Send To a Friend    Print friendly

Technology

Which Way RHIO?
By Dagmara Scalise

Providers must decide where they want to store patient data before joining a network

With Americans increasingly concerned about the privacy of their medical information, providers face a major question when forming regional health information organizations: Where to store patient data?

Two RHIO models are emerging, each with its own answer to that question:

  • Centralized—The entity that runs the RHIO stores patient data on its own servers.
  • Federated—Each RHIO participant stores patient information on its own system and allows other providers access.

So far, there’s no consensus among hospital leaders about which model works better. 

Joseph Casper, senior vice president, First Consulting Group, Long Beach, Calif., says that the federated model gives providers more direct control over security protocols.  “With a centralized model, someone in the RHIO would have the ability to search the database. That individual could target patient information,” Casper says. “You can’t make a single mistake with privacy, otherwise, you’ve corrupted the entire model.”

It’s also harder to keep patient records up to date in a centralized model, says Mike Skinner, executive director of the Santa Barbara (Calif.) Care Exchange, one of the nation’s first functioning RHIOs. Data has to be copied from each organization and stored on the RHIO’s server.

“It’s like the game telephone,” says  Skinner, whose group follows the federated model. “The message changes along the way. As a lab, I might update a patient’s information, but it can take six months to send that updated information to the RHIO. We want to direct users to clinical data that already exists.”

But a federated RHIO has the challenge of dealing with multiple systems, each of which differs in how it stores and transmits information. For example, there are no guarantees that individual systems remain turned on and accessible to other RHIO members at all times.

“That makes it impossible to access patient data. A glitch in a single system could produce an incomplete or erroneous medical record at a critical moment,” says John Quinn, chief technology officer in Accenture’s Provider, Health & Life Sciences practice. Accenture was awarded a grant from the Department of Health & Human Services last November to  create a prototype of a nationwide health information network. The company is working on a centralized model to connect three RHIOs in Kentucky, Tennessee and West Virginia.

And because it stores data on a single server and in a consistent format, a centralized model can also transmit data more quickly. A physician who requests patient data wants to see it instantly, within 3 to 5 seconds, but a federated RHIO may transmit the data in as many as 30 seconds. “If I’m the physcian, that’s the same as having that data unavailable. I may get so frustrated, I close out and get the information the old-fashioned way, from a paper chart,” Quinn says.

The federal government is not pushing either model, but instead is letting the market sort itself out. Some experts like Casper believe that privacy concerns will outweigh many other factors when it comes to choosing a federated versus a centralized model. Others suggest that a hybrid approach may be the answer.

At least one RHIO—in Indianapolis—has taken that approach: It stores data on seven separate servers, one for each of its participants, which allows it to better ensure data security, as in the federated approach, while taking advantage of the centralized model’s data reliability and speed.

RHIO vs. RHIO

 
Centralized    
Federated
Pros 

• Simplicity (a single entity manages the process)
• Data is consistent in appearance 
• Quicker access to data
• May be easier to bring to market because most databases with clinical data are not Web-enabled

• Greater assurance of privacy
• Proven working examples
• May be easier to get provider buy-in because the  politics of storing and accessing data are less of an issue
 Cons • Higher hardware costs
• High operational costs
• Difficult to manage scores of different RHIO participants
• Potential problems displaying data in user-friendly way since each organization has different protocol
• Delays in retrieving data for users

This article 1st appeared in the June 2006 issue of HHN Magazine.



To respond to this article, please click here.