The Study Protocol Part Two
The Study Protocol Part Two
Contents
1. Contents
2. Course Objectives and Content
3. Course layout
4. Data Collection and Management
1. Data Collection and Management
5. Monitoring Study Safety, Quality and Data
1. Monitoring Study Safety, Quality and Data
6. Adverse Event Reporting
1. Adverse Event Reporting
7. Analysis Strategies and Statistical Considerations
8. Setting up a Research Team
9. Post-Recruitment Retention Strategies
10. Economic Evaluation
11. Clinical Laboratories
12. Results Reporting, Dissemination and Notification
13. References and Appendices
14. Protocol Amendments and Modifications
1. Protocol Amendments and Modifications
15. Common Problems
16. Considerations When Protocol Writing
17. Key Points to Remember
18. References and Resources
19. Quiz
Contents
What a protocol is
The topics that are covered by a protocol
Common problems encountered when writing a protocol
Making amendments to a protocol
This section provides the references used in this course and resources that you may find useful
for further reading.
- Quiz
This section provides questions that will allow you to test what you have learned from the
course.
Course layout
This e-course is divided into parts one and two.
1. General Information
2. Background and rationale
3. Aims and objectives
4. An overview of the research question and the study design
5. Methodology, interventions and allocations
6. Outcome measures
7. The sample size
8. Study population
9. Ethical aspects
10. Participant and community engagement
11. Research governance
12. Study management
13. Key points to remember
14. References and resources
15. Quiz
As recording the study data is considered to be the most crucial stage of the data management
process, high data quality is essential. To improve the reliability and validity of the source data,
it is important to get the processes and procedures for data collection correct from the start, as
later on there is little you can do to improve the data quality. Defining the source data is
important, although this might be captured in detail in the data management plan rather that
the protocol.
A data capture form (DCF) turns the protocol into the data capture system. The 2002 ICH GCP
guidelines define a DCF as a printed, optical, or electronic document that ‘is designed to record
all of the protocol required information to be reported to the sponsor on each trial subject’. A
DCF can also be used to document where the source information can be found e.g. medical
notes, scans and X rays. The term 'case record form' or CRF is also commonly used for the data
capture tool.
As McFadden (2007) points out ‘if using paper data capture forms, there is a need for systems
for (a) distributing blank paper forms to the participating sties and (b) for returning completed
forms in a timely way to the Coordinating Centre. If data is captured electronically, hardware
and software must be developed and fully tested and validated. If samples are being collected
as part of the study (e.g. X rays, blood/tissue samples), mechanisms should be in place for
shipping, receipt, and logging of sample’.
Some of the issues that you may want to consider when deciding upon your data management
systems include:
It is important to remember that no matter how well a study is run, if the study does not capture
the correct data, meaningful analysis may not be possible. It is therefore important to have
quality control systems in place for data checking, audit trails and source verifications
procedures.
This section is often split into quality management and safety monitoring.
In terms of safety of the participants the Principal Investigator is responsible for monitoring,
assessing, recording and reporting on all adverse events related to the study to the sponsor and
data safety monitoring board (DSMB). Procedures for continually reviewing the risk-benefit
status should also be provided.
Interventional studies may wish to evaluate efficacy or safety issues around the intervention(s)
during the study’s progress. This evaluation is called an interim analysis. If interim analyses are
planned the timings and processes should be explained. If applicable, formal statistical stopping
guidelines can be applied to the interim analyses in case any evidence of unacceptable risk or
actual harm is uncovered by the analyses. Any stopping rules should be laid out and justified.
The study set up must have clear, comprehensive internal and external
quality systems which ensure that the quality of the procedures, and the
results generated, are reliable and valid. The description of the data
handling procedures should include plans for data checking (how much will
be checked, by whom, and how); audit trails (detailed log showing which
data have been changed, why it was changed, who changed it and when);
and source verification procedures (how much data will be checked
against the source documents e.g. medical records).
The difficulties anticipated throughout the duration of the study should be listed. The potential
solutions to these anticipated problems should also be outlined.
All clinical research should have an appropriate level of safety monitoring and quality
management to protect the rights of the individual and to ensure that the data and procedural
qualities are reliable. It is important to also remember the resource and cost implications when
committing to levels of monitoring and quality assurance in the protocol. They should be
proportionate to the risks of the study.
Mulay (2001) stresses that 'reporting SAEs is your highest priority after participant care'.
The WHO’s 2005 Draft Guidelines for Adverse Event Reporting and Learning Systems advises
that non-serious events are recorded in the study's data collection forms alongside the other
data. SAEs are usually collected on a specially designed form. If the study plans to use these
forms, the data they will collect should be outlined, described, and justified. Actual and potential
risks of participating must also be described in detail. Efforts to minimize these risks must be
clearly explained.
SAEs should usually be reported to the coordinator (or principle investigator if there is no
coordinator) within 24 hours. The coordinator or principle investigator should inform the Data
Safety Monitoring Board (DSMB), Ethics Committee and sponsor, typically within a week.
Generally studies will only report serious, unexpected and possibly related events, to the DSMB,
ethics committee and sponsor.
It may be that a DSMB is required to independently review adverse events. These are by no
means always needed and often a local independent safety monitor is a very acceptable step
for monitoring safety. Trials evaluating investigational new products often have a local safety
monitor and a DSMB. Whether a DSMB is needed is dependent on the type of intervention, the
sponsor's requirements and local regulations.
Details of the proposed analysis of the primary and the secondary objective(s)
should also be included. If it is not possible to have a statistician write the
statistical analysis plan, it should at least be commented on by one.
The statistical analysis plan outline can be altered during the course of the study, for example
after requests for interim analyses, but it should be finalised before the database is ready for
the full data analysis as this avoids the risk of the data being manipulated. A statistical analysis
plan has the advantage of preventing unplanned analyses at the end of the study but it should
not stop further examination if unusual features of the data are identified during the analyses.
Toto and McPhaul (2011) state ‘the investigator must invest in building a research team that
consists of individuals (e.g., study coordinator, co-investigators, biostatistician, research and
laboratory technicians, and administrative support) that have the level of expertise required to
navigate through the various study challenges'. Of course this depends upon the size and
complexity of your study. The requirements of a year long complex multi-centre trial of an
experimental new malaria vaccine will be vastly different from a small phase II 'proof of
concept' trial with 40 participants evaluating whether to treat pneumonia at home or on the
ward.
Having the right staff is essential to the success of your study. The following should be taken
into consideration when deciding on the make-up of your research team:
Once you have funding secured and have an outline of the research team you will need, you can
begin to hire staff members in order of requirement e.g. laboratory staff won’t be needed until
you have samples to work on but the study co-ordinator will need to be in place as soon as
possible to begin tasks like developing the protocol, seeking regulatory approvals and
establishing the various groups.
Continued data collection is only possible with the on-going interest and interaction of both the
study’s recruits and the study centres, this therefore must be planned for before the study
begins.
Options such as planning collaborator meetings, producing newsletters and posters, and sending
out merchandise should be listed if you intend on using them during the course of the study.
Economic Evaluation
An economic evaluation is the comparison of the costs and consequences of the
alternative courses of action with regards to issues such as a treatment, a health
care provision and drug options. Some studies may want to consider including a
health economic evaluation in their protocol. These can be highly beneficial if a
trial is evaluating a new intervention in a resource-limited environment.
The point is to decide whether or not one intervention is more cost effective than another by
weighing up both efficacy and costs. The most inexpensive treatment is not always the
recommended one. Hackshaw (2009) suggests that the ‘treatment of choice is likely to be both
cheaper and more clinically effective’.
Clinical Laboratories
Many clinical studies involve the collection and analysis of samples. The
protocol therefore needs to state how these will be collected and by whom.
Sample collection time can be critical, either for diagnostics or often for
pharmacokinetics. It is often helpful to ensure sample collection is included in
the study flow diagram.
A laboratory management plan is typically written as a separate document but key points should
be written in the protocol.
Authorship should also be discussed in the protocol. Large collaborative studies have group
authorship with a list of contributors at the end of the paper e.g. the participants, the Steering
committee, the DSMB and the collaborating researchers.
As the WHO's Research Ethics website points out that ‘the protocol
should specify not only dissemination of results in the scientific media,
but also to the community and/ or the participants, and consider
dissemination to the policy makers where relevant. Publication policy
should be clearly discussed - for example who will take the lead in
publication and who will be acknowledged in publications, etc’.
Disseminating results to the public is essential. If you do not disseminate the results, the time,
resources, and commitment which have been invested into the study will have been in vain as
no one will get to learn from your research.
Some funders may require that any resulting publications are Open Access. You should be
aware of your funders policy with regard to this and outline an appropriate plan in the protocol if
this is something they require.
A sample of each of the study DCFs that will be used can be included
Budget details
Study timeline
Dummy tables
Statistical analysis plan
A record of the amendments with version numbers and dates on all relevant documents must
be kept. For example, if it becomes obvious that altering the method of statistical analysis
would be beneficial, or that adding an extra statistical test would produce a more credible result,
the section describing this in the protocol will need to be amended. Once approval has been
sought and granted, the protocol should be amended. The updated protocol should be marked
‘Version 2’ and should be dated. If a further change is ever made to the document it will then
become version 3, with the appropriate date included.
A well written protocol, developed through careful consultation, will help reduce the need for
amendments. However, sometimes amendments are important and fully warranted to improve
a study. For example, sample size calculations are often set using assumptions and until the
study is underway these cannot be fully predictive. In practice, the nature of a disease or the
numbers presenting may be different and so a change in the number of participants is needed.
Common Problems
There are a series of problems that commonly occur with protocols which lead to the weakening
or failure of a study. When preparing your protocol you should be aware of the
following:
The protocol being too ambitious e.g. trying to answer too many questions, it is best
to have one clear primary objective.
Lack of background knowledge and previous research in your topic of interest,
combined with lack of discussion with all collaborators, partners, the community and
the research team.
An inadequate description of the proposed study and a clear description of the aim.
Poor justification for the study. You must be able to clearly justify the need for the
research, show why is it important to answer this question and explain the impact it will
have on public health.
Inappropriate statistical analysis e.g. using tests that will not produce the required results
Lack of piloting or testing any of the study methods and instruments.
1. Planning is essential, many months may be required to prepare and finalise a protocol.
2. It is important to be aware of the funding policy and procedures of your funder and to
ensure your budget is as realistic as possible.
3. Background knowledge in the proposed area of research is key so it is important to know
what research has already been carried out on the study topic.
4. It is advisable to have the tasks and responsibilities off all partners clearly outlined to avoid
confusion later on.
5. The protocol is not finalised until all peer-review suggestions are obtained, and where
applicable, incorporated.
Not every study protocol will need to cover all of the topics covered in parts one and two of this
course. The study design will dictate which will be relevant and which will not e.g. a complex
clinical trial of a new investigational drug should have all of the sections discussed within this
course included in the protocol but a small cross sectional survey will not need to include
sections on, for example, ‘Economic Evaluation’ or ‘Interventions and allocations’.
Websites:
Resources
Quiz
Summary
1. Once a protocol is finalised it should not be altered under any circumstance:
True
False
2. Which of the following statements is false:
You should avoid trying to answer too many questions
It is important to be aware previous research in your topic of interest
An inadequate description of the proposed study and its aim lends to a poor
protocol
Poor justification for the study will result in failure to obtain funding and regulatory
approvals
Piloting or testing study methods and instruments is considered to be a waste of
resources.
3. When writing a protocol which of the following do you not need to consider:
The length of time required to prepare and finalise a protocol
The types of contracts required for the research team
Providing a sound background to the study
Including the tasks and responsibilities of all partners
4. Generating dummy tables can help you to decide on how many participants you need to
recruit:
True
False
5. A strong research team is one that:
Has a lot of time to spend on the study
Consists of experienced individuals with different areas of relevant expertise
Is not expensive to employ
6. To avoid wasting resources such as participant time and study funds it is important to
collect only data required to answer the study objectives:
True
False
7. Long term studies or those that plan to have participant follow-up should not use which of
the following to encourage continued participation:
Newsletters
Meetings
Monetary bribes
Merchandise
8. Which of the following statements is false:
High data quality is essential because the poorer the quality of your data the less
reliable the study results will be.
It is important to get the processes and procedures for data collection correct from
the start, as later on there is little you can do to improve the data quality.
Well-designed data capture forms (DCFs) mean that you will still have valid results
even if you are missing a lot of data
9. Funding bodies and health care decision-makers are interested in economic evaluations
because:
They can decide on where they would like the study to be run based on how much
it would cost
They require information on both the effectiveness of the particular intervention,
and on its effectiveness relative to other interventions that already exist.
10. Which of the following is not a component of a well-designed DCF:
Clear and easy to read
Understandable instructions making it easy to complete
Collects additional information that might be useful for extra analysis
11. With clinical laboratory samples, labelling, tracking and management is very important
because:
Often the laboratory samples measure a critical endpoint in the study
Taking and processing samples is expensive so you don't want to lose any
12. When deciding upon your data management systems, which of the following does not have
to be considered:
Time needed to complete the DCF
Collection and entry
The database to be used
Processing and coding
Validation and cleaning
Analysis
13. Which of the following statements is false:
Failure to disseminate the study results means the time, resources and
commitment invested will have been in vain as no one will get to learn from your
research.
You should only publish your results if you achieve a positive result.
14. Clear, comprehensive internal and external quality systems should be established to:
Demonstrate the strength of the study design
See that as many participants as possible are recruited as quickly as possible
Enable the investigator to carry out the study in as short a timeframe as possible
Ensure that the quality of the study procedures, and the results generated, are
reliable and valid
15. Stopping rules can be put in place in case of:
Evidence of unacceptable risk or actual harm uncovered by interim analysis
The study running out of money
Participants withdrawing their consent to continue in the study
New regulatory approvals being required
16. Which of the following statements is true:
All adverse events must be reported to the ethics committee, even if they are only
minor
It will depend on the individual study which types of adverse events are reported,
by whom and within which timeframe
Only serious adverse events that result in death need to be reported to the ethics
committee
17. The statistical analysis plan sets out how you will analysis your data:
True
False
18. Which of the following should not be included in the statistical analysis plan:
The assessment of treatment compliance, efficacy and safety
The quality assurance processes
The reasons for the sample size selected, the power of the study, the level of and
significance to be used
The procedures for accounting for any missing or spurious data etc
The statistical tests that will be used to test the hypotheses and how the data will
be analysed
Submit