Collapse All Expand All

Goals

Often in healthcare, understanding the patient journey is a vital first step in the developing personalized treatments. The goal is to get a quick understanding of the journey a patient takes from a primary diagnosis (like cancer or pre-diabetes) to a clinical outcome (like tumor progression or advanced diabetes). By understanding this journey, clinicians and researchers can devise improved treatment plans and even propose new drugs.

Users

Population health experts and clinical researchers.

Problem

The real world data that describes the patient journey in the healthcare system is messy and the each patients journey is often unique. Machine learning approahces can often help pick "optimal" paths, but these algorithms can ignore the detailed variation that give the clinical researchers a fuller picture of the patient population.

Solution

The visualization above shows the possible journeys from "primary diagnosis" to "present with clinical outcome" for a hypotheical patient population. Instead of relying solely on machine learning to pick the "optimal" journey, we can use similar approaches to select and highlight possible patient journeys. A knowledgeable clinical user can then explore these journeys to decide which might be actionable for treatment plans and drug development.

The complexity and noisiness of real world healthcare data often leads to many more patient journeys than have been displayed above. Thus, the clinical researcher requires the ability to focus on certain journeys or events in order to pursue specific hypotheses. Path highlighting and collapsing allows the user to focus only on those paths of interest. In addition, since some journeys may be more common than others, the number of patients that follow each journey is proportionate to the width of each path and size of each node.