SDC Insights: The Biggest Mistakes CROs and Clinical Research Sponsors Make With Biostatistics & Data Management

Todd Bonta is SDC's VP of Clinical Data Management and has over 22 years of experience in CDM spanning all phases (I-IV) of clinical research. His responsibilities have included all aspects of the CDM field from CRF design through database lock across numerous therapeutic areas. 

Quan Doan is SDC’s Sr Director of Technology Solutions and has spent over eighteen years in clinical data management within biotech, pharmaceutical, medical devices, and CRO environments. He is experienced in leading early phase multi-center and multi-national clinical trials as well as managing data management activities for Phase I-III clinical and/or non-clinical research studies.

Adam Hamm serves as SDC's Sr VP of Biostatistics and Strategic Consulting and has over 17 years of experience in clinical trials research. Adam has technical expertise in adaptive designs and statistical software platforms that utilize clinical trial simulations for efficient and statistically rigorous designs for innovative trials, including early phase oncology trials. For over fifteen years, Adam has had successful and collaborative interactions with regulatory agencies, communicating and consulting on various regulatory and statistical issues on behalf of clients over many therapeutic areas.

Bonta: The biggest mistake I see teams make is failing to allocate enough time to engage the full spectrum of stakeholders in vetting their Case Report Form (CRF) design against the study protocol.

From my perspective, the most important part of proper data management in a trial begins with the design of the CRF. 

It's almost as important as protocol design and regulatory approval and it can be the hardest part of the process.  Without investing that time, items can be omitted, or teams lose the opportunity to ask questions in a clearer way.

Doan: I would add that it's also important to make sure that whatever data we expect to collect is done so for the right purpose. Depending upon the type of study being conducted or who is involved, the sponsor may want to explore an additional area of interest but if it's not relevant to the protocol or the statistical analysis of that specific trial, capturing this data can lead to a database filled with extraneous information that may be interesting to the researchers but not necessarily relevant to the trial.  It can also add to the complexity of keeping the database clean. 

Our teams are skilled at keeping everyone's attention focused on the objectives of the trial.

Another challenge is the redundancy of data. Because there are so many new systems being engaged now, data management teams need to understand what systems are being used to store and transfer data, what the IRT system is delivering and how it’s being integrated and the RBM strategy decided upon so SDC can properly develop the clinical database to support these activities.

Hamm: As a statistician, the biggest mistake I see is research sponsors or CRO's not involving the biostatistics team early enough. If a client is planning to use a consultant or managing biostatistics in-house but they're going to outsource SAP writing, programming, etc. to another vendor, it can happen that the research teams fail to engage the statistics team until well into the trial. This is a very common mistake that can negatively impact studies.

🗨️  Why do these mistakes happen?

Bonta: Teams often make these mistakes due to timeline constraints. Regulatory approval of protocols takes time and sponsors have corporate goals, perhaps they have made promises to investors, shareholders, or others. They're excited for the database to be ready and for that first enrolled patient so they crunch timelines not realizing the inevitable costs of wasted time and money from not vetting the CRF design early.

Doan: When we're doing early phase studies, sometimes sponsors have a clinician, engineer, or a scientist who wants to know more about their drug or device than perhaps what the protocol entails. They may be studying their product for use in epilepsy patients, for example, but also have some curiosity around four or five other indications. It’s our job to keep them focused on capturing data that helps support the current study scope, not building a database that houses data to mine for future studies. Sometimes sponsors don't realize that we can't just update the database without making sure that the protocol is amended, as well.

Hamm: Smaller companies make the mistake of not involving the stats team early because they don't realize the impact that biostatistics has on their study. They're very focused on the clinical details and don't necessarily plan out the timeline and strategy around the analysis and the explanation of their data. They may have a consultant who they believe is contracted to do it all, including the programming, but unfortunately, that's very rarely the case.

Another reason for not involving stats early is that as timelines are crunched, it can feel easy to push statistics and programming to the back end. Sometimes, if you’re working on a long-running study, programming can start later, however, in terms of the design and such, the statistics team should be involved early. When they aren’t, we see issues.

🗨️  What's the best way to avoid these mistakes, and who are the vital stakeholders to have engaged in order to steer around them?

Bonta: Whenever you are making critical decisions about your trial data, you need the full attention and engagement of the right people. The earlier you can do this in your process, the smoother it will go.  I've also seen situations where we believe we have the key stakeholders involved but after hammering out all the details, it’s decided to include input from an additional key opinion leader. These mistakes can add delays and possibly increase the costs of your program.  The vital voices are, first and foremost, the data manager responsible for the timeline and team oversight for the database architecture and structure, the biostatisticians, clinical project leadership team, pharmacovigilance, medical affairs, and the IRT project manager.

Recently, I worked with a customer on a CRF design for a complex, rare disease study. We had the right players at the table and it literally took six months for the entire team to agree on what needed to be collected.

At the time, the process felt incredibly laborious but, in the end, timelines were met, and we had the database ready for first patient enrollment.

Doan: I often hear sponsors say, “we can have the full database build completed in two weeks, right?” and my response is, “not if you would like to ensure a high-quality database and/or you have a recent study identical to your current one where you are only tweaking very minor details.”

Hamm: We strive to ensure that our clients have a good understanding of industry-standard timelines as well as an awareness of the key components, review cycles, etc. required to guarantee a high-quality clinical database.

I’ve never encountered a time when we regretted spending the time to set the right foundation.

🗨️  Can you work your way back from these mistakes once they occur?

Certainly, you can work your way back from most mistakes but there are always timeline and often financial implications. 

Doan: Problems can also be amplified when you start talking about all the new technology pieces being introduced. IRT impacts the Role-Based Security (RBS) system so your whole operational timeline can be impacted. 

Hamm: Sometimes a protocol amendment is necessary if there's something grievously wrong with the study design the sample size, etc.  This can cause you to do an interim analysis that you didn’t have planned but that's rare.   At the end of the day, mistakes can happen but usually, we can help our clients untangle themselves and get on the right path for successful regulatory submission.

If you would like to learn more about this topic, please refer to the following SDC resources:

✔️How to Prevent, Identify, and Implement a Clinical Trial Rescue | Webinar | SDC (sdcclinical.com)

✔️Managing the Successful Execution of the Clinical Trial Rescue Process | Webinar | SDC (sdcclinical.com)

✔️How Quality and Innovation in Your Data Services Leads to Success | SDC Watch Now (sdcclinical.com)

Other resources:

✔️The Society for Clinical Data Management

✔️Basics of case report form designing in clinical research (nih.gov)

If you would like to connect with Todd Bonta, Quan Doan, or Adam Hamm directly or anyone from SDC, click the Connect with Us button below.