?)
• DSMB may be even
counterproductive in
some situations
UK NHS: Issue in Data
Management and Interim
Analysis of Trials:
• Not possible for a DMC to
make a contribution
• Where any observed
differences would not prompt
a protocol change (i.e.. stop
the trial)
• Where there is no reason why
a DMC decision would differ
from internal monitoring
(open-label studies?)
Questions for the participants
1. Are there any recent/additional
views of the CHMP on the
requirements/indications for a
DSMB?
2. Is the EMA/PDCO guidance
sufficient?
3. Should there always be an explicit
justification if a DSMB is not
proposed/mentioned in the
opinion for any paediatric study?