Two national health information exchange organizations are soliciting input on ideas for improving how providers and others healthcare organizations match patients to their electronic medical records.
And no, they're not recommending a national patient identifier. That controversial approach to patient matching was side-stepped when developing the group's 55-page white paper, “A Framework for Cross-Organizational Patient Identity Management.”
The difficulty in matching patients to their records is seen by many health IT experts as a barrier to achieving full interoperability of healthcare information. It also highlights issues of patient safety and cost inefficiency.
The authors of the new white paper came up with nearly two dozen recommendations that are in part gleaned from and validated by a case study of an actual patient-matching improvement project between Intermountain Healthcare, the Utah Health Information Network, a statewide health information exchange, and other local information exchange participants.
During the project, participants improved the rate of “true matching” of records across the exchanges from 10% to 95% by identifying a key set of the most useful patient data attributes – such standards as first and last names, sex, dates of birth; but also Social Security numbers, home phone numbers, race, and home addresses – and standardizing the way those attributes are recorded.
“They were really stuck by the high failure rate, and when they got down to the root cause it had nothing to do with patient matching but data quality,” said Mariann Yeager, CEO of the Sequoia Project, a McLean, Va.- based not-for-profit that teamed with the Care Connectivity Consortium to produce the report.
For example, according to the study, telephone numbers were inconsistently recorded among exchange participants, some with dashes between number blocks and others without. Another area for improvement is consistency when recording and matching the patient's name (e.g., preferred name over legal name). Inconsistencies also arose when a person's first or middle initials were combined with their first names, and names with non-alphabetic characters, as O'Toole, or with spaces, such as Van Der Camp.
Differing formats for addresses also generate matching errors. Yeager suggests everyone use the U.S. Postal Service's format, for example.
Still, a national patient identifier, what many see as a solution to all of these problems, is not in the document.
“Of course it came up,” Yeager said. But, she added, “We felt it was important to be grounded in the practical realities of today, with the technologies at hand, so we can do better.”
A national patient identifier has been the third rail of health IT since privacy advocates persuaded the Clinton Administration and Congress in 1998 to effectively ban any federal work on development of an NPI system, which was called for in The Health Insurance Portability and Accountability Act.
A persistent corps of health IT developers, users and other true believers in an NPI, however, have attempted to resurrect it, but all attempts thus far have failed.
The Care Connectivity Consortium is a collaboration of the Mayo Clinic, Geisinger Health System, Group Health Cooperative, Intermountain Healthcare, Kaiser Permanente and OCHIN, the Portland, Ore.-based, not-for-profit health information exchange.
The group is asking for feedback on the white paper to be posted online and has set up website to receive it.
“We really are serving in a neutral, convening role,” Yeager said. “I don't know that we have a specific goal, but we think this paper is going to help us make great strides across the industry. But it's progress, not a panacea. We hope this actually informs discussions” to be had in the future.