Data management in observational studies
By: Zia Haque | September 16, 2016
Maximizing the value of real-world research through the proper data management strategy.
In the landscape of observational research studies, there are nuances to data management practices that need to be understood if research teams hope to be able to extrapolate results from these studies to real-world settings.
It is tempting to assume that an observational study is similar to an early phase study in terms of how data is captured and managed. Data managers in both settings review data, send queries, resolve queries, and code data. But observational studies operate in very different settings, with different guidelines and oversight that demands a unique data management approach. Where randomized clinical trials exist in a highly controlled and monitored environment with very specific patient demographics, observational studies are conducted in the real-world where the studies aim to replicate real world patient – physician settings, with minimal to no interventional practices. They also have to make the most of limited data collection opportunities and infrequent on-site monitoring.
These data collection ‘obstacles’ are what make observational studies so valuable, because they reflect the real-world experience in which the medication is used. But they also impact how the study is defined, and how data is collected and analyzed.
To avoid problems and increase the chance that real-world trends can be confidently extrapolated from the data, decisions need to be made at the protocol design stage to ensure the process is fit for purpose. That begins with the research design. To be effective, study teams need to rein in their desire to answer every research question. Gathering as much information as possible is always a goal, but asking too many questions or demanding too much time can alienate both sites and patients, resulting in loss of site engagement. The fact that sites participating in observational studies are typically therapeutically focused, while being rather research naïve is an important consideration in designing easy to use data collection tools.
Study teams also need to recognize that site staff on Observational studies typically only log into the Electronic Data Capture (EDC) periodically, after a certain time or number of patient visits. This can make managing the trial more efficient for their purposes, though it may mean there will be lags in data currency and outstanding queries. The site staff’s level of familiarity with clinical research may also be limited, which will translate to limited knowledge of electronic case report form (eCRF) completion, query resolution turnaround times, and the need for precise responses to queries. As a result, eCRFs should be designed to ensure maximum data cleaning activity can be achieved at time of data entry into the EDC. eCRFs should be offered in formats that are easy to navigate for sites with limited EDC experience, and feature optimized conditional branching of data fields, drop down menus and radio button options to assist with streamlined data collection.
To ensure safety issues are promptly addressed, sites should be trained appropriately on safety reporting requirements to avoid risks such as inadvertent entry of serious adverse events (SAE) in free text fields that could lead to under reporting of safety events. The EDC should also be programmed to launch SAE forms and send automated alert e-mails when serious adverse events are entered in the EDC.
Finally, study teams must embrace raw data that conform to early phase standards. The quality of data in observational studies should be a true representation of the area of research from where the data has been collected. Observational studies afford opportunities in the real-world setting that cannot be replicated in the strictly controlled environment of a clinical trial. However the study teams need to acknowledge and relinquish a certain level of control over tight data conformity for these trends to emerge.
Observational research has become a vital part of the drug development process. Understanding the unique benefits and constraints of observational research, and factoring these issues into the study design and technology choices will help study teams generate the most value from these vital real investigations that offer niche opportunities to study medicines in real-world settings.