Do you remember the old paper case report form (CRF)? I
remember working on an oncology trial. It was ‘massive’. It was massive because
the CRF was 300 pages long, in a binder you could barely lift with one hand, and
there were 300 subjects.
Let’s say, on average, there were 30 fields per page, which
would equate to about three million data points collected over a five year
clinical trial. We thought of this as a large trial, and yet from a data
perspective, it could easily be stored on my smartphone. I’m calling this period Clinical Data Management v1.
Then, there was a revolution called Electronic Data Capture
(EDC). This was a fantastic concept of gathering clinical data into the drug
development data life-cycle in real time. While there have been significant gains from EDC, still thirty years after
its inception, its key tenant, bedside data capture, has yet to be realized.
This was Data Management v2.
Then, the accountants got involved. They wanted more
productivity and reduced costs. At the
same time, the scientists and statisticians got smarter by inventing new
scanners, devices, biomarkers, and adaptive trial techniques.
To this end, there was an explosion of outsourcing to contract
research organizations (CROs), along with new data source providers and tools, such
as specialist labs, imaging, and electronic patient reported
outcomes (ePRO), to name just a few. What had started as an initiative to move
from paper CRF to electronic CRF, had suddenly turned into a much more complex
data management challenge.
This is today. It is a world of multiple data sources, from
multiple providers, which must be standardized and integrated quickly to
deliver statistical conclusions on efficacy and safety. This is Data
Management v2.5. In software terms, this is the .5 patch, a partial upgrade on our way to all things electronic (e).
Pausing for a moment on this journey, let’s take a quick look outside our office window
and see what other industries are doing. In just one minute
Uber starts 694 rides, YouTube users upload 300 hours of video, and Facebook
publishes four million
“Likes” from friendly folks. Of course,
we should also mention the CERN Large Hadron Collider producing over 30
petabytes of data a year. My ‘massive’ oncology trial, of approximately 100
megabytes, suddenly becomes a grain of sand in a beach of data.
Can you see Data
Management v3 coming? It’s in development. It is possible. It can happen.
It will be a world far removed from data managers who focus
on identifying data inconsistencies, such as ‘missing gender’. It will be a
world removed from clinical monitors flying in person to sites to verify the
true clinical record against the transcribed, EDC data; and it will be a world without
statisticians worrying about limited statistical power for their analyses.
Data Management v3
will embrace eSource, sourcing data directly from electronic health records
(EHR) systems, devices, and more. The concept of a clinical visit will change
significantly, become event based, and focus on real life data captured as a
The enablers of this vision are here today, and are evolving
rapidly. Devices, sensors, and wearables are raining down on consumers. They
will become the norm for clinical trials. Official regulator groups, such as
the FDA, already are requesting dialogues with the industry on how best to maximize
the sensor/mHealth revolution. Further, the
convergence of EHR systems around interoperability is being driven by
government/payer demands, which the clinical trials industry can leverage to
achieve true eSource.
Finally, technologies, such as big data analytics and
artificial intelligence, are advancing at an exponential rate. Initiatives in
other industries, such as improving sports performance, predicting flu
epidemics, and fully autonomous cars, are creating dynamic new tools and
methods which we can exploit to gain further insight from our clinical data.
Disruption is coming. Some facets of clinical trials driving
data management (transcription, data lag, and discrete, visit-based encounters)
are about to disappear.
Database Lock, a concept based on an arbitrary definition of
completeness and cleanliness of data, will be banished to the history
books. Clinical data management will
evolve quickly into the management of data specification, acquisition, and
The true winners in this new world will be those who learn
to exploit technology across the continuum of clinical data. The winners will
be those who can find, explore and
discover new hypotheses, patterns, trends, and conclusions to drive out new
therapies for better patient health.
Clinical trials just got exciting. Are you ready to embrace the