This is Healthcare… so why the map?

This is Healthcare… so why the map?


Recently I spoke at a seminar on provider directories hosted by the California Association of Health Plans (CAHP).  An audience member and asked me “So why the map?”  It had not occurred to me that the use of a map-based interface for provider directories as well as integrated care coordination would be a novel idea.   

The use of geographical references in medicine started with Hippocrates around 485 BCE.  Hippocrates attributed physique and general demeanor to the location in which an individual was raised.  Large, hearty, and perhaps aggressive individuals came from rugged mountainous regions with distinct seasons, while shorter rounder folks hailed from the bucolic meadows of the flat low-lying areas.    Al Razi, a noted Persian Physician hung slabs of meat in various locations around Baghdad with the theory that the slab that took the longest to rot must be in a place of the cleanest healthiest air and as a result present the best location for a Hospital.  And of course most of us know of Dr.John Snow if only because he shares the name with a prominent Game of Thrones character.  Dr. Snow mapped the location of the deaths attributed to Cholera in London and then “Layered” on that the location of the city’s wells.  In doing so he was able to identify the well closest to the most deaths and had the wells handle removed.   Cholera related deaths soon abated. 

Figure 1 OnePercentNetwork

But does the introduction of a map actually help the user?  Is it worth the investment?  John Zheng wrote a  great article “Visualizing Healthcare Provider Network using SAS Tools” that really drives home the value of the geospatial representation of data.  Figure 1 represents a network involving roughly 1,100 providers.  The shades of blue represent provider types, the greater the number of patients shared by a provider with other providers the larger the node.  Nodes are connected when the two providers involved share more than 1% of their total patient load.  As John points out in his article, this figure can, “barely tell an insightful story”. 

Figure 2 shows how the network is “cleaned up” a bit when the shared patient percentage is raised to 5% or 10%.  While doing so clears things up viewing the data within this context



Figure 2   5_10Percent



hardly means anything without reading the entirety of John’s article.   I think that we can all agree that having a frame of reference enables us to better understand information.  This holds true for a user of an application, an analyst of data, an individual walking down a street, or a patient looking at a network of providers.  Figure 3 illustrates this point well. 


Figure 3   BothMaps


The introduction of a map places the data in a context all of us can relate to.  With navigation systems now standard in most vehicles mapped based interfaces have become commonplace.  Here the map allows us to easily wrap our heads around what was previously a meaningless pattern of dots and lines. 


There are an estimated 325 million people in the U.S. 251 million are over the age of eighteen.  Let’s assume half of these use an online provider directory of some sort.   That is a lot of consumers all facing the difficult task of locating health providers. And yet almost every online provider directory begins with the same drop-down boxes.  These filters allow us to par down the list of providers contracted with our health plan by type of provider, specialty, gender, accepting patients, and languages spoken.  Once entered we are shown a list which, if we are lucky, allows us to click on an entry and have it mapped.  But this must be done repeatedly for each type and specialty of provider you’re looking for. 


Figure 4




Figure 4, in contrast is a landing page with an online map that uses your location and displays color coded pins that identify all the providers within 15 miles of your home that are accepting patients with just a single click – “use my location”.   As healthcare consumers we should accept nothing less.  Health plans faced with increasing competition should do everything possible to maintain their current member base as well as gain new members.  A mapped based provider directory would be a great start. 

It's The 21st Century. Do you know where your providers are?

It’s the 21st Century, do you know where your Doctor is?


At the beginning of each year many of us are asked to stay with our current health plan or select a new one.  As a father of four trying to identify a plan that meets the needs of my family with the variety of doctors we see, prescriptions we receive, as well as our individual tastes is no simple task.  As a technologist I am certain that the tools available to select a plan, determine if plans contract with certain providers, and see ratings associated with new physicians we may need, are horribly outdated both in terms of technology and data.  Thankfully, the industry as well as federal and state authorities all agree that it is a problem.  It is the solution that may not be so easy to agree on. 


Federal legislation introduced a solution by threatening to fine plans up to $25,000 for each error in the plan’s directory.  It made me wonder if somehow the public could get in on this.  Plan directory bounty hunting came to mind.  Identify and report an entry that is in error and we will pay you $1,000.  Physicians by and large pay little attention to entries in plan directories.  However, groups that represent them like the AMA have partnered with folks who know a lot about data, such as LexisNexis, to tackle the problem.  They are using the power of technology to bounce all the providers demographic data contained in directories against other sources of data like claims, to try and reconcile the differences so that plans can identify and correct errors. 


But come on.  How hard can it be?  It’s like a phone book isn’t it?  Well, let’s do the math. There are nearly 8,000,000 providers in the United States.  A typical practice is party to 13 managed care contracts, a contract management system requires about 150 different provider data elements.  So, a small practice with five providers will be managing almost 10,000 data points.  Also, we must take into consideration churn.  How often does this data change?  Unlike a phone book, where people tend to stay in one place with the same number for an average of five years; this data changes and it changes a lot.  25% of all providers change location every year, 12% retire or pass away, and 20% incur some change to their credentialing information. Nationwide that amounts to roughly 35 billion changes a year.  Nearly 4 billion involving California providers alone.  Obviously, a promise to healthcare consumers of up to date records in near real time should not be delivered lightly. 


So, what is to be done? As already mentioned the federal government intends to levy fines against plans that have inaccurate directories. On a more constructive note several states including the state of California are working to create a single source of Provider Data that can be shared with plans and providers.    The idea being that then the plans can tap this information and provide it via their own Online and Printed Directories.  If you would like to know more about California’s efforts head here.  I think these are steps in the right direction.  However, under this approach, the data will still be delivered through the same tired platforms.  


I am a firm believer that it will be plan members, aka you and I, as consumers of health care, that will drive the industry to deliver.   If I look to be active in the coordination of my family’s care or decide to turn it over to the plans to do it. I will be looking for a single source that can hold and display my medical records, like 1upHealth , combined with something like google earth to show me where all my providers or potential providers are located along with hours of operation or and  better yet open appointment slots.  Perhaps include some quality scores and reviews, and if they offer concierge medical – their fee schedule.   At the point provider data becomes a business driver the data will become clean. I’ll be keeping you up to date. In the meantime, directory bounty hunting might be a good option.