Dooloo Case Study

E-Health Company Creates a Knowledge Graph Solution to Help Patients with Chronic Pain

The Challenge

Nearly one person in five suffers from chronic pain, a complex, multi-factorial, multi-dimensional disease whose symptoms are individual and subjective. Often disabling, it is also frequently associated with numerous comorbidities that worsen the prognosis and rule out overly standardized treatment.

Dooloo sought to create a platform that would feature available resources on chronic pain while also providing recommendations to patients based on their self-reported health and behavior.

Such a platform would need to synthesize a vast amount of data, including all of a patient’s medical records across providers and their prescription history as well as information on the latest treatments and their efficacy. Dooloo also added a key dimension that classical approaches focused on diseases and treatments often omit: patient behaviors.

Laurence Sergheraert, CEO and founder of Dooloo, put it this way: “Our challenge was to be able to take into account everything that happens in the patient’s life to understand them better and to support them effectively in their journey, and also guide caregivers in their decisions.”

Creating this platform required consolidating data from many different sources that came in an array of disparate formats.

At first, the team tried building a solution in a relational database. “We spent a lot of time arranging our data in two-dimensional tables, in an SQL database,” said Laurence Sergheraert, “but we quickly realized the limits.” Such platforms cannot flex to absorb dynamic data and answer new questions.

The Solution

The team turned to an approach that is increasingly used in medical research: creating knowledge graphs that assemble and organize large quantities of data relevant for solving complex problems. Knowledge graphs support open-ended queries and are ripe for advanced analytics using AI and ML.

The Dooloo team heard about the Neo4j Graph Data Platform and began looking into it. "We were amazed by Neo4j's ability to visualize the data we were storing and the relationships between the different data points,” said Laurence Sergheraert. “This would transform our own analyses and show patients and healthcare professionals the richness of the data and the kinds of queries we could make.”

Neo4j conducted an Innovation Lab with Dooloo that brought together patients, healthcare providers, subject matter experts, and technical staff to brainstorm the kinds of questions they would want to answer.

Through an iterative process, the team coordinated the various points of view of the participants and created a property graph model in Neo4j that enabled them to import the data and rapidly build a prototype.

The process highlights the ease of evolving the graph model to incorporate new data sources. Dooloo saw how quickly graph-based solutions can be created – critical to compete in e-health.