As an alternative to the force-layout network diagram, the Adjacency Matrix can provide a cleaner, more intuitive view of complex relationships. In this graph, 250 providers are arranged in the same order on the X and Y axes. The color intensity of each box in the resulting matrix represents the number of patients that have seen both providers in a given year, or how much the two providers’ patient bases overlap. The providers are then clustered by specialty, shown as the colored boxes along the 45° line.
This visual highlights some trends that we would expect. There is very little overlap between Primary Care specialties, as patients typically lock into a single provider. Conversely, radiology services see significant overlap, and show very little discrimination for which other providers are involved in a patient’s care.
Beyond such expected outcomes, this visual also has the potential to provide useful insights. High-volume providers appear as dark horizontal and vertical bands, and aberrations – a lone dark or light spot – highlights the close relationship between two providers. Examining such outliers offers an opportunity to identify root causes for relationships. Are referrals flowing in one direction primarily, or does there seem to be a mutual flow? Does the receiving provider offer particularly good quality, value, or other intangibles, or could inertia be the greatest motivating factor?
Among all the figures shown in this series, most were entered upon with no sense of what the outcome would be. Some merely provide functionality, others offer a new perspective. Mostly, the opportunity to create dynamic (and attractive) visualizations of healthcare data is just that, an opportunity, one with potential that cannot be known until it unfolds before ones eyes.