Later this year, Baylor plans to launch the next phase of the project, which will involve checking for a patient identifier match in real time during the registration process, Schoenvogel said.

The key goal of the project is to solve one of the most vexing problems facing health care organizations looking to share patient data internally or with other entities: ensuring that data in e-health records is matched to the correct patient.

"We have to be able to link literally hundreds of different data repositories together to create that continuous record," Schoenvogel said. "We need to be sure that the information we have produces a match to another record."

So far, the project has not been without its challenges - including coming to terms with the fact that no system, including this one, can guarantee that the correct patient records will be accessed every time, Schoenvogel said.

Instead, the software provides a probability that certain records should be associated with a patient. As part of the first phase, the organization had to establish an acceptable probability level associated with different processes and put in place manual review processes for when that level is not met, he added.