Launching a Patient Registry: Seven Factors for Success


 

 

Charles Makin,
Vice President & Global Head, Late Phase & Observational Research

Because they are economical and flexible, registries are ideal for collecting data from a wide range of patients and practice sites, generating the real-world evidence regulators and payers require to demonstrate effectiveness and safety, and to support label and coverage extensions.

However, the flexibility of registries also creates risk. To generate useable, uniform data, registries must be carefully designed and administered. Consider these seven factors to design and launch a successful registry.

1. Consider potential analytical issues
The goal of a registry is to answer questions about how treatments perform in the real world. These questions are typically specific and often examine diverse factors, such as effectiveness by patient group, effectiveness compared with standard of care or other treatments, patient function and quality of life, and impact on the use of other services and overall care costs. Not collecting the right data, or enough data, to support valid statistical analyses to answer these questions can lead to study failure. It is important to carefully consider analytical issues as you design your registry to ensure it can address critical questions.

First, clearly define all the variables to be collected and identify how they will be collected. At the same time, develop your statistical analysis plan, to understand what data are required to answer your questions, and the timeframe and sample size needed to adequately power your analysis. Second, think about how you will address any confounding issues, for example, through stratification, propensity scores or matching, to ensure your analytic design is fit for purpose.

2. Address known biases
It is critical to identify and address all known biases at the design stage, to help prevent interpretation problems later. Potential bias sources include patient selection practices within sites, as well as systematic differences among sites, such as between practice specialties, patient populations served, and community and academic settings.

3. Broad entry criteria
Ensuring your disease registry is not limited to a specific treatment will facilitate a more robust range of comparisons, and enhance your registry’s credibility and generalizability to the target population. However, using such broad entry criteria may show your treatment is not the best, so be prepared to address this risk.

4. Optimal ratio of incident and prevalent cases
In many cases, it would be ideal to enrol only newly diagnosed cases to ensure comparability of results across the study population. However, you may need to enrol previously diagnosed cases to reach an adequate sample size or broaden a study’s scope for comparison purposes. Consider your study objectives and analytic requirements when balancing between newly diagnosed and existing patients. Identify patients by diagnosis status at enrolment so analyses can be run with or without either group.

5. Targeted data collection
Carefully balance how much data you collect for your study objectives against the cost of collecting those data. Collecting only the data you plan to analyse will increase efficiency. Gathering more data might allow you to answer additional interesting questions, but can also burden sites and patients, which may drive up attrition and curtail recruitment. Collect both clinical and patient-reported data, since neither can substitute for the other.

6. Automate data collection wherever possible
You can enrich your registry at a relatively low cost by automating the capture of objective medical data such as labs, prescription refills, and hospital and emergency room utilization. Automating capture of broader data, such as progress notes and discharge summaries, directly from electronic medical records also may reduce ongoing data collection costs when it is possible, though initial costs to set up this capability may be higher if it doesn’t already exist in your study sites.

7. Control data access
Create a single analysis team and restrict data access to this team only. This helps prevent data misinterpretation and poor-quality analysis that may undercut your findings.

A registry’s success hinges on well-reasoned design and sound data collection practices. Considering the factors above will help ensure your registry will meet research and product development requirements, and maximise product value – though this requires considerable late-stage study expertise.

An experienced partner can help. ICON has developed and successfully conducted more than 100 registries, including clarifying critical research questions, developing study designs that address them, recruiting and monitoring sites, and collecting and analysing registry data. Contact our experts for a consultation on your registry and other post-market clinical study needs.

 

OTHER ARTICLES CATEGORISED UNDER

  • Commercialisation & Outcomes
  • Real World Evidence