Changing regulations and new requirements about how data needs to be gathered and presented in a submission is creating challenges for life sciences companies, and in particular for smaller companies that simply don’t have the resources to keep up with the requirements.
By Mark Tumelty, Manager Biometrics and CDISC, Life Sciences, CSC
As I mentioned in my last blog on CDISC, “Dress Rehearsal over for CDISC’s datasets standard SDTM,” there is often a lack of understanding about the wider implications and intentions of various industry standards. The same is true with biometrics. This is a very broad term, but loosely biometrics refers to the analysis of study data using various statistical methods to reflect the impact on the body of a drug, a piece of equipment, a procedure and so forth.
A thorough biometrics approach is about planning, preparing and validating comprehensive data packages for regulatory submissions, and incorporates CDISC data and other data that regulators use to review applications and mine data.
Avoiding Red Flags
Regulators are looking for quality data that demonstrate thorough analysis and research. For example, when submitting safety information, a company will need to collect data on adverse events, concomitant medications, medical history, lab data, and so on, and present it clearly in the reviewer’s guide, which is designed to assist reviewers with the process. If reviewers feel some data is missing or not thorough enough – for example, perhaps the company didn’t collect concomitant drug data cleanly so reviewers don’t know what other meds patients are taking while taking the product in question – that could send up a red flag.
The problem is once regulators question one thing about a submission, they question everything. So every piece of data needs to be built into the submission in support of the company’s analysis of the product. It’s not just the safety data, but also the efficacy data – does the drug work? That is why the reviewer’s guide is critical as the data can get very complex. Companies need to ensure the data is in a format the regulators expect and is thoroughly explained. A biometrics team spends a lot time working with data; summarizing, analyzing, cleaning, etc., that they often become the key contributor to provide a high-quality reviewer’s guide. The process of updating the reviewer’s guide needs to occur throughout: before the study gets under way, as the study is going on when there are a lot of check points that might require a company to adjust the study and how it is collecting data, and of course within the submission.
Since this process brings together so many pieces of the puzzle, it’s important to have a team or partner with a provider that can help fit the pieces together. These elements include CDISC data conversion and validation, data conversion of tabulation data (SDTM) and analysis data (analysis data model – ADaM), production of data definition files for CDISC data sets (define.xml, etc.) and of data definition files for legacy data (define.pdf), and interpretation and summarization of trial results.
A Positive Outcome
Recently, the biometrics team at CSC worked with a small pharmaceutical company on the submission of their product, a particularly important moment for the company as the product in question was their main focus. The company had some experienced people on the medical side, including clinicians and submission people, so they had a good idea of what was required of them, but they weren’t familiar with the finer details, for example, what versions of SDTM were required, nor did they have the staff to manage the process.
We contracted with them to take on that work, managing the SDTM specs and the analysis data model (ADaM) specs, and worked with another contracting company to do statistical work. The work involved handling the table package for the company, all the figures, and listings – the whole reporting piece. In addition, we put together their submission package, including the reviewer’s guide. We did the required define package, which is an explanation of the actual data sets, and took them all the way to submission.
We submitted that whole piece and stayed on to support and answer questions from FDA, and after a couple of rounds the company ended up winning approval of the drug. This was an important and significant success for them, and one that would have been difficult to achieve without proper management of the data.
The fact is that if the regulators aren’t satisfied with a company’s submission they could reject the file. And while denial of approval is the most extreme case, it’s not unprecedented. Even delays for questions and resubmission will be costly for a company due to the loss of revenue for every day a product is not on the market. So getting the biometrics data right is crucial.
To learn more about biometrics in regulatory submissions, contact our biometrics team.