Case Study | Heor

It is time to use an Interactive Flow Diagram to Tell Patient Narratives

Published On

With the ever-increasing degree of digitalization of health records of various kinds, it has never been easier to track different facets of a patient’s journey. Not only it is now possible to trace all the different events associated with one particular medical encounter, there are also ways and means to probe into what happens before and after the event. As a result of the emergence and popularization of Internet of Things (IoT), it is conceivable that we will soon be able to tap into even more streams of information recorded and generated by all those wearable devices and smart sensors.

The richness of the scope of the knowledge contained in such convoluted time-series events can be extremely hard to visualize using simple charts and graphs. We at KMK Consulting believe it is no longer merely a novelty, but rather a necessity to use an interactive smart diagram to present narratives associated with patient journeys.

Take the example of a disease treatment pathway. It may be easy to follow how a single patient switches from one treatment to another, and move on to the next. When millions of patient records are compounded, the transition sequence from one drug to the next may parallel a multi-state stochastic event. It is still possible to depict all these information in a cross frequency dashboard, listing all possible treatment pathways and associated flows. It may require some exceptional acumen and knowledge to interpret such results.

In the dashboard above, it is evident that when the number of possible drug switches increases, the number of potential treatment pathway explodes exponentially. It becomes excessively hard to read, and extremely cumbersome to make comparison and inference. The underlying trend of the pattern could be well camouflaged in thousands of different possibilities, and may not be easily identifiable. To confront such obscurity, we will outline two simple, yet elegant interactive diagrams with dynamic flow analysis to showcase the power of interactive graphics.

Tree Diagrams (Interactive Tree Diagram)

A tree diagram may be the simplest way to visualize a series of probability events. Such diagrams can be conveniently enhanced to expose more hidden relationships than just plotting the possible pathways. In the example above, the size of each circle/node is directly proportional to the number of patients following the treatment path. It provides an explicit visual reference to compare the distribution of different treatment pathways. Furthermore, additional information can be triggered to display in floating boxes.

Sankey Diagrams (Interactive Sankey Diagram)

Sankey diagrams were originally invented to illustrate the flow of steam in a steam engine. It has since found a variety of applications in detailing the flow of any quantifiable matter. Sankey diagrams inherently carry a very strong visual cue of the flow quantity over different stages or time. In the example below, researchers can draw immediate conjecture of the embedding drug switch pattern in a large cohort, all thank to the competence of this interactive Sankey diagram.

At KMK Consulting, we believe a picture is worth more than a thousand words. A well-suited data visualization will not only be better to articulate the multitude of information, but also give a vantage point to healthcare practitioners and researchers to decipher any hidden message.

Both these two applications utilize the popular open source D3.js JavaScript library, and can be readily rendered/implemented on any web-enabled devices or platforms. For more information on the topic of interactive tree diagrams and Sankey diagrams, you can read our paper at PharmaSUG 2018.

Get The Latest Updates

Subscribe To Our Monthly Newsletter

No spam, only the content you’ll want to read.

Details about how we process your Information is available in our 

Privacy Policy

Watch our Latest Event

Ready, Set, Launch

Why Technology is Key to a Successful Product Launch

Kun Liu

Patrick Retif