Recruiting patients is one of the most challenging—and costly—aspects of rare disease research. It is estimated that about one-third of clinical trial failures overall may be due to enrollment challenges, and with rare disease research the obstacles are even greater.
Traditional methods for recruitment of patients are frequently not adequate to populate these trials. Small numbers of patients who are spread over large geographic areas, as well as singular aspects of specific diseases that may not be well understood make it difficult to find and recruit enough patients to meet enrollment goals. And as more companies pursue rare disease and precision medicine research, these recruiting challenges will only increase.
One way to mitigate these risks is by leveraging real-world data and innovations in advanced analytics and machine learning technology to develop more targeted recruiting strategies. Access to big data and advances in data analytics tools are enabling an information revolution in the biopharma industry. These analytics tools can be applied from early trial design and protocol planning through enrollment and study conduct to tailor the research process to the specific patient populations. We have already seen applications of these technologies result in dramatic improvements to identify hard-to-find patients.
However, there are hurdles to incorporating this data effectively. At DIA 2017 Annual Meeting in Chicago (June 18-24), I will participate in a session on rare disease recruiting and the use of big data with investigator and patient advocacy outreach to provide clarity on this topic. The session will explore unique recruiting strategies for rare disease research, and will share specific examples from a trial of Idiopathic Pulmonary Fibrosis patients to highlight how combining real world data and collaboration with patient groups in early planning stages, can improve trial design and accelerate patient recruitment.
Data is only part of the process
Raw data on its own is only a resource, and requires expertise, context, and perspective to add value to any clinical research project. One common challenge in using data to inform trial design or recruiting is that a single database may not provide information that is consistent with the known disease landscape, or it can be biased, which may cloud decision-making. For example, patients in a specific community or healthcare system may be more likely to be prescribed a certain medication or to follow a unique treatment path that is not reflective of the broader patient population. In other cases, physician data may suggest a high percentage of patients are receiving a specific prescription however pharmacy data suggests they aren’t filling these prescriptions. Researchers may also fail to access the right sets of data if they don’t know what to look for. If they only use data from centers of excellence, for example, they may miss patients being treated by local physicians; or in relying on outdated data, they can make false assumptions about current standards of care.
To generate the most benefit from any of these data sets, research teams need to enhance them with insights from other trusted sources.
In the DIA session, I will discuss how database-derived information has been used in combination with feedback from physicians and patient groups to generate an improved understanding of the disease landscape in Idiopathic Pulmonary Fibrosis. By combining raw data with more ‘traditional’ sources, we were able to enrich the information and improve the context of the data, to the benefit of the trial and the patient community.
Integrating multiple data sources with ‘soft’ knowledge from patients and investigators can help researchers gain context from the data, and decide when and whether to expand the data search or modify the interpretation. It is this combination of information and expertise that brings the most value from big data to the clinical research process.