The Right Data Can Address the Challenges of Rural Healthcare 

Updated on January 29, 2024
medical doctor holding senior patient's hands and comforting her

Specialty and subspecialty healthcare services are less likely to be available in rural areas and are less likely to include highly sophisticated or high-intensity care. This exacerbates problems for rural patients seeking specialized care who are faced with significant travel distances for treatment.

It comes as no surprise a 2019 policy brief from the University of Minnesota Rural Health Research Center found that 64% of surveyed rural health clinic staff members reported difficulties finding specialists for patient referral. 

A functioning rural health system relies on legions of specialty care doctors doing outreach visits across a wide geography. In theory, that’s a very effective way to ensure that rural patients have access to specialty care without traveling to a major metro area. But bad data is keeping us from achieving complete access to specialty care in rural areas. 

This “hub and spoke” model of specialty care, where a provider is based in one area but may serve a very wide geography, presents unique data challenges that require a different approach to analyzing information and modeling data. In this care model, a provider may go long periods without seeing patients at a given location simply because their services are only needed intermittently. A data model that doesn’t understand that will show that that provider is no longer serving that location.

Without taking this business model into account, a payer may underestimate the amount of network availability to serve members in a specific region. That’s a problem because you’d be underestimating the network when the region is already underserved in terms of provider network. The last thing we want to do is make it appear that there are network adequacy gaps where they may not exist.

Part of the problem is that legacy manual attestation practices can’t easily track all of the locations these rural providers can serve. Current data practices would show a provider at his or her primary location and would likely not show them as a referral option in the rural areas he or she serves. The system (and the data feeding that system) has to know a provider could practice at a location and is taking appointments. 

Even using typical AI models, the location information isn’t accurate because it isn’t reflective of a health insurance plan’s whole network or a provider’s ability to see patients. Once appointments in an area stop, AI may say “this provider isn’t at this location anymore, do not offer this provider or location” when really, they’ll be back in a month and have appointments available for patients to book.

The answer to the issues around rural health providers is simple yet complicated: get more data. This isn’t the “easy” data to deal with. These are the unruly problems no one wants to address. Providers share all sorts of ambient data all the time from state and federal licensures, registrations, claims patterns, and more. These data points all can be triangulated to create a more accurate picture of the entire geography that a provider serves—not just their home base or frequently visited locations.

The good news is that all of this data can be gleaned from existing sources—there is no need for time-consuming and error-prone manual calls.

Any health plan and payer that is serving rural populations must take these data challenges into account. Not only because it’s the right thing to do to achieve member access and satisfaction but because we know unsatisfied members are likely to switch health plans. Look at the number of available Medicare Advantage plans, there are plenty of competitors if your plan can’t serve a rural population. 

Every health plan should ask themselves: Can my technology detect and track providers that work at multiple locations? Can it reflect accurate provider contact information? If the answer to either of those questions is no, it’s time to act.

Bob Lindner grey backdrop headshot copy
Dr. Bob Lindner
Chief Science and Technology Officer at Veda

Dr. Bob Lindner is Veda’s Chief Science and Technology Officer, co-founder, and a full-stack data scientist, who specializes in cloud-based machine learning systems. He earned his PhD in Astrophysics from Rutgers University. As a post-doctoral researcher in radio astronomy at the University of Wisconsin Madison, Bob invented Gauspy, an AI program used at NASA and 40 other research facilities across the globe.

Author of five technology patents on AI, entity resolution, and machine learning, Bob also has over 16 years of experience writing and publishing scientific and academic papers in the artificial intelligence field.