There has been much talk recently about the value registries can bring to clinical research. These online data platforms give patients and providers a place to record information about their condition and treatment that may not be available in a patient’s medical record, including how they were diagnosed, how their treatment is working, what non-prescription and complementary treatments they are using, etc. These data can offer vital insights into the broad patient experience that, in addition to knowledge gathered through the clinical trial, can better support trial results and help researchers understand the value and limitations of a particular treatment.

The challenge for biopharma companies is building registries that gather the right information from the right patients in the desired timeframe. As with any study, these projects face operational, logistical and methodological hurdles that have to be addressed during planning to ensure the registry delivers the quality and quantity of data sought.

To help industry stakeholders address some of these challenges, my colleague, Nancy Dreyer, and I are hosting a training program on registries and prospective cohort studies at the International Conference on Pharmacoepidemiology & Therapeutic Risk Management in Dublin, discussing how to design, execute and interpret data gathered in registries and prospective studies.

In the session, we’ll explore many of the regulatory, methodological and logistical challenges that organizations face when designing registries, along with best practices for avoiding common pitfalls. My section of the program focuses on considerations related to defining patient populations, including how to choose the right patients for your comparator group and how to reduce bias through patient selection, which are further outlined below. I also focus on new-to-market treatments throughout my talk, because we face additional challenges when studying these treatments that might rapidly disseminate, sometimes unevenly, throughout the patient population. With prior awareness and acknowledgement of these challenges we can work to reduce them through early planning.

These topics can be complex, but they are critical decision points that researchers need to factor into their study design strategies if they want the registry to generate robust, value-driven results.

  • Make sure your study population includes a diverse set of patients, but in certain cases, is also appropriately restricted to patients who will have similar treatment paths and responses that tie back to the research goals of the registry. For example, if you were building a registry to understand effectiveness of treatment for patients with colon cancer, you may want to limit enrollment to older patients only; colon cancer patients diagnosed in their 20s and 30s tend to have a very different disease experience, which could influence your end results. 
  • Use active comparators to create a clear baseline, and ensure overlap of patient characteristics between the treatment under study and that of the comparator. When designing a registry or cohort study, you want to be sure your population includes patients who have used or are using an active comparator – i.e., another therapy used to treat the same condition as your treatment. By limiting the participants in this way, you will generate data that will best highlight how your treatment works against current standards of care. In the case of our colon cancer example, you would want all participants to have received chemotherapy of some type (excluding those receiving no treatment or radiation) so that you could demonstrate the comparative outcomes of your treatment of focus within a similar population. 
  • Focus on new users of the treatment to increase study validity. When building a registry to assess the impact of a particular treatment, it is important to populate it with patients who have recently started on the treatment – “new users” – so as to provide a common treatment initiation date and circumvent under-ascertainment of early outcome events. If you include patients who have been on the treatment for an extended period of time – “prevalent users” – it may unintentionally create bias because those patients are more likely to have had a successful experience with the drug, otherwise they would have stopped using it and/or experienced the outcome event of interest, precluding their enrollment in your study. This means your population might unfairly show positive results while missing out on events experienced by earlier users.

These are just a few of the many factors that should be considered when building a registry or cohort study. It can seem overwhelming, but when pharma companies take the time to understand the intricacies that can impact study results and factor these issues into their design plan, they will build registries that deliver the greatest value in support of their research goals.

I look forward to seeing you at our session.