What data do we have?
Clinical trials get more and more complex with the demands on data collection, data privacy and trackability of any decisions and steps made. When we think about the data we will have at the end of the trial, by no means can we only speak about the patient data collected during visits and entered into case report forms (CRFs). In a simplified way, it could be categorized as in Figure 1.
With a trial requiring 6 to 11 years to move from Phase I to III, and large investments still facing a low success rate (30% for Phase I, 14% for Phase II and 9% for Phase III)(1), very few companies have the luxury of not knowing the status of their clinical study on an ongoing basis and allowing avoidable delays and risks. Let’s look at the different sources of data and how having real time access can help reducing costs, resources and time needed to finalize results.
Data entered into (e)CRF
While 10-15 years ago the term “EDC” was still relatively new (based on a web-survey done on Canadian studies only up to 20% of trials prior to 2007 used Electronic Data Capture (EDC))(2), tools used were complex and the main function was to collect, manage and report; the expectations we have now are far more advanced. At a minimum, an EDC tool should be able to:
- support fast design and updates of CRF pages,
- allow real-time access to designated study team,
- allow user-friendly data entry,
- support automated data checking,
- support fast creation on manual data listings,
- support review by different parties (e.g. DM, safety, CRAs) and tracking of the status,
- allow fast creation of printouts and data export,
- support medical encoding,
- allow import from external sources (e.g. central laboratories, central ECG provider, electronic Patient Reported Outcomes (ePRO),
- serve as Interactive Web Response System(IWRS), medication inventory, unblinding tool (as applicable),
- support eSignatures,
- be compliant with the industry regulations.
This might sound like a tall order, but with sponsors requiring clean and structured data from the moment the first patient is entered into the system, the expectations are high for all. With this said – is there any more room for paper studies? Conducting your study on paper will add multiple additional layers of complexity and the overall flow of data must be carefully planned. Every step needed from patient visit to having clean data in an electronic system means time, resources and costs. Also, tasks which are avoidable when an EDC tool is used must be instituted for paper studies (e.g. tracking of pCRFs and DCFs, entry or study data from pCRF to a collector, QC activities). It is very complex to analyse the study data or make any conclusions before data is in any type of electronic format (especially in the case of a multi-site study), and it is also easy to miss critical items, e.g.:
- sites’ inconsistency in understanding CRF completion guidelines,
- not having a clear picture of missing data and/or dirty datapoints,
- trial-wide early conclusions on critical datapoints,
- re-training or audit needs for sites.
The risks listed above can be minimized with an EDC tool. A well-designed eCRF will guide the site user through the process of completing the required data fields, point out any obvious data errors or possible deviations via automated validations, provides manual listings which can be run on the data on an ongoing basis allowing the identification of any protocol deviations and other critical findings as data is flowing in. Capitalizing on EDC capabilities, it is important to establish the rules on ongoing data entry and review to ensure data is made available and cleaned in EDC shortly after being made available. This requirement should not be underestimated, as this is the only way to ensure issues are identified while there is still time to act. It requires a lot less effort to re-train a site or modify eCRF (even protocol, if applicable) in the beginning of a trial, if any change needs become apparent. Detecting issues only at the end will mean greater time investment from many more study team members to fix the issue. And a most extreme case scenario could jeopardize critical analysis capability impacting the outcome of a trial.
Faster access to data also has a positive effect on other areas in clinical research. In the latest ICH GCP addendum the monitoring expectations have also changed. Previously the standard approach was to have 100% SDV done during site visits; however, the latest version is suggesting development of a systematic, prioritized, risk-based approach, also allowing a combination of on site and centralized, or, where justified, centralized monitoring.(3) This would only be possible with ongoing data entry, cleaning, transfer from external databases into (ideally) one clinical database which supports smart reporting features. More efficient monitoring efforts are also supported by CTMS tool, which has become a vital system when planning needed site visits, ensuring data collected is available fast, found issues can be tracked and metrics/KPIs can be collected. It will be the final decision of each Sponsor how they want their clinical data to be collected, on paper or with EDC. However, with continuous EDC tools’ development and improvements in the overall user experience and functionality, while offering a compatible price, when it comes to the speed of clean data availability, the two methods are in different categories, with EDC offering a much more strategic approach.
Data from external sources
Clinical data is not only data entered into (e)CRF. Depending on the study phase and needs, data from external sources such as central laboratory, ECG or other vendor data, as well as IWRS and ePRO functionalities can be a standard requirement. As data from these sources is often linked to endpoint data, having access to it close to real-time is key to successful analysis of the study from very start until finish.
Studies are different, therefore the number of laboratory tests and level of ECG or other vendor data needed can also vary. However, if considered critical, this data must be made available to applicable study team members as soon as results are available. Transferring laboratory and other external vendor data to EDC is nothing new. In addition, it is also critical to establish system alerts to the safety team in case any flagged values are collected, as making data available on time only for site users is not always sufficient. It is a small extra step, but an important one when ensuring patient and timely action.
Ensuring externally collected data is made available in EDC will also allow faster data reconciliation, issue solving (e.g. data discrepancies are identified) and more speedy data review.
Features like IWRS (together with medication inventory) and encoding tool are often kept separate from the EDC tool, as more cost-efficient means are sought after. Every system vendor has their own costs linked to each feature, but from time and user-friendliness perspective it only makes sense to link this with the EDC tool as well. Randomization is a critical milestone in each study and efforts are made to keep the number of incorrectly randomized patients as low as possible. In-built IWRS tool will allow adding basic background checks (e.g. inclusion/exclusion criteria, abnormal laboratory and vital signs values) prior to allocating the subject to a study arm. Also, confirming randomization can be connected to alerts being sent out to the study team, meaning again that in real time – everyone who needs to know will know.
While it doesn’t seem like such a critical item to encode adverse event, medical history and concomitant medication data (as applicable) using an external tool, having this in EDC and allowing designated study team members access to the common terminology data always has its benefits. Not only will an inbuilt encoding tool enable easy reporting, it will also ensure consistency and cleaner data when doing the analysis, changes made can be tracked, encoding can be done in batches and via autoencoder, and queries issued within the tool in case questions arise during the process. Continuous encoding of data will make monitoring and assessing entered data possible, and ensuring SAE reconciliation can be done per study requirements.
A hot topic now is also ePRO and mHealth. Paper diaries have been used for a long time, but per research, paper diary compliance is proven to be as low as 11%.(4) As data collected using this means is again of important value, such low compliance must be challenged. One way is by using ePRO solutions. Not only will this allow access to collected data in real time, it also supports use of (validated) questionnaires, opening a communication line between site and study participant, notifications in case timely entries have been missed and alerts if contradictory data is entered. Added costs which can apply when study participants must be provided with tablets or smartphones can again be decreased with a BYOD approach (Bring Your Own Device), as many possible study participants own devices with needed requirements (device analysis can be done to ensure its functionality meets the needs) eliminating the need to carry multiple devices with you to report data.
There is a very large amount of documentation created during the clinical trial. Even a few years ago it was standard practice to collect data into a paper Trial Master File (pTMF), sometimes spending months cleaning and organizing the folders prior to finalization, while having no clear picture of the status during the trial. This too is changing very rapidly.
It is a huge adjustment in mindset to move from paper to electronic TMF (same as with CRF), but again, the benefits are not to be underestimated. An eTMF can give you a status of your study documents at any time. It is easy to see what is pending or outstanding, what are the quality issues, compare a document between versions, apply electronic signatures, report on the ongoing status and pinpoint any areas which can be improved. Not only will you know when a document was created and finalized, it can be easily tracked who reviewed, approved or even accessed the document and when. Every step can be reported on, and metrics and Key Performance Indicators (KPIs) can be collected.
eTMF can also decrease security risks. Documents are stored in “the cloud” with regular back-ups, there is less need for sending documents via emails, and the study team can always be sure that they are working based on the latest version of documents (as available in eTMF).
An active TMF approach decreases the time needed when preparing for audits and inspections, not to mention the possibility of conducting the audits remotely, which again speeds up the process while decreasing the costs. To standardize the approach to eTMF even further, many companies in the industry have implemented the DIA TMF Reference Module as the basis for their TMF structure, again cutting the time needed to familiarize oneself with the company specific folder structure, especially when it comes to audits and inspections.
Work on TMF is no longer a task meant for the close-out phase of the study. As Veeva TMF Maturity assessment from 2016 also states – companies who have moved to eTMF applications are seeing better visibility into performance metrics (55%). In addition, organizations using metrics have reported:
improved audit and inspection readiness (67% vs 29% of those not using metrics),
use of automated tracking /reporting of documents (53% vs 14%),
better visibility into performance metrics (53% vs 14%),
cost savings (47% vs 10%). (5)
There are very many benefits when speaking of an active eTMF and having data related to study documentation available at all times.
With all the demands on study execution, and all the requirements and guidelines to follow, who has time to think about how to take advantage of instant data availability? Wouldn’t it be nice if after recruitment starts, half of the study team could take a break and only worry about the study quality shortly before study end? The reality is that today we have everything at our disposal to build in oversight from the very beginning; from CRF data management, vendor data and documentation compliance. As such, we all need to think about fast data availability and how to best take advantage of the tools available to use the data effectively, as this is the only way to ensure all parties are working based on the same assumptions and targets, and are making decisions and taking action not only efficiently but also effectively. The cost of not knowing is far too high. As we can’t hide from the data, we should make it work for our benefit.