Improving the Quality of Real World Data

Updated on February 22, 2023
Business woman drawing global structure networking and data exchanges customer connection on dark background

The volume of data is soaring, and so is its value. With approximately 30% of the world’s data volume currently being produced by the healthcare sector, it is unquestionable that the industry is approaching a golden age of data. The expanded clinical use cases for the data have spurred innovation in the sector, presenting new opportunities for its use by clinical researchers, hospitals, and life sciences companies. As a result, the FDA has accelerated the importance of Real World Data (RWD), defined as data relating to a patient’s health status not captured within a clinical or prospective trial. However, with greater volume comes increased scrutiny and new challenges. 

An ongoing challenge in leveraging real world data, particularly clinical data, to improve healthcare outcomes and drive research is determining whether the data being used is reliable and of high quality. This requires that the data retains its integrity throughout its lifecycle, from the capture of the information to the data transformations to the usage and derivative works.,  Given the relatively recent interest in large clinical RWD datasets, there continues to be a lack of standardization around best practices in producing and using clinical data at scale, coupled with the existing challenges within the healthcare ecosystem such as lack of continuity in patient records across systems, incongruent coding systems, different EMRs, and other healthcare software.

As a result, all parties should have some healthy skepticism when obtaining and leveraging data sources, even within their own organizations. Understanding where the data originated from, how it was refined, who did the work according to what set of guidelines, and how it was quality checked are key questions to ask prior to leveraging data. The lack of focus on data quality is seen across hospitals and life sciences alike, where datasets that are not well-understood are deployed against a wide range of use cases – data is sometimes inaccurate, incomplete, or not suitable for the use case at hand. Oftentimes, these are expensive purchases or initiatives where trust in the data and stakeholders is eroded. 

While there are risks involved in ambitious initiatives involving clinical real world datasets, adopting stringent clinical data management practices can help ensure that high-quality datasets are available to further the future of medicine and care.

First, because significant effort is required to navigate unstructured notes and search through volumes of data, time and resources must be devoted to prioritizing the training and retention of clinical data management staff. As the healthcare industry battles against labor shortages, a recent survey found that nearly nine in 10 hospital leaders surveyed mentioned labor challenges have changed how they manage their clinical data. Hospitals can take action to prevent burnout among staff, and by leveraging third parties and automation to alleviate workloads, clinical data management can continue to be prioritized. Similarly, companies with access to clinical data often avoid the path that hospitals have solidified as the best approach to create datasets for research and reporting, expert-driven chart review. Without clinical data management experts, data quality is nearly impossible to maintain.

Second, while the above approach may receive criticism for being manual, technology is an integral part of an organization’s ability to successfully produce high-quality data. The right data capture tools, coupled with sophisticated data engineering and other automation efforts can enable consistency in how millions of patient records are produced. Technology can be used to manage the clinical staff involved in the work, drive efficiencies, and programmatically assess the data for quality, whether that’s data accuracy, fill-rates, clinical quality checks, and more. With a strong end-to-end tech-driven process for managing clinical data and its associated abstraction and standardization steps, data provenance over every patient record and every individual data element can be established, creating transparency and ensuring trust in the event of skepticism.

Lastly, healthcare leaders need not only to say that using real world data for clinical research is a priority; they need to provide the proper framework, resources, and processes across their organization to make data quality a pillar in the organization. Part of this pillar is messaging across the organization for the sake of improving internal collaboration, aligning project goals, and discussing patient privacy protections. By elevating quality discussions, strategies around clinical data can be more pragmatic and aligned with what’s achievable, moving goals towards projects and initiatives where high-quality data can make an impact as opposed to lofty agendas unlikely to be supported by the current state of data. 

Because using real world data in clinical research is new territory for many hospitals and health systems, it’s imperative that the strategy of the organization considers data quality early to set a solid foundation of trust for all subsequent data-driven initiatives.

The potential of RWD is endless when it comes to clinical research and advancing patient outcomes–but only with the proper tools and practices. The highest standards must be met in all instances and throughout the entire process, including implementing a strong clinical data management process. Hospitals and health systems must begin prioritizing these practices in order to fully reap the benefits of real world and clinical data.

Victor Wang
Victor Wang

Victor Wang is SVP, Real World Data at Q-Centrix.