Hands-On Healthcare Data

Chapter 8. Graphs, Harmonization, and Some Final Thoughts

We have spent the past several chapters diving deeply into various aspects of working with real-world healthcare data. We started with a pretty high-level view of healthcare data, starting with an overview of common sources of healthcare data. One of the most challenging aspects of working with RWD is that the number and types of data sources is virtually unbounded. As clinicians, researchers, entrepreneurs, and tech companies continue to innovate, we will continue to see an increase in the different types of RWD available.

When we use the phrase real-world data, most of us immediately think of electronic health records, claims, and clinical registries as the most common sources of data. Many also include noninterventional studies, often referred to as observational studies. All of these studies are similar in that they typically focus on collecting patient demographics, conditions and diagnoses, medications, and other interventions.

However, as we continue to dive deeper into the world of data generated (and collected) for the purpose of caring for patients, we quickly venture into a whole new frontier. Whether it’s an app that uses your typing speed to detect changes in mental health, a phone that uses your gait to track movement disorders, or a baseball cap to detect seizures, the possibilities are endless. However, it does mean that nearly anything can be considered real-world data if it can be used to detect, treat, ...