|
|
|
International Clinical Trials
|
As the processing and managing of clinical data continues to evolve with
the advent of new technologies, the approach taken to risk-based
monitoring is being challenged and modernised.
More than 50 years ago, the clinical trial industry was shaken to
its very core by the thalidomide research disaster. The reactions were
widespread and included a renewed focus on data quality, patient source
notes, investigator management and, in particular, a demand for
extensive processes to obviate the collection and reporting of
fraudulent data. The follow-up to the disaster also marked the birth of
modern-day monitoring as we now know it. Safety in checking led to
confidence, confidence led to best practice, and best practice led to an
all-consuming acceptance that monitors must go to sites every four to
six weeks, check every data point against the source notes, and
‘guarantee’ that data is valid and correct at all times.
During the latter stages of the 20th century, with profits at a
record high, financial constraints took second place to speed and
inherited best practice. People accepted historical processes as
standard practice and did not seek more timely solutions. Clinical
operations and data management worked two different jobs; handling the
same pieces of paper, but rarely combining their efforts beyond a
transitory collaboration around database lock time. There is, of course,
a touch of sarcasm in this position, but having spent 15 years working
in data management, I feel confident that many of you will agree with
me. Today, we find this paradigm under extreme threat from many
directions: increasing investigator fees; tightening regulatory
requirements; the impact of new technology; and the growing pressure on
R&D time and cost. We should also not forget the patient, who today
demands lower costs for treatments and strives for unprecedented levels
of personalised medicine. The net impact is that we need to do much
more, with much less, much faster.
In August 2011, the US Food and Drug Administration (FDA) released
the ‘Guidance for Industry Oversight of Clinical Investigations – A
Risk-Based Approach to Monitoring’ (1). This draft guidance addresses
the misconception that 100 per cent source document verification (SDV)
is required by the FDA, encourages sponsors to tailor monitoring plans
to the needs of the trial, and even describes strategies for monitoring
activities that reflect a modern, risk-based approach. The guidance
encourages sponsors to focus on critical study parameters and rely on
well-defined, yet flexible processes to oversee a study effectively. The
European Medicines Agency (EMA) released a similar paper, also in
August 2011.
Although risk-based site monitoring is a hot topic, the
counter-considerations are equally strong – we operate in a risk-adverse
industry, leading to the stifling of innovation by restrictive business
practices, preconceived ideas and incorrect perceptions. The resulting
failure to evolve processes and resistance to new approaches or
technologies is the subject of this paper. This article examines some
major industry concerns in adopting a riskbased approach and showcases
how technology paves the way for clinical researchers to take it on.
The Challenge: How Do We Ensure Trial Integrity and Optimise Trial Costs?
Traditionally, four of the most significant considerations when initiating a trial are:
- Trial design (and statistical planning)
- Site selection and patient recruitment
- Monitoring
- Data collection and management
These considerations can all be intrinsically linked by
consolidating our focus on the lowest common denominator: data. If we
accept that the main objective of clinical trial is to collect the data
that we require to prove our hypothesis and confirm the safe and
efficacious long-term usage of the solution under examination (chemical
entity or device), we can then review our clinical trials’ best
practices from a more holistic perspective.
Indeed, by examining these four trial components, we can trace the
data through these phases, reviewing how, when and why we interact with
the data generated. We can then truly identify how a streamlined
strategy can not only simplify trial execution, but actually enhance
quality by focusing on what truly matters, and consequently deliver upon
a critical theme in clinical trial execution – putting the right people
in the right place, at the right time. It is through this analysis that
we will address the true challenges to ensuring trial integrity and
optimising trial costs.
Trial Design (and Statistical Planning)
The key question every study team should be asking is ‘are we
collecting the data we need and only the data we need?’ More often than
not, the truth is that we are collecting too much data – data that we
will not even analyse in a final study report.
A recent analysis completed by the Tufts Center for the Study of
Drug Development revealed that 15 to 30 per cent of all data collected
by some sponsors are never used or incorporated into a new drug approval
submission (2). If we condense the volume of data collected to what is
truly important, there must be fewer data points to verify, and
therefore we reduce the volume of time, resources and money required to
manage that trial.
However, reviewing the trial design also permits us a chance to
prioritise that data – most typically key efficacy and safety data, or
data linked directly to a primary or secondary objective. This analysis
also allows us to identify the less important data – the data that we
can make a conscious decision to expend less energy upon. This is the
first step in our risk-based, or targeted, data strategy.
Site Selection and Patient Recruitment
As a clinical development programme progresses, we can learn from
one study to the next, gaining insight into common mistakes, challenges
and high value data points. Sites participating in early phase studies
might be expected to carry forward their lessons learnt and experiences
in a positive way. That means we can consider categorising sites
according to their relative experience with the clinical development
programme under discussion.
Does this knowledge offer us a further decision on our risk-based
data strategy? Put simply, yes. If we select the most appropriate sites
and train them appropriately, we can expect significantly higher data
quality in return. Using technology to gain feedback on data issues in
‘real-time’ early in the process is key; and of course, sharing feedback
across sites, so that mistakes are not replicated across the entire
site base, is fundamentally important.
Monitoring, Data Collection and Management
The wide adoption of data management technologies such as electronic
data capture (EDC) in the last decade certainly made its mark in the
area of processing and managing data. It is the advanced use of these
technologies, in conjunction with trial design and site selection, which
can enable further step changes in the classic site monitoring
strategy. Perhaps the simplest example of this is site-based data entry;
as sites enter data, they are immediately subject to validation checks,
promoting immediate feedback to the site, which in turn promotes the
real fixing of basic data errors. Anyone viewing the data can assess the
data quality at a glance without having to review the content manually.
This is just one simple example of technology performing in seconds – a
task that traditionally could take minutes or even hours. Can we
challenge other activities in the same manner?
Changing Paradigm
It is important to challenge the classic functional labels of
monitoring and data management, and instead consider the tasks that need
to be performed. In my experience, I have always looked at
monitors/clinical research associates (CRAs) as requiring two very
different skill sets:
- Site motivators: skilled at site management, motivation and patient recruitment
- Data monitors: skilled at reviewing data and finding the errors
Monitors with a preference towards site motivation tend to deliver
lots of patients and lots of data, but might lack the attention to
detail when it comes to reviewing that data. By comparison, monitors
with a preference towards the data generally delivered fewer patients
(and data), but what they delivered was immaculate. This clearly
describes the clinical challenges – how to find patients to meet
recruitment targets or deliver perfect data. In an ideal world, we would
do both.
If we also overlay the advent of technology to this role map – in
particular EDC – we can identify a further fundamental change in our
strategy. Utilising a paper case report form (CRF), the first person to
see the data was the monitor, who would then make provision for that
data to be shipped to data management after performing the required SDV.
Over time, this non-direct data transmission, was evolved to enable
sites to pull pages and submit them direct to data management (direct
data transmission).
Under direct CRF or data transmission, the data manager would review
the data, generate queries and provide direct site feedback, often
before the next monitoring visit would occur. Today, using EDC, it is
almost always the data manager who reviews the data first, enabling us
to take this process even further, challenging the traditional division
of responsibilities and allowing us to further streamline data checking.
Are we merging the roles of data management and clinical monitoring,
or should we now simply decide that data management is responsible for
the quality of
the data? I certainly do not think the former is true, but more a case
of maximising strategy. The site management component is absolutely key
and cannot be underestimated. Perhaps a more valid approach is to
re-examine the real steps involved and align our data strategy around
these. In essence, we combine the classic monitoring plan with the
classic data management and validation plan, and allocate full
responsibility for each activity to a single individual.
New Approach
There are two major considerations when we build and execute our
monitoring plan: volume of SDV and frequency of visits. By controlling
and understanding the volume of SDV, we can allow the accrual of data to
drive events at the site. Instead of defaulting to setting monitoring
visits to every four to six weeks, the data can dictate that frequency.
High enrolling sites generating greater volumes of data would therefore
likely need more frequent or longer visits. With these considerations in
mind, we can employ a very different mentality to scheduling and
managing standard monitoring visits.
Site Management Activities
Each time a monitor visits a site, there are certain activities that
must be performed. If we review these activities, we can anticipate they
will take a fixed time (for example, four hours) for the monitor to
complete – assuming that there are no major deviations from normal
practice. Therefore, our goal is to optimise our return on investment
(ROI) for this visit by ensuring there is sufficient work required in
other areas to allow the CRA to spend a full day (or days) on site.
Data Handling and Review (for Example, SDV and Query Management)
If we can track the volume of data that requires SDV in real-time, we
can project when there will be sufficient data to maximise the monitor’s
time on site. By examining a patient’s visit schedule, with an
understanding of the data flow and our SDV requirements, we can project
when this optimal time will be. Furthermore, the monitor can work with
data management to ensure that all data-related activities are
concurrent in time for that visit, again maximising the ROI for that
visit. In terms of a monitoring plan, there are other considerations
that we must consider – such as serious adverse events, deaths, protocol
violations, site personnel changes – all of which may cause us to
require a visit before the data projects, but using the data as an
overall projector brings significant benefits, saving time and effort.
With this type of planning, it is now common for sponsors to estimate
that monitors will be on site (post-initiation and pre-lock), on
average, every eight to 12 weeks (as opposed to four to six weeks). The
reduction in direct (labour) and indirect (travel) costs are
significant.
Does This Safeguard Our Data?
As an industry, our mission remains unchanged – to deliver safe and
effective treatments to patients and those in need of clinical
assistance. It is clearly stated and understood that sponsors of
clinical investigations are required to provide oversight to ensure
adequate protection of the rights, welfare and safety of human subjects
and the quality and integrity of the resulting data submitted to
regulatory authorities. With the advent of technology and an updated
regulatory landscape, our standard approach is being challenged and
modernised.
It is important that we recognise the difference between quality
assurance and quality check. Risk-based or targeted site monitoring
enables a systematic, total data quality assurance approach to be
implemented, which considers the overall responsibilities shared by
clinical project management, clinical monitoring, data management and
other reviewers, and then re-deploys them in a manner that obviates
duplications and streamlines clinical data acquisition and management.
The ability to identify data trends early and to share these learnings
enables us to re-deploy study personnel, placing the right person in the
right place at the right time. The industry must stop performing site
visits with no data, and stop burdening high performing sites with
unnecessary visits on a fixed schedule, as they serve only to distract
and delay the process.
By using the data to drive those decisions, we promote our chances
of catching errors early and learning from mistakes first time around.
Technology plays a fundamental role in every current clinical trial and
brings with it the ability for real-time visibility. If we can view
real-time data, and most importantly utilise that data in a standardised
format, we can direct field operations in a logical and appropriate
manner.
Sites are happy because they will not be visited as often, or
perhaps their average visit duration will be reduced. Monitors will be
happy because they can perform more centralised management of sites,
spend less time on travelling and more time on valued added activities.
They can also prioritise appropriately, visiting sites that need
assistance in meeting enrolment targets or data quality standards, not
visiting low recruiting sites ‘just because my last visit was six weeks
ago’. Data management assumes more of a daily role in the site
management activities, facilitating the site, the monitor and the rest
of the clinical study team.
Conclusion
None of this is a replacement for thinking. The data does not always
tell the whole story. In addition to data-driven factors, we must also
look for other signals – serious adverse events, deaths, protocol
violations, site feedback – events that can require more immediate
action. The data helps us identify global trends, as well as localised
issues, but the data cannot be used in isolation to drive every critical
decision. You can, however, use the data to challenge decisions and
verify the need, or lack thereof, for activity.
The key to all of this is technology. Real-time, complete and
coherent data flow shows us where we are being successful and where we
need to invest our time and effort. The advent of technology has
promoted these ideas from potential to standard procedure, turning
risk-based approach into reality. The critical piece to this puzzle was,
and remains, a flexible platform that centralised clinical researchers
can use to dynamically adjust the SDV level in real-time. Based on the
observed quality, these researchers can effectively manage sites and
their field-based operations.
References
- 1Guidance for industry oversight of clinical investigations – A
risk-based approach to monitoring, US Department of Health and Human
Services, Food and Drug Administration, Center for Drug Evaluation and
Research, Center for Biological Evaluation Research, Center for Devices
and Radiological Health, August 2011
- Getz K, Assessing the downstream impact of protocol design complex
city, Tufts Center for the Study of Drug Development, August 2009
|
Read full article from PDF >>
|
|
 |
|
| Rate this article |
You must be a member of the site to make a vote. |
|
Average rating: |
0 |
| | | | | |
|
|
 |
News and Press Releases |
 |
Pistoia Alliance Launches Global Community of Interest to Assess the Environmental Impact of Clinical Trials
More info >> |
|
 |
White Papers |
 |
|
Keeping Abreast of the Competitive Environment in Clinical Development
Thomson Reuters
Strategic management of a therapy area requires hard decisions to be made about which programs to invest in and which to let go. Understanding the rapidly changing competitive environment for your drug candidates is a key input into these decisions. It is easy to lose track of the broader picture when a clinical program is in progress and get blindsided by external events.
More info >> |
|
|
