By Julie Horan
October 7, 2014 | | To trust trial outcomes, you must verify all data. This
is an old adage that continues to be used in drug development, where
traditional standards require that to ensure patient safety and data quality, clinical
trial sponsors need to verify 100 percent of the trial data with on-site source
data review.
Due to advances in digital data capture and data
modeling tools, sponsors, and the clinical research organizations (CROs) that
support and monitor clinical trials on their behalf, now have more ready access
to data than ever before. But, where electronic data capture has progressed,
data validation techniques remain outdated and inefficient. As a result,
sponsors still feel compelled to devote much of their limited resources to the source
verification of every data point. This issue has brought into question current
approaches to source data verification (SDV) and whether 100 percent SDV is
necessary.
To put this into perspective, a study on
data verification by the Swedish Association of the Pharmaceutical Industry found
that SDV alone can consume about 25 percent of a trial budget*. Meanwhile, other
studies support the contention that the effect on data quality of performing
SDV on all data is negligible. TransCelerate BioPharma Inc., an international
pharmaceutical industry consortium, analyzed nine studies from six member
companies, and found that on average only 7.8 percent of all queries were generated
by SDV. Their research concluded that 100 percent SDV made no meaningful
improvements to data quality.
Yet patient safety remains a top priority, and it
is not immediately clear how sponsors can craft new SDV review objectives to
ensure a data set ultimately meets the requirements of regulatory authorities,
while streamlining this part of the data review process. The key to
successfully transitioning from 100 percent SDV is to develop an integrated
data review plan that helps sponsors better understand which data points matter
most.
Identifying critical data requires taking a step
back and considering the trial’s larger purpose — allowing the end to define
the means. The endpoints, both primary and secondary, should guide all review
processes. Some endpoints are defined by one data point, while others
involve secondary and tertiary data. Mapping case report form fields to
the endpoints they support can help reveal which are the most important data
fields to clean. A clear plan can then emerge and more attention can be
focused on cleaning the information that is critical to the study objectives.
Define your teams and project timing
Clearly defining the roles within the data review
team creates efficiencies that balance quality data review objectives with
available resources. Traditionally, sponsors have contracted CROs to handle
routine aspects of trial conduct and deliver a data set for the sponsor to later
assess. However, a partnered collaboration between sponsor and CRO can prove
more effective in a risk-based monitoring setting, particularly for smaller or
mid-size biopharma companies that have less bandwidth to clinically analyze
data. With an integrated data review team, a clinical research associate (CRA)
travelling to a clinical trial site to review and source verify patient data can
work with an in-house clinical data monitor who reviews the data in advance of
the visit. This centralized review streamlines the on-site review and allows
the CRA to focus their efforts on the critical data points. Subsequent reviews
by a data manager are in turn shortened, as these processes help to remove many
data inaccuracies and inconsistencies early in the life of the trial.
In addition to assembling the right team, rethinking
the timing of data reviews can also bring efficiencies. Reviews should be planned
long before data review begins, ideally shortly after identifying the necessary
data fields and designing the case report form. An early and ongoing review
process can identify issues with protocol design or interpretation, and avert problems
before they could potentially derail a trial.
Spot trends faster with data visualization
Faster access to digital data has brought with it an
array of data modeling and visualization tools. Crucial data trends or changes,
which can go unnoticed when reviewing standard tables or listings, become
pronounced when presented as an image or graph. Newer software tools offer
sponsors the ability to create individual patient reports, which integrate data
from different sources. Programs like these assist in spotting data values outside
the expected norm, leading to more effective and targeted data review.
The future of data review and SDV in trial
management lies both in setting intelligent data review objectives, and using
tools to enhance execution. Taking these steps will ultimately lead to a
process that will decrease data handling and reduce errors. Creating more
efficient data review processes will require sponsors and CROs to work together
to create smart solutions that save resources without sacrificing data quality.
Dr. Julie Horan, DVM, is Senior Director of Clinical Science at Novella Clinical. As the head of Novellas’ Clinical Science
Department, Dr. Horan has more than 20 years of progressive experience in
clinical research across various indications and phases, with the past 10 years
exclusively in the management of oncology clinical trials. Dr. Horan has
strong experience in the scientific aspects of protocol development and
execution as well as data review and analysis. For more information visit www.novellaclinical.com.
* Funning S, Grahnén A, Eriksson K, Kettis-Linblad A (2009)
Quality assurance within the scope of Good Clinical Practice (GCP) — what is
the cost of GCP related activities? A survey within the Swedish Association of
the Pharmaceutical Industry (LIF)’s members. The Quality Assurance Journal.
12(1): 3–7. Available at http://onlinelibrary.wiley.com/doi/10.1002/qaj.433/pdf.