Top 5 Data Pitfalls in Oncology Clinical Trials and How to Avoid Them

Top 5 Data Pitfalls in Oncology Clinical Trials and How to Avoid Them

Introduction

Pitfall No. 1 & Recommendations

When Protocol Design Doesn’t Consider the Patient Experience

Pitfall No. 2 & Recommendations

Inflexible Database Builds

Pitfall No. 3 & Recommendations

Data Collection that is Not Fit-for-Purpose

Pitfall No. 4 & Recommendations

Not Addressing the Unique Data Cleaning Challenges of Oncology Studies

Pitfall No. 5 & Recommendations

Insufficient Preparation for Study Close

Conclusion

Trust the eClinical Solutions Oncology Experts

of eClinical Biometrics Services

Clients’ Portfolios are Oncology Based

85% 90%

of eClinical Solutions Data Management Leads are Experienced in Oncology

“The (eClinical Solutions) team’s knowledge of Karyopharm data standards and processes, combined with their expertise in data collection platforms and clinical data management practices, helped us meet our accelerated timelines for this important new (oncology) research.”

SVP Development Operations at Karyopharm Therapeutics

Kristan Gallitano

Pitfall No. 1

When Protocol Design Doesn’t Consider the Patient Experience

For patients participating in oncology clinical trials, this often represents one, if not the only remaining treatment option for their illness. While this is a powerful incentive for patients to remain committed to a study, oncology trials are lengthy and complex. Studies can take years to move from early to late phase and require a great deal of effort from patient participants and their caregivers. If the protocol design requires a heavy site-visit schedule and overly complex procedure requirements, retention may suffer. For example, if patients are simply too sick for frequent travel, or are so fatigued that they cannot fill out questionnaires with regularity, they may opt to drop out of the trial. Even if they remain enrolled, requirements that ignore the day-to-day needs of patients will result in poor data quality.

It is critical to consider how to make

Oncology patient trial participation easier during the protocol design stage

Recommendations for Avoiding This Pitfall

The Importance of Patient Centricity in Early Protocol Design

It is critical to consider the patient experience during the protocol design stage of planning. Designing study protocols that reflect the oncology patient’s experience can make participation easier, leading to better retention and higher quality data.

How to build more patient-centric protocols

Making Data Collection Easy for Trial Participants

Data collection strategies that include decentralized clinical trial (DCT) approaches can help minimize burden for oncology patients and their families. DCT elements allow for data to be collected remotely via computer or smart device. Patients or caregivers can submit eConsents and eDiary entries from home, reducing the requirement for frequent travel. Data can also be collected via wearable devices or sensors, allowing for frequent or continuous data submissions. Additionally, some check-ins may be possible via telehealth, or perhaps via an in-person visit at a clinic near the patient’s home versus at the central trial site. While most oncology studies will always require some in-clinic visits due to the need for things like infusion therapies and periodic imaging tests, allowing some data to be collected and submitted remotely can make a significant impact on patient quality of life.

Identifying Risk Mitigation Strategies Early

Risk-based quality management (RBQM) can help you dive into past study data to better understand the kinds of challenges that make participation difficult for patients. Existing data can show patient dropout trends and decipher what factors led to withdrawals. By analyzing this data through the lens of RBQM, risks and effective mitigation strategies can be identified early on, allowing teams to de-risk protocols during their design and as part of any necessary protocol amendments over the course of the study.

Pitfall No. 2

Inflexible Database Builds

As stated earlier, oncology trials can last years. This makes protocol amendments inevitable. It is common to see new cohorts, arms, and other parts added throughout the course of the study. Adaptive approaches are needed to set a foundation for making these downstream modifications.

Recommendations for Avoiding This Pitfall

Preparing for Oncology Trial Protocol Amendments

A static casebook is a recipe for disaster because large and/or complicated amendments can force drastic changes to database design, adding cost, increasing cycle times to amend the database as needed, and potentially put data at risk. Thus, it is important to build databases that are flexible and enable the ability to more easily accommodate for amendments, large or small.

Statistical Review and Input on eCRF Design

Early planning is crucial for building more effective databases for oncology trials. To this point, performing a statistical review during electronic case report form (eCRF) design is a critical step towards successfully achieving trial objectives. If, for instance, statistical review is delayed until just a few days or weeks before a large data cut, any data accuracy issues or collection errors can result in the need for expensive and time-consuming corrections, putting the entire study in jeopardy. Including statistical review as an integral part of the eCRF design process helps ensure the right data is being collected and reduces the possibility of downstream issues.

Building the Database with Research Sites in Mind

When building electronic data collection (EDC) databases, it is difficult to predict what changes will be made to the protocol over the course of the study and what ramifications amendments will have. Setting the precursors early in the build design phase can reduce the need to overhaul the structure later in the study. For example, having to repeatedly look back at retrospective data can be problematic. A way to avoid this is by keeping track of data at the visit level, which allows teams to compare visit dates with protocol versions to ensure that patients are current.

Site burden should also be a consideration. Flexible databases that can easily allow site teams to carry over data, such as tumor identification information, from initial screenings and baseline visits. This helps to eliminate the need for transcribing or repeated manual data entry. This also reduces errors while allowing study teams to focus on other important trial responsibilities. Further, tools that provide easy visibility throughout the database are beneficial, allowing teams to get the information they need quickly to stay on top of study progress, ultimately leading to cleaner data.

Data Collection that is Not Fit-for-Purpose

It is very important that the EDC database is designed based on the specific data collection strategy to be employed, as well as the needs of trial participants and regulatory requirements. Input should be sought from all relevant members of the study teams across all operational aspects of the trial to ensure that the critical data needed to meet the protocol objectives is captured appropriately. This level of collaboration is vital – if it does not take place, it can result in poor data quality, incomplete data and failure to answer the research questions being investigated. or greater issues that result in costly delays to data delivery.

Target the right data and ensure proper access

Recommendations for Avoiding This Pitfall

Making Data Access Easy, Despite Multiple Data Sources

Having to look at volumes of study data from multiple sources is often inefficient. Luckily there are solutions that help study teams to break through the clutter and get right to the useful study insights they need. For example, the elluminate® platform gathers data from a broad range of data sources common in oncology research and transforms the data to provide valuable insights through analytics, listings, patient profiles and reports. Study team members can take action on potential discrepancies and identify clinical and operational risks and data trends at a glance.

An oncology specific example would be a data manager seeking to easily find and track the use of prohibited concomitant medications and any instances of treatment emergent adverse events. The elluminate platform allows data managers to do this easily.

Getting Vendor Selection Right

In oncology trials, it is common for sponsors to engage with multiple external vendors for the purpose of data collection. To set the stage for success, vendors should be evaluated based upon how their technology and processes will be able to adapt to the changing needs of the trial. For instance, if an additional cohort is required, is the interactive response technology (IRT) system able to accommodate the necessary adjustments? If additional central lab or specialty lab collections are added or if visit collection time points are changed, are the selected vendors and their solutions flexible enough to handle these occurrences?

Additionally, having access to the data on an ongoing basis is also critical for effective communication to and from sites in oncology trials. Vendor solutions need to allow for seamless sharing of timely data among all relevant study stakeholders. This helps to reduce query burden based on reconciliation and avoid any kind of delays in data deliveries.

Pitfall No. 4

Not Addressing the Unique Data Cleaning Challenge of Oncology Studies

Oncology studies involve huge volumes of data from many different sources over long periods of time. This makes data cleaning a particular challenge compared to other therapeutic areas.

Recommendations for Avoiding This Pitfall

RBQM Revisited

The utility of RBQM in oncology trials does not stop at protocol design. It is also extremely useful in the data cleaning process. Implementing a risk-based approach to the data cleaning process prioritizes the intense interrogation of critical data while review of non-critical data is still performed, but with a lower priority. This approach reduces the time, effort and cost of reviewing non-critical data and shifts resource efforts to focus on the data that matters. This process should include the setup of regular meetings with relevant team members to document critical data sets and ensure that appropriate quality checks are in place. The data cleaning strategy should involve collaboration with the entire study team to ensure data is continuously reviewed, data quality issues are proactively identified, and risk mitigation steps are in place. This helps to maintain data integrity over the course of the study.

Utilize Near Real-time Analytics to Evaluate Data Trends

Study teams working to manage oncology trials are typically looking at multiple external vendor files every day. They are opening up data transfer agreements and files from numerous data sources and often different software platforms in order to see data status and understand how their study is trending. This

can be a tedious and slow process. However, tools like elluminate give study team members access to all data from all different sources, in one place. They have access to EDC data sets and external data sources that allows them to dive in on any particular subset of data, such as the most recent adverse event data, and compare that with other trial data such as labs. This information is refreshed as frequently as needed, but daily at minimum. Study team stakeholders maintain constant access and visibility into the data. Data managers can perform quality oversight and share trends and issues with other study stakeholders to ensure that issues are addressed in a timely fashion.

Additionally, elluminate gives data managers and study team stakeholders access to exception listings and reconciliations via role-based workflows, patient profiles and analytics for data review, exploration and operational oversight. Because this data is refreshed on an ongoing basis, data managers are getting frequent updates, allowing them to see issues in near real time as data is entered. This helps to get data reviewed and queries out quickly. This is critically important for oncology trials, as teams are able to achieve accelerated timelines for deliverables, keep studies on track and always be prepared for analysis.

Pitfall No. 5 Insufficient Preparation for Study Close

Many factors can delay data delivery. As discussed already, the volumes of external data collected from various sources over the course of a given oncology trial can be difficult to reconcile and aggregate. Database discrepancies and unresolved data cleaning issues can also have a significant impact on data delivery at the end of a clinical trial. Finding out about data problems at the end of the study can have catastrophic results, potentially putting huge amounts of data and the study itself at risk.

Recommendations for Avoiding This Pitfall

Finding and Fixing Issues as they Happen Saves Time and Cost at Closeout

Having an experienced data management team, is important, as is having a sound data management strategy designed to identify issues in as close to real time as possible over the course of the study — this can make or break an oncology trial. Part of an effective closeout plan is ensuring that clear and comprehensive data transfer agreements are in place with all vendors at the beginning of the relationships and that ongoing data reconciliations are performed, reducing problems caused when errors are only found at the very end of the study.

The right process and platform equals timely delivery of clean data

Choosing Cloud-based Digital Data Delivery

For many reasons, the research industry has largely moved away from delivering data at closeout via physical media. Embracing a digital approach that utilizes cloud storage and connectivity, study stakeholders can be more confident that data is as up to date as possible. Data can be shared directly among sites, sponsors and other relevant stakeholders, saving a great deal of time. This helps accelerate study close, as data can be shared, and receipt confirmed in a matter of moments.

Conclusion

Oncology clinical trials pose unique challenges. The sheer volumes of data, taken from multiple disparate sources over periods of years make early planning essential to avoiding the five data pitfalls outlined in this eBook. Focusing protocol design on the needs of patients (patient centricity), developing adaptive database builds, and having solutions in place to provide access and visibility into current data for all relevant study team members help set the foundation for study success. Collaborative and strategic data cleaning strategies, such as the approaches discussed here can help ensure successful final delivery of clean study data.

About eClinical Solutions

eClinical Solutions’ products and biometrics services are helping life sciences companies adopt a more modern data infrastructure to break down data silos, automate processes and accelerate the speed of drug development. With the elluminate Clinical Data Cloud, life sciences organizations can begin to maximize the value of their clinical data, help their teams accelerate development processes, and serve patients more efficiently throughout the clinical trial lifecycle.

For more information, contact eClinical Solutions at eclinicalsol.com