grant application, protocol , and other documents Flashcards
what is murphys law?
“what can go wrong, will got wrong”
whta is a trial protocol?
A document that describes the objective(s), design,
methodology, statistical considerations, and organization
of a trial.
Can also give background and rationale for the trial
what makes a good protocol
- Clear rationale of the trial
- Include all relevant preclinical and clinical data in the
protocol - Details about the trial process: recruitment, follow-up,
and collection of outcomes - Able to replicate
- Detailed who will be doing what
- Safety issues and monitoring
- Data collection and main analysis
standard formats for protocols
> SPIRIT guidance
SEPTRE
which documents come out of the trial protocol
- randomisation
- data managment
- unblinding procedure
- TMS charter
- DMC charter
- Statistical analysis plan
what is a master protocol?
a different type of protocol containing within it several sub-studies
- can include more than 1 disease
- can include more than 1 treatment intervention
- can canswer more questions efficiently in less time
what is a Platform trial: Overarching master
protocol with Intervention Specific
Appendices (ISA)
more wishy washy
allows new new interventions to be added and old ones either futile or effective to be replaced. without having to start up a new trial
isa’s will have intervention specific appendancies
Statistical analysis plan
seperate document to be completed after finishing the protocol
what things are included in a statistical analysis plan?
- Details of the planned statistical analysis
- elaborating non the principal features of the technical analysis
- trial objectives
- data sources
- populations studied
- study endpoints
- statistical methodology
- sensitivity analyses and handling of missing data
why dod we need a SAP (Statistical a nalysis plan?)
unplanned or post hoc anaalyses can lead to data fishing or selective reporting
important to pre-define the analysis population, statistical methods, and methods for handling spurious or missing data
when do we write a SAP?
the plan should be reveiwed and possibly updated as a result of the blind review of the data and should be formalised before breaking the blind (knowing treatment allocation of subjects)
a formal record should be kept of when the SAP was finalised and the blind broken
things to specify in the SAP: Primary outcome
o Main outcome of interest in a trial
o Outcome on which any decision should be based
o Sample size calculations should be based on
o Usually specify single primary outcome
o Some cases may be justification for co-primary outcomes,
consider:
- Issues surrounding multiplicity
- Whether statistical significance required for all or at least
one in order to establish success of treatment
o Analysis should be pre-specified both in the protocol and in more
detail in the statistical analysis plan
things to specify in the SAP: secondary analysis
o Additional analysis of the primary outcome, e.g. adjusting for
additional covariates
o Analysis of additional outcomes of interest, e.g. safety outcomes/
potential harms of the intervention, additional aspects of the
condition that is being targeted
things to specify in the SAP: sensitivity analysis
o Robustness of results to different assumptions, e.g.:
Assumptions about missing data mechanism
Excluding extreme/ potentially spurious data values
Excluding ineligible participants /randomised in error
Impact of protocol deviations
Methods accounting for clustered /correlated data
Different censoring types in time-to-event data
Adjusted vs unadjusted effects
Different specifications of the outcome, e.g. different cutoffs
o Some can be pre-specified but some potential issues will only be
identified during review process
things to specify in the SAP: subgroup analysis
- does treatment effect differ in a certain subpopulation?
- should be prespecified if possible
- Subgroups of interest may only become apparent after review,
in which case it should be conducted as an exploratory analysis
